thdltools/Jskad
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity
Jskad/source/org/thdl/tib/text/ttt/ACIPTshegBarScanner.java

1053 lines
54 KiB
Java
Raw Normal View History

I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
/*
The contents of this file are subject to the THDL Open Community License
Version 1.0 (the "License"); you may not use this file except in compliance
with the License. You may obtain a copy of the License on the THDL web site
(http://www.thdl.org/).
Software distributed under the License is distributed on an "AS IS" basis,
WITHOUT WARRANTY OF ANY KIND, either express or implied. See the
License for the specific terms governing rights and limitations under the
License.
The Initial Developer of this software is the Tibetan and Himalayan Digital
Library (THDL). Portions created by the THDL are Copyright 2003 THDL.
All Rights Reserved.
Contributor(s): ______________________________________.
*/
package org.thdl.tib.text.ttt;
Two things: One, TMW->EWTS gives dbas and dngas instead of dabs and dangs because Chris Fynn's e-mail from today has dbas and dngas. Second, Down with ACIPRules. Long live ACIPTraits. EWTS->Tibetan conversion is closer still.
2005-02-22 04:36:54 +00:00
import java.io.IOException;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
import java.util.ArrayList;
import java.util.Stack;
ACIP->TMW and ACIP->Unicode now allow for Unicode escapes like K\u0F84. This means that the lack of support for ACIP's backslash, '\\', is mitigated because you can turn ACIP {K\} into ACIP {K\u0F84}. Support for U+F021-U+F0FF, the PUA that the latest EWTS uses, is not provided.
2003-11-29 22:56:18 +00:00
import org.thdl.util.ThdlOptions;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
/**
Refactored so that there can be an EWTS scanner and an ACIP scanner.
2005-02-21 05:37:01 +00:00
* This singleton class is able to break up Strings of ACIP text (for
* example, an entire sutra file) into tsheg bars, comments, etc. Folio
* markers, comments, and the like are segregated (so that consumers
* can ensure that they remain in Latin), and Tibetan passages are
* broken up into tsheg bars.
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
*
* <p><b>FIXME:</b> We should be handling {KA\n\nKHA} vs. {KA\nKHA} in
* the parser, not here in the lexical analyzer. That'd be cleaner,
* and more like how you'd do things if you used lex and yacc.
*
Two things: One, TMW->EWTS gives dbas and dngas instead of dabs and dangs because Chris Fynn's e-mail from today has dbas and dngas. Second, Down with ACIPRules. Long live ACIPTraits. EWTS->Tibetan conversion is closer still.
2005-02-22 04:36:54 +00:00
* This is not public because you should use {@link ACIPTraits#scanner()}.
*
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
* @author David Chandler */
Two things: One, TMW->EWTS gives dbas and dngas instead of dabs and dangs because Chris Fynn's e-mail from today has dbas and dngas. Second, Down with ACIPRules. Long live ACIPTraits. EWTS->Tibetan conversion is closer still.
2005-02-22 04:36:54 +00:00
class ACIPTshegBarScanner extends TTshegBarScanner {
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
/** True if those ACIP snippets inside square brackets (e.g.,
"[THIS]") are to be passed through into the output unmodified
while retaining the brackets and if those ACIP snippets inside
curly brackets (e.g., "{THAT}") are to be passed through into
the output unmodified while dropping the brackets. (Nesting
of brackets is not allowed regardless.) */
public static final boolean BRACKETED_SECTIONS_PASS_THROUGH_UNMODIFIED
= true; // Robert Chilton's e-mail from April 2004 calls for 'true'
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
/** Useful for testing. Gives error messages on standard output
* about why we can't scan the document perfectly and exits with
* non-zero return code, or says "Good scan!" otherwise and exits
* with code zero. <p>FIXME: not so efficient; copies the whole
* file into memory first. */
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
public static void main(String[] args) throws IOException {
Jskad's converter now has ACIP-to-Unicode built in. There are known bugs; it is pre-alpha. It's usable, though, and finds tons of errors in ACIP input files, with the user deciding just how pedantic to be. The biggest outstanding bug is the silent one: treating { }, space, as tsheg instead of whitespace when we ought to know better.
2003-08-24 06:40:53 +00:00
if (args.length != 1) {
System.out.println("Bad args! Need just the name of the ACIP text file.");
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
System.exit(1);
}
StringBuffer errors = new StringBuffer();
Using less memory and time to figure out if warnings occurred.
2003-12-14 07:41:15 +00:00
int maxErrors = 1000;
Refactored so that there can be an EWTS scanner and an ACIP scanner.
2005-02-21 05:37:01 +00:00
ArrayList al = instance().scanFile(args[0], errors, maxErrors - 1,
"true".equals(System.getProperty("org.thdl.tib.text.ttt.ACIPTshegBarScanner.shortMessages")),
"All" /* memory hog */);
Fixed a couple of small bugs. Only 250 errors are reported now; this is important if you try to convert an English document.
2003-08-17 02:12:49 +00:00
if (null == al) {
System.out.println(maxErrors + " or more errors occurred while scanning ACIP input file; is this");
System.out.println("Tibetan or English input?");
System.out.println("");
System.out.println("First " + maxErrors + " errors scanning ACIP input file: ");
System.out.println(errors);
System.out.println("Exiting with " + maxErrors + " or more errors; please fix input file and try again.");
System.exit(1);
}
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
if (errors.length() > 0) {
System.out.println("Errors scanning ACIP input file: ");
System.out.println(errors);
System.out.println("Exiting; please fix input file and try again.");
System.exit(1);
}
System.out.println("Good scan!");
System.exit(0);
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
/** Helper. Here because ACIP {MTHAR%\nKHA} should be treated the
same w.r.t. tsheg insertion regardless of the lex errors and
lex warnings found. */
{DAN,\nLHAG} used to be treated like {DAN, LHAG} but that got broken. Fixed. Added tests for lexer's handling of ACIP spaces etc.
2003-12-10 06:55:16 +00:00
private static boolean lastNonExceptionalThingWasAdornmentOr(ArrayList al, int kind) {
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
int i = al.size() - 1;
while (i >= 0 && (((TString)al.get(i)).getType() == TString.WARNING
|| ((TString)al.get(i)).getType() == TString.ERROR))
--i;
return (i >= 0 && // FIXME: or maybe i < 0 || ...
{DAN,\nLHAG} used to be treated like {DAN, LHAG} but that got broken. Fixed. Added tests for lexer's handling of ACIP spaces etc.
2003-12-10 06:55:16 +00:00
(((TString)al.get(i)).getType() == kind
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
|| ((TString)al.get(i)).getType() == TString.TSHEG_BAR_ADORNMENT));
}
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
/** Helper function that increments numErrorsArray[0] by one and
adds an ERROR to the end of al and appends to the end of
errors if it is nonnull. (Nothing else is mutated.)
@return true if and only if the error count has gone too high
and caller should abort scanning */
private static boolean queueError(int code,
String translit,
boolean shortMessages,
int i,
int numNewlines,
int maxErrors,
ArrayList al,
StringBuffer errors,
int numErrorsArray[]) {
String errMsg;
al.add(new TString("ACIP",
errMsg = ErrorsAndWarnings.getMessage(code,
shortMessages,
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
translit,
ACIPTraits.instance()),
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
TString.ERROR));
if (null != errors)
errors.append("Offset " + ((i < 0) ? "END" : ("" + i))
+ ((numNewlines == 0)
? ""
: (" or maybe " + (i-numNewlines)))
+ ": ERROR "
+ errMsg + "\n");
if (maxErrors >= 0 && ++numErrorsArray[0] >= maxErrors)
return true;
else
return false;
}
{DAN,\nLHAG} used to be treated like {DAN, LHAG} but that got broken. Fixed. Added tests for lexer's handling of ACIP spaces etc.
2003-12-10 06:55:16 +00:00
// DLC FIXME "H:\n\n" becomes "H: \n\n", wrongly I think. See
// Tibetan! 5.1 section on formatting Tibetan texts.
Refactored so that there can be an EWTS scanner and an ACIP scanner.
2005-02-21 05:37:01 +00:00
/** See the comment in TTshegBarScanner. And note that this not
* only scans; it finds all the errors and warnings a parser
* would too, like "NYA x" and "(" and ")" and "/NYA" etc. */
public ArrayList scan(String s, StringBuffer errors, int maxErrors,
boolean shortMessages, String warningLevel) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
// FIXME: Use less memory and time by not adding in the
// warnings that are below threshold.
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
// the size depends on whether it's mostly Tibetan or mostly
// Latin and a number of other factors. This is meant to be
// an underestimate, but not too much of an underestimate.
ArrayList al = new ArrayList(s.length() / 10);
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
int numErrorsArray[] = new int[] { 0 };
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
boolean waitingForMatchingIllegalClose = false;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
int sl = s.length();
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
int currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
int startOfString = 0;
Stack bracketTypeStack = new Stack();
int startSlashIndex = -1;
int startParenIndex = -1;
TMW->Wylie conversion now takes advantage of prefix rules, the rules that say "ya can take a ga prefix" etc. The ACIP->Unicode converter now gives warnings (optionally, and by default, inline). This converter now produces output even when lexical errors occur, but the output has errors and warnings inline.
2003-08-23 22:03:37 +00:00
int numNewlines = 0;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
for (int i = 0; i < sl; i++) {
if (i < startOfString) throw new Error("bad reset");
char ch;
ch = s.charAt(i);
TMW->Wylie conversion now takes advantage of prefix rules, the rules that say "ya can take a ga prefix" etc. The ACIP->Unicode converter now gives warnings (optionally, and by default, inline). This converter now produces output even when lexical errors occur, but the output has errors and warnings inline.
2003-08-23 22:03:37 +00:00
if (ch == '\n') ++numNewlines;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
if (TString.COMMENT == currentType && ch != ']') {
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
if ('[' == ch) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(102, "" + ch,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
continue;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
switch (ch) {
ACIP->Tibetan converters now warn every time {%} is encountered that U+0F14 might've been intended. The Unicode for ACIP {o} is U+0F37.
2003-11-09 23:15:58 +00:00
case '}': // fall through...
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
case ']':
if (bracketTypeStack.empty()) {
// Error.
if (startOfString < i) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, i),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
if (!waitingForMatchingIllegalClose) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(103, "" + ch,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
}
waitingForMatchingIllegalClose = false;
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
if (queueError(BRACKETED_SECTIONS_PASS_THROUGH_UNMODIFIED ? 140 : 104,
"" + ch,
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i+1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
} else {
int stackTop = ((Integer)bracketTypeStack.pop()).intValue();
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
int end = startOfString;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
if (TString.CORRECTION_START == stackTop) {
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
// This definitely indicates a new token.
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
char prevCh = s.charAt(i-1);
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
if (prevCh == '?')
end = i - 1;
else
end = i;
if (startOfString < end) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, end),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
if ('?' != prevCh) {
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.PROBABLE_CORRECTION;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
} else {
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.POSSIBLE_CORRECTION;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
}
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(end, i+1), currentType));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i+1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
break; // end ']','}' case
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
case '{': // NOTE WELL: KX0016I.ACT, KD0095M.ACT, and a
// host of other ACIP files use {} brackets like
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
// [] brackets. I treat both the same if
// BRACKETED_SECTIONS_PASS_THROUGH_UNMODIFIED
// is false.
ACIP->Tibetan converters now warn every time {%} is encountered that U+0F14 might've been intended. The Unicode for ACIP {o} is U+0F37.
2003-11-09 23:15:58 +00:00
// fall through...
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
case '[':
// This definitely indicates a new token.
if (startOfString < i) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, i),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
if (BRACKETED_SECTIONS_PASS_THROUGH_UNMODIFIED) {
int indexPastCloseBracket = i;
boolean foundClose = false;
while (++indexPastCloseBracket < sl) {
if ((('[' == ch) ? '[' : '{')
== s.charAt(indexPastCloseBracket)) { // "[i am [nested], you see]" is not allowed.
waitingForMatchingIllegalClose = true;
if (queueError(141, "" + ch,
shortMessages, indexPastCloseBracket, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
} else if ((('[' == ch) ? ']' : '}') == s.charAt(indexPastCloseBracket)) {
al.add(new TString("ACIP",
s.substring(startOfString + (('[' == ch) ? 0 : 1),
indexPastCloseBracket + (('[' == ch) ? 1 : 0)),
TString.LATIN));
startOfString = indexPastCloseBracket + 1;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
i = startOfString - 1;
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
currentType = TString.ERROR;
foundClose = true;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
break;
}
}
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
if (!foundClose) {
// FIXME: duplciated code, search for 106:
{
String inContext = s.substring(i, i+Math.min(sl-i, 10));
if (inContext.indexOf("\r") >= 0) {
inContext = inContext.substring(0, inContext.indexOf("\r"));
} else if (inContext.indexOf("\n") >= 0) {
inContext = inContext.substring(0, inContext.indexOf("\n"));
} else {
if (sl-i > 10) {
inContext = inContext + "...";
}
}
if (queueError(139, inContext,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
}
if (queueError(117, "-*-END OF FILE-*-",
shortMessages, -1, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
// we're done here:
{
i = sl;
startOfString = sl;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
}
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
}
} else {
String thingy = null;
if (i + "[DD]".length() <= sl
&& (s.substring(i, i + "[DD]".length()).equals("[DD]")
|| s.substring(i, i + "[DD]".length()).equals("{DD}"))) {
thingy = "[DD]";
currentType = TString.DD;
} else if (i + "[DD1]".length() <= sl
&& (s.substring(i, i + "[DD1]".length()).equals("[DD1]")
|| s.substring(i, i + "[DD1]".length()).equals("{DD1}"))) {
thingy = "[DD1]";
currentType = TString.DD;
} else if (i + "[DD2]".length() <= sl
&& (s.substring(i, i + "[DD2]".length()).equals("[DD2]")
|| s.substring(i, i + "[DD2]".length()).equals("{DD2}"))) {
thingy = "[DD2]";
currentType = TString.DD;
} else if (i + "[DDD]".length() <= sl
&& (s.substring(i, i + "[DDD]".length()).equals("[DDD]")
|| s.substring(i, i + "[DDD]".length()).equals("{DDD}"))) {
thingy = "[DDD]";
currentType = TString.DD;
} else if (i + "[DR]".length() <= sl
&& (s.substring(i, i + "[DR]".length()).equals("[DR]")
|| s.substring(i, i + "[DR]".length()).equals("{DR}"))) {
thingy = "[DR]";
currentType = TString.DR;
} else if (i + "[LS]".length() <= sl
&& (s.substring(i, i + "[LS]".length()).equals("[LS]")
|| s.substring(i, i + "[LS]".length()).equals("{LS}"))) {
thingy = "[LS]";
currentType = TString.LS;
} else if (i + "[BP]".length() <= sl
&& (s.substring(i, i + "[BP]".length()).equals("[BP]")
|| s.substring(i, i + "[BP]".length()).equals("{BP}"))) {
thingy = "[BP]";
currentType = TString.BP;
} else if (i + "[BLANK PAGE]".length() <= sl
&& (s.substring(i, i + "[BLANK PAGE]".length()).equals("[BLANK PAGE]")
|| s.substring(i, i + "[BLANK PAGE]".length()).equals("{BLANK PAGE}"))) {
thingy = "[BLANK PAGE]";
currentType = TString.BP;
} else if (i + "[ BP ]".length() <= sl
&& (s.substring(i, i + "[ BP ]".length()).equals("[ BP ]")
|| s.substring(i, i + "[ BP ]".length()).equals("{ BP }"))) {
thingy = "{ BP }"; // found in TD3790E2.ACT
currentType = TString.BP;
} else if (i + "[ DD ]".length() <= sl
&& (s.substring(i, i + "[ DD ]".length()).equals("[ DD ]")
|| s.substring(i, i + "[ DD ]".length()).equals("{ DD }"))) {
thingy = "{ DD }"; // found in TD3790E2.ACT
currentType = TString.DD;
} else if (i + "[?]".length() <= sl
&& (s.substring(i, i + "[?]".length()).equals("[?]")
|| s.substring(i, i + "[?]".length()).equals("{?}"))) {
thingy = "[?]";
currentType = TString.QUESTION;
} else {
// We see comments appear not as [#COMMENT], but
// as [COMMENT] sometimes. We make special cases
// for some English comments. There's no need to
// make this mechanism extensible, because you
// can easily edit the ACIP text so that it uses
// [#COMMENT] notation instead of [COMMENT].
String[] englishComments = new String[] {
"FIRST", "SECOND", // S5274I.ACT
"Additional verses added by Khen Rinpoche here are", // S0216M.ACT
"ADDENDUM: The text of", // S0216M.ACT
"END OF ADDENDUM", // S0216M.ACT
"Some of the verses added here by Khen Rinpoche include:", // S0216M.ACT
"Note that, in the second verse, the {YUL LJONG} was orignally {GANG LJONG},\nand is now recited this way since the ceremony is not only taking place in Tibet.", // S0216M.ACT
"Note that, in the second verse, the {YUL LJONG} was orignally {GANG LJONG},\r\nand is now recited this way since the ceremony is not only taking place in Tibet.", // S0216M.ACT
"text missing", // S6954E1.ACT
"INCOMPLETE", // TD3817I.INC
"MISSING PAGE", // S0935m.act
"MISSING FOLIO", // S0975I.INC
"UNCLEAR LINE", // S0839D1I.INC
"THE FOLLOWING TEXT HAS INCOMPLETE SECTIONS, WHICH ARE ON ORDER", // SE6260A.INC
"@DATA INCOMPLETE HERE", // SE6260A.INC
"@DATA MISSING HERE", // SE6260A.INC
"LINE APPARENTLY MISSING THIS PAGE", // TD4035I.INC
"DATA INCOMPLETE HERE", // TD4226I2.INC
"DATA MISSING HERE", // just being consistent
"FOLLOWING SECTION WAS NOT AVAILABLE WHEN THIS EDITION WAS\nPRINTED, AND IS SUPPLIED FROM ANOTHER, PROBABLY THE ORIGINAL:", // S0018N.ACT
"FOLLOWING SECTION WAS NOT AVAILABLE WHEN THIS EDITION WAS\r\nPRINTED, AND IS SUPPLIED FROM ANOTHER, PROBABLY THE ORIGINAL:", // S0018N.ACT
"THESE PAGE NUMBERS RESERVED IN THIS EDITION FOR PAGES\nMISSING FROM ORIGINAL ON WHICH IT WAS BASED", // S0018N.ACT
"THESE PAGE NUMBERS RESERVED IN THIS EDITION FOR PAGES\r\nMISSING FROM ORIGINAL ON WHICH IT WAS BASED", // S0018N.ACT
"PAGE NUMBERS RESERVED FROM THIS EDITION FOR MISSING\nSECTION SUPPLIED BY PRECEDING", // S0018N.ACT
"PAGE NUMBERS RESERVED FROM THIS EDITION FOR MISSING\r\nSECTION SUPPLIED BY PRECEDING", // S0018N.ACT
"SW: OK", // S0057M.ACT
"m:ok", // S0057M.ACT
"A FIRST ONE\nMISSING HERE?", // S0057M.ACT
"A FIRST ONE\r\nMISSING HERE?", // S0057M.ACT
"THE INITIAL PART OF THIS TEXT WAS INPUT BY THE SERA MEY LIBRARY IN\nTIBETAN FONT AND NEEDS TO BE REDONE BY DOUBLE INPUT", // S0195A1.INC
"THE INITIAL PART OF THIS TEXT WAS INPUT BY THE SERA MEY LIBRARY IN\r\nTIBETAN FONT AND NEEDS TO BE REDONE BY DOUBLE INPUT", // S0195A1.INC
};
boolean foundOne = false;
for (int ec = 0; ec < englishComments.length; ec++) {
if (i + 2 + englishComments[ec].length() <= sl
&& (s.substring(i, i + 2 + englishComments[ec].length()).equals("[" + englishComments[ec] + "]")
|| s.substring(i, i + 2 + englishComments[ec].length()).equals("[" + englishComments[ec] + "]"))) {
al.add(new TString("ACIP", "[#" + englishComments[ec] + "]",
TString.COMMENT));
startOfString = i + 2 + englishComments[ec].length();
i = startOfString - 1;
foundOne = true;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
break;
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
}
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
if (!foundOne && i+1 < sl && s.charAt(i+1) == '*') {
// Identify [*LINE BREAK?] as an English
// correction. Every correction not on this
// list is considered to be Tibetan.
// FIXME: make this extensible via a config
// file or at least a System property (which
// could be a comma-separated list of these
// creatures.
// If "LINE" is in the list below, then [*
// LINE], [* LINE?], [*LINE], [*LINE?], [*
// LINE OUT ?], etc. will be considered
// English corrections. I.e., whitespace
// before and anything after doesn't prevent a
// match.
String[] englishCorrections = new String[] {
"LINE", // KD0001I1.ACT
"DATA", // KL0009I2.INC
"BLANK", // KL0009I2.INC
"NOTE", // R0001F.ACM
"alternate", // R0018F.ACE
"02101-02150 missing", // R1003A3.INC
"51501-51550 missing", // R1003A52.ACT
"BRTAGS ETC", // S0002N.ACT
"TSAN, ETC", // S0015N.ACT
"SNYOMS, THROUGHOUT", // S0016N.ACT
"KYIS ETC", // S0019N.ACT
"MISSING", // S0455M.ACT
"this", // S6850I1B.ALT
"THIS", // S0057M.ACT
};
int begin;
for (begin = i+2; begin < sl; begin++) {
if (!isWhitespace(s.charAt(begin)))
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
break;
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
}
int end;
for (end = i+2; end < sl; end++) {
if (s.charAt(end) == ']')
break;
}
int realEnd = end;
if (end < sl && s.charAt(end-1) == '?')
--realEnd;
if (end < sl && begin < realEnd) {
String interestingSubstring
= s.substring(begin, realEnd);
for (int ec = 0; ec < englishCorrections.length; ec++) {
if (interestingSubstring.startsWith(englishCorrections[ec])) {
al.add(new TString("ACIP", s.substring(i, i+2),
TString.CORRECTION_START));
al.add(new TString("ACIP", s.substring(i+2, realEnd),
TString.LATIN));
if (s.charAt(end - 1) == '?') {
al.add(new TString("ACIP", s.substring(end-1, end+1),
TString.POSSIBLE_CORRECTION));
} else {
al.add(new TString("ACIP", s.substring(end, end+1),
TString.PROBABLE_CORRECTION));
}
foundOne = true;
startOfString = end+1;
i = startOfString - 1;
break;
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
}
}
}
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
if (foundOne)
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
break;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
}
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
if (null != thingy) {
al.add(new TString("ACIP", thingy,
currentType));
startOfString = i + thingy.length();
i = startOfString - 1;
} else {
if (i + 1 < sl) {
char nextCh = s.charAt(i+1);
if ('*' == nextCh) {
currentType = TString.CORRECTION_START;
bracketTypeStack.push(new Integer(currentType));
al.add(new TString("ACIP", s.substring(i, i+2),
TString.CORRECTION_START));
currentType = TString.ERROR;
startOfString = i+2;
i = startOfString - 1;
break;
} else if ('#' == nextCh) {
currentType = TString.COMMENT;
bracketTypeStack.push(new Integer(currentType));
break;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
}
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
}
It is now a compile-time option whether to treat []- and {}-bracketed sequences as text to be passed through (without the brackets in the case of {}) literally, which is the case by default because Robert Chilton requested it, or the old, ad-hoc mechanism which could be useful for finding some ugly input. Made a couple of error messages a little more verbose now that we have short-message mode.
2004-06-06 21:39:06 +00:00
// This is an error. Sometimes [COMMENTS APPEAR
// WITHOUT # MARKS]. Though "... [" could cause
// this too.
if (waitingForMatchingIllegalClose) {
if (queueError(105, "" + ch,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
}
waitingForMatchingIllegalClose = true;
// FIXME: duplciated code, search for 139:
{
String inContext = s.substring(i, i+Math.min(sl-i, 10));
if (inContext.indexOf("\r") >= 0) {
inContext = inContext.substring(0, inContext.indexOf("\r"));
} else if (inContext.indexOf("\n") >= 0) {
inContext = inContext.substring(0, inContext.indexOf("\n"));
} else {
if (sl-i > 10) {
inContext = inContext + "...";
}
}
if (queueError(106, inContext,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
}
startOfString = i + 1;
currentType = TString.ERROR;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
break; // end '[','{' case
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
case '@':
// This definitely indicates a new token.
if (startOfString < i) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, i),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
// We look for {@N{AB}, @NN{AB}, ..., @NNNNNN{AB}},
// {@[N{AB}], @[NN{AB}], ..., @[NNNNNN{AB}]},
// {@N{AB}.N, @NN{AB}.N, ..., @NNNNNN{AB}.N}, {@N,
// @NN, ..., @NNNNNN}, and {@{AB}N, @{AB}NN,
// ... @{AB}NNNNNN} only, that is from one to six
// digits. Each of these folio marker format occurs
// in practice.
for (int numdigits = 6; numdigits >= 1; numdigits--) {
// @NNN{AB} and @NNN{AB}.N cases:
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
if (i+numdigits+1 < sl
&& (s.charAt(i+numdigits+1) == 'A' || s.charAt(i+numdigits+1) == 'B')) {
boolean allAreNumeric = true;
for (int k = 1; k <= numdigits; k++) {
if (!isNumeric(s.charAt(i+k))) {
allAreNumeric = false;
break;
}
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
if (allAreNumeric) {
// Is this "@012B " or "@012B.3 "?
int extra;
if (i+numdigits+2 < sl && s.charAt(i+numdigits+2) == '.') {
if (!(i+numdigits+4 < sl && isNumeric(s.charAt(i+numdigits+3))
&& !isNumeric(s.charAt(i+numdigits+4)))) {
String inContext = s.substring(i, i+Math.min(sl-i, 10));
if (inContext.indexOf("\r") >= 0) {
inContext = inContext.substring(0, inContext.indexOf("\r"));
} else if (inContext.indexOf("\n") >= 0) {
inContext = inContext.substring(0, inContext.indexOf("\n"));
} else {
if (sl-i > 10) {
inContext = inContext + "...";
}
}
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(107, inContext,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
startOfString = i+numdigits+3;
i = startOfString - 1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
break;
}
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (i+numdigits+4 < sl
&& (s.charAt(i+numdigits+4) == '.'
|| s.charAt(i+numdigits+4) == 'A'
|| s.charAt(i+numdigits+4) == 'B'
|| s.charAt(i+numdigits+4) == 'a'
|| s.charAt(i+numdigits+4) == 'b'
|| isNumeric(s.charAt(i+numdigits+4)))) {
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
String inContext = s.substring(i, i+Math.min(sl-i, 10));
if (inContext.indexOf("\r") >= 0) {
inContext = inContext.substring(0, inContext.indexOf("\r"));
} else if (inContext.indexOf("\n") >= 0) {
inContext = inContext.substring(0, inContext.indexOf("\n"));
} else {
if (sl-i > 10) {
inContext = inContext + "...";
}
}
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(108, inContext,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
ACIP->Unicode was broken for KshR, ndRY, ndY, YY, and RY -- those stacks that use full-form subjoined RA and YA consonants. ACIP {RVA} was converting to the wrong things. The TMW for {RVA} was converting to the wrong ACIP. Checked all the 'DLC' tags in the ttt (ACIP->Tibetan) package.
2003-11-09 01:07:45 +00:00
startOfString = i+1; // FIXME: skip over more? test this code.
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
break;
}
extra = 4;
} else {
extra = 2;
}
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(i, i+numdigits+extra),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
TString.FOLIO_MARKER));
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
startOfString = i+numdigits+extra;
i = startOfString - 1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
break;
}
}
// @{AB}NNN case:
if (i+numdigits+1 < sl
&& (s.charAt(i+1) == 'A' || s.charAt(i+1) == 'B')) {
boolean allAreNumeric = true;
for (int k = 1; k <= numdigits; k++) {
if (!isNumeric(s.charAt(i+1+k))) {
allAreNumeric = false;
break;
}
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
if (allAreNumeric) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(i, i+numdigits+2),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
TString.FOLIO_MARKER));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i+numdigits+2;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
i = startOfString - 1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
break;
}
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
// @[NNN{AB}] case:
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
if (i+numdigits+3 < sl
&& s.charAt(i+1) == '[' && s.charAt(i+numdigits+3) == ']'
&& (s.charAt(i+numdigits+2) == 'A' || s.charAt(i+numdigits+2) == 'B')) {
boolean allAreNumeric = true;
for (int k = 1; k <= numdigits; k++) {
if (!isNumeric(s.charAt(i+1+k))) {
allAreNumeric = false;
break;
}
}
if (allAreNumeric) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(i, i+numdigits+4),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
TString.FOLIO_MARKER));
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
startOfString = i+numdigits+4;
i = startOfString - 1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
break;
}
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
// This case, @NNN, must come after the @NNN{AB} case.
TMW->Wylie conversion now takes advantage of prefix rules, the rules that say "ya can take a ga prefix" etc. The ACIP->Unicode converter now gives warnings (optionally, and by default, inline). This converter now produces output even when lexical errors occur, but the output has errors and warnings inline.
2003-08-23 22:03:37 +00:00
if (i+numdigits+1 < sl && (s.charAt(i+numdigits+1) == ' '
|| s.charAt(i+numdigits+1) == '\n'
|| s.charAt(i+numdigits+1) == '\r')) {
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
boolean allAreNumeric = true;
for (int k = 1; k <= numdigits; k++) {
if (!isNumeric(s.charAt(i+k))) {
allAreNumeric = false;
break;
}
}
if (allAreNumeric) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(i, i+numdigits+1),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
TString.FOLIO_MARKER));
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
startOfString = i+numdigits+1;
i = startOfString - 1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
break;
}
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
if (startOfString == i) {
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
String inContext = s.substring(i, i+Math.min(sl-i, 10));
if (inContext.indexOf("\r") >= 0) {
inContext = inContext.substring(0, inContext.indexOf("\r"));
} else if (inContext.indexOf("\n") >= 0) {
inContext = inContext.substring(0, inContext.indexOf("\n"));
} else {
if (sl-i > 10) {
inContext = inContext + "...";
}
}
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(109, inContext,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i+1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
break; // end '@' case
case '/':
// This definitely indicates a new token.
if (startOfString < i) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, i),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
if (startSlashIndex >= 0) {
ACIP->Unicode, without going through TMW, is now possible, so long as \, the Sanskrit virama, is not used. Of the 1370-odd ACIP texts I've got here, about 57% make it through the gauntlet (fewer if you demand a vowel or disambiguator on every stack of a non-Tibetan tsheg bar).
2003-08-18 02:38:54 +00:00
if (startSlashIndex + 1 == i) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(110, "" + ch,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
ACIP->Unicode, without going through TMW, is now possible, so long as \, the Sanskrit virama, is not used. Of the 1370-odd ACIP texts I've got here, about 57% make it through the gauntlet (fewer if you demand a vowel or disambiguator on every stack of a non-Tibetan tsheg bar).
2003-08-18 02:38:54 +00:00
}
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(i, i+1),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
TString.END_SLASH));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i+1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startSlashIndex = -1;
} else {
startSlashIndex = i;
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(i, i+1),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
TString.START_SLASH));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i+1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
break; // end '/' case
case '(':
case ')':
// This definitely indicates a new token.
if (startOfString < i) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, i),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
// We do not support nesting like (NYA (BA)).
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
if (startParenIndex >= 0) {
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
if (ch == '(') {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(111, "" + ch,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
} else {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(i, i+1), TString.END_PAREN));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startParenIndex = -1;
}
startOfString = i+1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
} else {
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
if (ch == ')') {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(112, "" + ch,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
} else {
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startParenIndex = i;
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(i, i+1), TString.START_PAREN));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
startOfString = i+1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
break; // end '(',')' case
case '?':
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
if (bracketTypeStack.empty() || i+1>=sl
|| (s.charAt(i+1) != ']' && s.charAt(i+1) != '}')) {
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
// The tsheg bar ends here; new token.
if (startOfString < i) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, i),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
}
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(113, "" + ch,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
startOfString = i+1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
} // else this is [*TR'A ?] or the like.
break; // end '?' case
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
case '.':
// This definitely indicates a new token.
if (startOfString < i) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, i),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
startOfString = i;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
// . is used for a non-breaking tsheg, such as in
Jskad's converter now has ACIP-to-Unicode built in. There are known bugs; it is pre-alpha. It's usable, though, and finds tons of errors in ACIP input files, with the user deciding just how pedantic to be. The biggest outstanding bug is the silent one: treating { }, space, as tsheg instead of whitespace when we ought to know better.
2003-08-24 06:40:53 +00:00
// {NGO.,} and {....,DAM}. We give a warning unless ,
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
// or ., or [A-Za-z] follows '.'.
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(i, i+1),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
TString.TIBETAN_PUNCTUATION));
I don't think warning level "None" was really doing the trick. Fixed that. You can now customize the severities of all warnings, even 504 and 510. When warning level is "None", scanning, i.e. lexical analysis, is faster.
2004-04-25 00:37:57 +00:00
if (ErrorsAndWarnings.isEnabled(510, warningLevel)
&& (!(i + 1 < sl
&& (s.charAt(i+1) == '.' || s.charAt(i+1) == ','
|| (s.charAt(i+1) == '\r' || s.charAt(i+1) == '\n')
|| (s.charAt(i+1) >= 'a' && s.charAt(i+1) <= 'z')
|| (s.charAt(i+1) >= 'A' && s.charAt(i+1) <= 'Z'))))) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
al.add(new TString("ACIP",
ErrorsAndWarnings.getMessage(510,
shortMessages,
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
"" + ch,
ACIPTraits.instance()),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
TString.WARNING));
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
startOfString = i+1;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
break; // end '.' case
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
// Classic tsheg bar enders:
case ' ':
case '\t':
case '\r':
case '\n':
case ',':
case '*':
case ';':
case '`':
case '#':
The ACIP {NYA%} is supported. {NYAo} and {NYAx} are confusing to me, because I don't know which glyphs o and x correspond to. For that reason, they cause ERRORs. The proposed THDL Extended Wylie ~X and X is now used for U+0F35 and U+0F37 respectively.
2003-09-07 16:19:50 +00:00
case '%':
case 'x':
case 'o':
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
case '^':
ACIP {&}, as in {KO&HAm,}, is supported.
2003-11-30 02:18:59 +00:00
case '&':
ACIP->TMW and ACIP->Unicode are now smart about when a newline is really a newline and when a space is really a tsheg. The space in {KA ,MDO} is a tsheg, but the space in {GA ,MDO} is not.
2003-09-04 04:04:21 +00:00
The ACIP {NYA%} is supported. {NYAo} and {NYAx} are confusing to me, because I don't know which glyphs o and x correspond to. For that reason, they cause ERRORs. The proposed THDL Extended Wylie ~X and X is now used for U+0F35 and U+0F37 respectively.
2003-09-07 16:19:50 +00:00
boolean legalTshegBarAdornment = false;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
// The tsheg bar ends here; new token.
if (startOfString < i) {
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
if (currentType == TString.TIBETAN_NON_PUNCTUATION
The ACIP {NYA%} is supported. {NYAo} and {NYAx} are confusing to me, because I don't know which glyphs o and x correspond to. For that reason, they cause ERRORs. The proposed THDL Extended Wylie ~X and X is now used for U+0F35 and U+0F37 respectively.
2003-09-07 16:19:50 +00:00
&& isTshegBarAdornment(ch))
legalTshegBarAdornment = true;
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, i),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
ACIP->TMW and ACIP->Unicode are now smart about when a newline is really a newline and when a space is really a tsheg. The space in {KA ,MDO} is a tsheg, but the space in {GA ,MDO} is not.
2003-09-04 04:04:21 +00:00
// Insert a tsheg if necessary. ACIP files aren't
// careful, so "KA\r\n" and "GA\n" appear where "KA
// \r\n" and "GA \n" should appear.
if (('\r' == ch
Fixed ACIP->Unicode spaces/tshegs and newlines, especially with shads. "NGA," becomes "NGA-tsheg-," automatically now.
2003-09-05 05:08:47 +00:00
|| ('\n' == ch && i > 0 && s.charAt(i - 1) != '\r'))
ACIP->TMW and ACIP->Unicode are now smart about when a newline is really a newline and when a space is really a tsheg. The space in {KA ,MDO} is a tsheg, but the space in {GA ,MDO} is not.
2003-09-04 04:04:21 +00:00
&& !al.isEmpty()
{DAN,\nLHAG} used to be treated like {DAN, LHAG} but that got broken. Fixed. Added tests for lexer's handling of ACIP spaces etc.
2003-12-10 06:55:16 +00:00
&& lastNonExceptionalThingWasAdornmentOr(al, TString.TIBETAN_NON_PUNCTUATION)) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", " ", TString.TIBETAN_PUNCTUATION));
Fixed ACIP->Unicode spaces/tshegs and newlines, especially with shads. "NGA," becomes "NGA-tsheg-," automatically now.
2003-09-05 05:08:47 +00:00
}
// "DANG,\nLHAG" is really "DANG, LHAG". But always? Not if you have "MDO,\n\nKA...".
if (('\r' == ch
|| ('\n' == ch && i > 0 && s.charAt(i - 1) != '\r'))
&& !al.isEmpty()
{DAN,\nLHAG} used to be treated like {DAN, LHAG} but that got broken. Fixed. Added tests for lexer's handling of ACIP spaces etc.
2003-12-10 06:55:16 +00:00
&& lastNonExceptionalThingWasAdornmentOr(al, TString.TIBETAN_PUNCTUATION)
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
&& ((TString)al.get(al.size() - 1)).getText().equals(",")
Fixed ACIP->Unicode spaces/tshegs and newlines, especially with shads. "NGA," becomes "NGA-tsheg-," automatically now.
2003-09-05 05:08:47 +00:00
&& s.charAt(i-1) == ','
&& (i + (('\r' == ch) ? 2 : 1) < sl
&& (s.charAt(i+(('\r' == ch) ? 2 : 1)) != ch))) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", " ", TString.TIBETAN_PUNCTUATION));
ACIP->TMW and ACIP->Unicode are now smart about when a newline is really a newline and when a space is really a tsheg. The space in {KA ,MDO} is a tsheg, but the space in {GA ,MDO} is not.
2003-09-04 04:04:21 +00:00
}
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
if ('^' == ch) {
// "^ GONG SA" is the same as "^GONG SA" or
// "^\r\nGONG SA". But "^\n\nGONG SA" is
// different -- that has a true line break in the
// output between ^ and GONG. We give an error if
// ^ isn't followed by an alphabetical character.
boolean bad = false;
if (i + 1 < sl && isAlpha(s.charAt(i+1))) {
// leave i alone
} else if (i + 2 < sl && (' ' == s.charAt(i+1)
|| '\r' == s.charAt(i+1)
|| '\n' == s.charAt(i+1))
&& isAlpha(s.charAt(i+2))) {
++i;
} else if (i + 3 < sl && '\r' == s.charAt(i+1)
&& '\n' == s.charAt(i+2)
&& isAlpha(s.charAt(i+3))) {
i += 2;
} else {
bad = true;
}
if (!bad)
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", "^", TString.TIBETAN_PUNCTUATION));
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
else {
if (queueError(131, "^",
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
}
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
} else {
// Don't add in a "\r\n" or "\n" unless there's a
// blank line.
boolean rn = false;
boolean realNewline = false;
if (('\n' != ch && '\r' != ch)
|| (realNewline
= ((rn = ('\n' == ch && i >= 3 && s.charAt(i-3) == '\r' && s.charAt(i-2) == '\n' && s.charAt(i-1) == '\r'))
|| ('\n' == ch && i >= 1 && s.charAt(i-1) == '\n')))) {
for (int h = 0; h < (realNewline ? 2 : 1); h++) {
if (isTshegBarAdornment(ch) && !legalTshegBarAdornment) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(132, "" + ch,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
} else {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", rn ? s.substring(i - 1, i+1) : s.substring(i, i+1),
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
(legalTshegBarAdornment
? TString.TSHEG_BAR_ADORNMENT
: TString.TIBETAN_PUNCTUATION)));
}
The ACIP {NYA%} is supported. {NYAo} and {NYAx} are confusing to me, because I don't know which glyphs o and x correspond to. For that reason, they cause ERRORs. The proposed THDL Extended Wylie ~X and X is now used for U+0F35 and U+0F37 respectively.
2003-09-07 16:19:50 +00:00
}
}
I don't think warning level "None" was really doing the trick. Fixed that. You can now customize the severities of all warnings, even 504 and 510. When warning level is "None", scanning, i.e. lexical analysis, is faster.
2004-04-25 00:37:57 +00:00
if ('%' == ch
&& ErrorsAndWarnings.isEnabled(504, warningLevel)) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
al.add(new TString("ACIP",
ErrorsAndWarnings.getMessage(504,
shortMessages,
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
"" + ch,
ACIPTraits.instance()),
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
TString.WARNING));
}
ACIP->Tibetan converters now warn every time {%} is encountered that U+0F14 might've been intended. The Unicode for ACIP {o} is U+0F37.
2003-11-09 23:15:58 +00:00
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i+1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
break; // end TIBETAN_PUNCTUATION | TSHEG_BAR_ADORNMENT case
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
default:
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
if (!bracketTypeStack.empty()) {
int stackTop = ((Integer)bracketTypeStack.peek()).intValue();
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
if (TString.CORRECTION_START == stackTop && '?' == ch) {
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
// allow it through...
break;
}
}
Fixed a couple of small bugs. Only 250 errors are reported now; this is important if you try to convert an English document.
2003-08-17 02:12:49 +00:00
if (i+1 == sl && 26 == (int)ch)
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
// Silently allow the last character to be
// control-Z (sometimes printed as ^Z), which just
// marks end of file.
Fixed a couple of small bugs. Only 250 errors are reported now; this is important if you try to convert an English document.
2003-08-17 02:12:49 +00:00
break;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
if (!(isNumeric(ch) || isAlpha(ch))) {
if (startOfString < i) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, i),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
TMW->Wylie conversion now takes advantage of prefix rules, the rules that say "ya can take a ga prefix" etc. The ACIP->Unicode converter now gives warnings (optionally, and by default, inline). This converter now produces output even when lexical errors occur, but the output has errors and warnings inline.
2003-08-23 22:03:37 +00:00
if ((int)ch == 65533) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(114, "unknown character",
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
TMW->Wylie conversion now takes advantage of prefix rules, the rules that say "ya can take a ga prefix" etc. The ACIP->Unicode converter now gives warnings (optionally, and by default, inline). This converter now produces output even when lexical errors occur, but the output has errors and warnings inline.
2003-08-23 22:03:37 +00:00
} else if ('\\' == ch) {
ACIP->TMW and ACIP->Unicode now allow for Unicode escapes like K\u0F84. This means that the lack of support for ACIP's backslash, '\\', is mitigated because you can turn ACIP {K\} into ACIP {K\u0F84}. Support for U+F021-U+F0FF, the PUA that the latest EWTS uses, is not provided.
2003-11-29 22:56:18 +00:00
int x = -1;
if (!ThdlOptions.getBooleanOption("thdl.tib.text.disallow.unicode.character.escapes.in.acip")
&& i + 5 < sl && 'u' == s.charAt(i+1)) {
try {
if (!((x = Integer.parseInt(s.substring(i+2, i+6), 16)) >= 0x0000 && x <= 0xFFFF))
x = -1;
} catch (NumberFormatException e) {
// Though this is unlikely to be
// legal, we allow it through.
// (FIXME: warn.)
}
}
if (x >= 0) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", new String(new char[] { (char)x }),
ACIP->TMW and ACIP->Unicode now allow for Unicode escapes like K\u0F84. This means that the lack of support for ACIP's backslash, '\\', is mitigated because you can turn ACIP {K\} into ACIP {K\u0F84}. Support for U+F021-U+F0FF, the PUA that the latest EWTS uses, is not provided.
2003-11-29 22:56:18 +00:00
TString.UNICODE_CHARACTER));
i += "uXXXX".length();
startOfString = i+1;
break;
} else {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(115, "\\",
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
ACIP->TMW and ACIP->Unicode now allow for Unicode escapes like K\u0F84. This means that the lack of support for ACIP's backslash, '\\', is mitigated because you can turn ACIP {K\} into ACIP {K\u0F84}. Support for U+F021-U+F0FF, the PUA that the latest EWTS uses, is not provided.
2003-11-29 22:56:18 +00:00
}
TMW->Wylie conversion now takes advantage of prefix rules, the rules that say "ya can take a ga prefix" etc. The ACIP->Unicode converter now gives warnings (optionally, and by default, inline). This converter now produces output even when lexical errors occur, but the output has errors and warnings inline.
2003-08-23 22:03:37 +00:00
} else {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(116, "" + ch,
shortMessages, i, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
startOfString = i+1;
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType = TString.ERROR;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
} else {
// Continue through the loop.
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
if (TString.ERROR == currentType)
currentType = TString.TIBETAN_NON_PUNCTUATION;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
break; // end default case
}
}
if (startOfString < sl) {
TString now has tracks what Roman transliteration system it is using. Next up is to make ACIPConverter handle EWTS or ACIP TStrings.
2004-01-17 19:28:54 +00:00
al.add(new TString("ACIP", s.substring(startOfString, sl),
Andres found that "THAG PA" caused a NullPointerException. That's fixed. Renamed ACIPString to TString -- we'll use this for EWTS and ACIP both. TMW->ACIP for TMW9.61 should work now.
2003-10-04 01:22:59 +00:00
currentType));
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
if (waitingForMatchingIllegalClose) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(117, "-*-END OF FILE-*-",
shortMessages, -1, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
}
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
if (!bracketTypeStack.empty()) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(((TString.COMMENT == currentType) ? 118 : 119),
"-*-END OF FILE-*-",
shortMessages, -1, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
if (startSlashIndex >= 0) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(120, "-*-END OF FILE-*-",
shortMessages, -1, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
}
if (startParenIndex >= 0) {
Robert Chilton's experience inspired me to make the handling of errors and warnings in ACIP->Tibetan conversion much more configurable. You can now choose from short or long error messages, for one thing. You can change the severity of almost all warnings. Each error and warning has an error code. Errors and warnings are better tested. The converter GUI has a new checkbox for short messages; the converter CLI has a new mandatory option for short messages. I also fixed a bug whereby certain errors were not being appended to the 'errors' StringBuffer.
2004-04-24 17:49:16 +00:00
if (queueError(121, "-*-END OF FILE-*-",
shortMessages, -1, numNewlines, maxErrors, al, errors, numErrorsArray))
return null;
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
return al;
}
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
/** See implementation. */
private static boolean isNumeric(char ch) {
return ch >= '0' && ch <= '9';
}
Improved the ACIP scanner (the part of the converter that says, "This is a correction, that's a comment, this is Tibetan, that's Latin (English), that's Tibetan inter-tsheg-bar punctuation, etc.) It now accepts more real-world ACIP files, i.e. it handles illegal constructs. The error checking is more user-friendly. There are now tests. Added some tsheg bars that Peter E. Hauer of Linguasoft sent me to the tests. Many thanks, Peter. I still need to implement rules that say, "This is not Tibetan, it must be Sanskrit, because that letter doesn't take a MA prefix."
2003-08-17 01:45:55 +00:00
/** See implementation. */
private static boolean isWhitespace(char ch) {
return ch == ' ' || ch == '\t' || ch == '\r' || ch == '\n';
}
The ACIP {NYA%} is supported. {NYAo} and {NYAx} are confusing to me, because I don't know which glyphs o and x correspond to. For that reason, they cause ERRORs. The proposed THDL Extended Wylie ~X and X is now used for U+0F35 and U+0F37 respectively.
2003-09-07 16:19:50 +00:00
/** See implementation. */
private static boolean isTshegBarAdornment(char ch) {
return (ch == '%' || ch == 'o' || ch == 'x');
ACIP % {MTHAR%} and o {Ko} and ^ {^GONG SA} are now supported. A % always causes a warning.
2003-11-11 03:43:11 +00:00
// ^ is a pre-adornment; these are post-adornments.
The ACIP {NYA%} is supported. {NYAo} and {NYAx} are confusing to me, because I don't know which glyphs o and x correspond to. For that reason, they cause ERRORs. The proposed THDL Extended Wylie ~X and X is now used for U+0F35 and U+0F37 respectively.
2003-09-07 16:19:50 +00:00
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
/** See implementation. */
private static boolean isAlpha(char ch) {
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
return ch == '\'' // 23rd consonant
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
// combining punctuation, vowels:
I'm now stricter about accepting alphabetic characters. F, Q, X, a, b, c, d, e, ... do not belong in ACIP, so the scanner rejects them. This should make it even easier to distinguish automatically between Tibetan and English texts.
2003-08-17 02:38:58 +00:00
|| ch == 'm'
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
|| ch == ':'
ACIP->Unicode was broken for KshR, ndRY, ndY, YY, and RY -- those stacks that use full-form subjoined RA and YA consonants. ACIP {RVA} was converting to the wrong things. The TMW for {RVA} was converting to the wrong ACIP. Checked all the 'DLC' tags in the ttt (ACIP->Tibetan) package.
2003-11-09 01:07:45 +00:00
// FIXME: we must treat this guy like a vowel, a special vowel that numerals can take on. Until then, warn. See bug 838588 || ch == '\\'
I've improved the ACIP tsheg bar scanner to handle a lot of illegal constructions that occur in practice.
2003-08-16 16:13:53 +00:00
|| ch == '-'
|| ch == '+'
I'm now stricter about accepting alphabetic characters. F, Q, X, a, b, c, d, e, ... do not belong in ACIP, so the scanner rejects them. This should make it even easier to distinguish automatically between Tibetan and English texts.
2003-08-17 02:38:58 +00:00
|| ((ch >= 'A' && ch <= 'Z') && ch != 'X' && ch != 'Q' && ch != 'F')
|| ch == 'i'
|| ch == 't'
|| ch == 'h'
|| ch == 'd'
|| ch == 'n'
|| ch == 's'
|| ch == 'h';
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
Refactored so that there can be an EWTS scanner and an ACIP scanner.
2005-02-21 05:37:01 +00:00
/** non-public because this is a singleton */
protected ACIPTshegBarScanner() { }
private static ACIPTshegBarScanner singleton = null;
/** Returns the sole instance of this class. */
public synchronized static ACIPTshegBarScanner instance() {
if (null == singleton) {
singleton = new ACIPTshegBarScanner();
}
return singleton;
}
I now have a function that takes as input a String of ACIP and breaks up that String into tsheg bars, punctuation, etc., while finding errors. I've tested it some, but I'm not yet committing the tests. Next step: a converter that takes an ACIP file as input and outputs TMW+Latin.
2003-08-14 05:10:47 +00:00
}
Reference in a new issue Copy permalink