summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAge
...
* Silly of me to overlook it: another obvious way you might like toSimon Tatham2012-07-19
| | | | | | | | | specify characters to 'confuse' is to just put them on the command line in the system multibyte encoding! In a UTF-8 terminal environment this may very well be the easiest thing. [originally from svn r9584] [this svn revision also touched charset,filter,timber]
* A slightly silly new utility: 'confuse'. You provide it with someSimon Tatham2012-07-18
| | | | | | | | | | | | | Unicode values (typically two of them), and it finds cases in which the provided characters are all encoded as the same thing in different charsets and prints those charsets. So if you encounter, for example, some piece of text which has U+0153 LATIN SMALL LIGATURE OE where you might have expected U+00A3 POUND SIGN, simply run 'confuse 153 a3' and it'll tell you which character sets the sender and receiver of the text might have got confused between. [originally from svn r9581] [this svn revision also touched charset,filter,timber]
* Mechanism for iterating over all supported charsets.Simon Tatham2012-07-18
| | | | | [originally from svn r9580] [this svn revision also touched charset,filter,timber]
* Patch from Maciej Adamczyk to set standard output to binary mode inSimon Tatham2012-06-03
| | | | | | | the standalone deflate utility modes, so you can redirect output to a file and end up with correct data in that file. [originally from svn r9552]
* Make the Deflate decoder correctly handle the special case of aSimon Tatham2012-06-01
| | | | | | | | | dynamic block in which the backward-distances Huffman table contains no entries at all. (In this situation it's impossible to use the distance table, and therefore the block must contain nothing but Huffman-coded literals.) [originally from svn r9550]
* OS X Lion's compiler spotted two missing casts, where we use %hu andSimon Tatham2012-05-06
| | | | | | pass something that isn't necessarily an unsigned short. [originally from svn r9493]
* Now that you can specify negative indents in Halibut text mode, it'sSimon Tatham2012-05-04
| | | | | | | | | | possible to specify an indent which is _overall_ negative, i.e. intended to be off the left-hand side of the text file. There's no nice way to handle this, but printing O(2^32) spaces in a while (n--) loop is particularly bad. Fix the loop to read while (n-- > 0), so that at least it doesn't get _that_ badly confused. [originally from svn r9491]
* Fix an integer-type mismatch between %04x in a printf format stringSimon Tatham2012-05-03
| | | | | | | | and a long int. Spotted by Ubuntu 12.04's gcc, and probably would have caused trouble on 64-bit machines. [originally from svn r9489] [this svn revision also touched charset,filter,timber]
* Apply r6972 from my other copy of deflate.c: don't lose the BFINAL bitSimon Tatham2012-05-03
| | | | | | | on an uncompressed block. [originally from svn r9482] [r6972 == 37af901cfe63ed062306e3835751e9b8e089e252 in misc repository]
* Remove a bunch of unused variables spotted by Ubuntu 12.04's gcc.Simon Tatham2012-05-03
| | | | [originally from svn r9478]
* Ahem. Don't invert the length _twice_.Simon Tatham2012-05-03
| | | | [originally from svn r9476]
* Add an error check for correct formatting in Deflate uncompressedSimon Tatham2012-05-03
| | | | | | | block headers. (Ubuntu 12.04's gcc spotted that nlen was unused, which it shouldn't have been.) [originally from svn r9475]
* Bug in utoi(), which made it ignore a leading minus sign whenSimon Tatham2012-05-03
| | | | | | | | | | | | converting a number! Oops. Now you can do things like \cfg{text-list-indent}{-2} \cfg{text-listitem-indent}{2} to get your bullets placed in the left margin rather than indenting the paragraphs of your list. [originally from svn r9474]
* Allow macro definitions to be terminated by tok_eof as well asSimon Tatham2012-02-24
| | | | | | | tok_eop, to prevent Halibut tight-looping when a definition is the last thing in a file. [originally from svn r9419]
* Correct a comment.Simon Tatham2011-11-09
| | | | | | | | | | | | | | | I had wrongly believed my TYPECHECK macro double-evaluated one of its arguments and hence would cause side effects to happen twice. But in fact I've just realised that although it double-_expands_ the argument, it doesn't double-_evaluate_ it: the two expansions occur in mutually exclusive branches of a ?:, and hence cannot both be executed. So I've removed the comment that says my macro is rubbish. My macro is in fact great :-) [originally from svn r9328] [this svn revision also touched charset,filter,timber]
* Merge PuTTY r9326, adding CP852 support.Simon Tatham2011-10-14
| | | | | | [originally from svn r9327] [r9326 == c72d4b413f024e3c50645caceaddbb65401fb06a in putty repository] [this svn revision also touched charset,filter,timber]
* First attempt at an Emacs mode that syntax-highlights Halibut markup.Simon Tatham2011-08-09
| | | | [originally from svn r9259]
* A user recently reported that Halibut's font-file handling doesn'tSimon Tatham2010-02-27
| | | | | | | | | work on Windows, because input files are never opened in binary mode. Introduce a small piece of compensatory mechanism, which opens files initially in binary mode and then reopens them as text if they're determined not to match any binary file type. I hope. [originally from svn r8889]
* Clarify the section about backend-generic config options to make itSimon Tatham2010-01-14
| | | | | | look less as if they're the _only_ config options. [originally from svn r8842]
* Escape &<> when they appear in href text.Simon Tatham2009-11-10
| | | | [originally from svn r8743]
* Allow a hastily selected subset of the output formats to also acceptSimon Tatham2009-10-24
| | | | | | | | | | | | | | | | | | | | "-" as a special file name meaning standard output. I've restricted it to just those output formats which can predictably output only one file, just for the sake of not having to faff too much with the others. Probably what I should have done for all of this would have been to write a set of wrappers around fopen, fclose and everything in between, and use them everywhere. Those wrappers would uniformly detect "-" and convert it into stdin or stdout as appropriate, would avoid fclosing those files for real when told to close them, and would also be able to handle reading a little bit of data from the start of a file and then pushing it all back even if the file were not seekable (which would allow input.c to lose the ugly special case whereby it can't currently read font files from standard input). [originally from svn r8729]
* Enable Halibut to read a .but file from standard input, by supplyingSimon Tatham2009-10-24
| | | | | | the special filename '-'. [originally from svn r8728]
* Revise r8309 to use \- rather than - when trying to emit a literal U+002D;Jacob Nevins2009-09-06
| | | | | | | | apparently this is preferred (it's what pod2man uses); cf Debian #496063, #159872. [originally from svn r8641] [r8309 == 73e8c7d1b4ac77ec1b5acc700cb3af277a150bcf]
* More portable attempt to get a literal U+0027 than r8321, pinched from pod2manJacob Nevins2009-09-06
| | | | | | | (via Colin in Debian bug#496063). [originally from svn r8640] [r8321 == 5e7a3a5b1bbaccb7ce7d61bb9bc4654924135dfb]
* I've just seen the MIME charset name 'x-sjis' in the wild. Add it toSimon Tatham2009-04-17
| | | | | | | the list. [originally from svn r8498] [this svn revision also touched charset,filter,timber]
* Fix missing docs for \cfg{paper-contents-indent-step} andSimon Tatham2009-01-20
| | | | | | | \cfg{paper-contents-margin}. Also correct the spelling of the former. [originally from svn r8424]
* Fix unintentional bash-specificity in Makefile. (Patch from ColinSimon Tatham2009-01-20
| | | | | | Watson.) [originally from svn r8422]
* Use of \lcont in the document preamble caused an assertion failure.Simon Tatham2009-01-20
| | | | | | Fix it. [originally from svn r8421]
* ctype functions require their argument to be EOF or representable as anBen Harris2009-01-11
| | | | | | | | | unsigned char. On platforms were char is signed, passing plain char won't cut it. Make sure we case chars to unsigned char before passing them to tolower(). [originally from svn r8404] [this svn revision also touched charset,filter,timber]
* Test for r8309/r8321.Jacob Nevins2008-11-23
| | | | | | [originally from svn r8322] [r8309 == 73e8c7d1b4ac77ec1b5acc700cb3af277a150bcf] [r8321 == 5e7a3a5b1bbaccb7ce7d61bb9bc4654924135dfb]
* Extend r8309 to try to ensure that single-quote/apostrophe characters in codeJacob Nevins2008-11-23
| | | | | | | | | | contexts get output usefully in the face of UTF-8 *roff implementations. Works on Debian-derived distributions (and hence along with r8309 should fix #496063), but I'm slightly worried about portability as I've used a named character "\(aq" which doesn't appear in the "classic" reference CS TR #54. [originally from svn r8321] [r8309 == 73e8c7d1b4ac77ec1b5acc700cb3af277a150bcf]
* Cut-and-paste error: reference to "man page" in Windows Help section.Jacob Nevins2008-11-23
| | | | [originally from svn r8320]
* Let's not use parentheses in test.but in a way that tempts HalibutSimon Tatham2008-11-21
| | | | | | | to try to put them into an Info node name, so as to avoid the error message when it fails. [originally from svn r8313]
* Correct mishandling of paragraphs beginning with "\#{".Simon Tatham2008-11-21
| | | | | | (Embarrassingly, this was showing up in inputs/test.but itself.) [originally from svn r8312]
* Oh, all right. Put in the implicit zero elements at the ends of theSimon Tatham2008-11-21
| | | | | | | initialisers, so that gcc stops whining. [originally from svn r8311] [this svn revision also touched charset,filter,timber]
* Clean up some compiler warnings.Simon Tatham2008-11-21
| | | | [originally from svn r8310]
* Don't translate hyphens into \(hy inside \c{...} or code paragraphs,Simon Tatham2008-11-21
| | | | | | | | | | because otherwise UTF-8 versions of man(1) tend to turn them into Unicode hyphen characters, with the result that if you cut and paste sample command lines out of man pages in the obvious way then the things that look like hyphens are not in fact the character that the program in question would recognise as an option introducer. [originally from svn r8309]
* I've just had some spam in Windows-874, a Thai SBCS. Add libcharsetSimon Tatham2008-08-21
| | | | | | | support for it. [originally from svn r8151] [this svn revision also touched charset,filter,timber]
* Just in case sbcsgen.pl is fed an sbcs.dat with the wrong lineSimon Tatham2008-07-09
| | | | | | | endings, remove \r from input lines. [originally from svn r8113] [this svn revision also touched charset,filter,timber]
* Turn the numeric parameter to err_sfntbadglyph from wchar_t toSimon Tatham2008-07-09
| | | | | | | | | | | | unsigned int, after a correspondent mentioned that DJGPP produces a very scary warning about integer promotion making use of the former in va_arg non-portable. Whether or not that's standards-justifiable, it's certainly the case that we were casting _to_ wchar_t from unsigned int in all cases of this error actually being used, so the simplest thing is just to transfer the number through va_arg as unsigned. [originally from svn r8112]
* Fix cut'n'paste error which had the effect of making the default forJacob Nevins2008-03-19
| | | | | | \cfg{html-section-shownumber}{n} undefined. [originally from svn r7929]
* New option *-section-shownumber, alongside *-section-numeric and inSimon Tatham2008-02-25
| | | | | | | | | both the back ends which currently support that, to leave out chapter and section numbers totally in section headings. Can be useful for publishing man pages (which don't normally want section numbers) on the web. [originally from svn r7892]
* Cope with TrueType fonts with slightly broken cmaps, just ignoring code pointsBen Harris2007-12-02
| | | | | | that can't be resolved (apart from warning about it). [originally from svn r7800]
* Typo.Simon Tatham2007-12-02
| | | | [originally from svn r7799]
* Recognise a couple more types of Unicode 'cmap'. Also fail if a format 4Ben Harris2007-12-02
| | | | | | 'cmap' refers to a character off the end of the glyphIndexArray. [originally from svn r7798]
* Add the ability to pass a NULL output buffer and/or an unlimitedSimon Tatham2007-08-05
| | | | | | | | | output length to charset_{to,from}_unicode, permitting convenient dry-running of conversions to determine the required output length and/or test for the presence of difficult characters. [originally from svn r7677] [this svn revision also touched charset,filter,timber]
* The length parameters in mbstowcs and wcstombs are limits on theSimon Tatham2007-06-21
| | | | | | | | | | | | _output_ length, not the input length. Adjust accordingly. This has the side effect of working around what I believe to be a bug in DJGPP's mbstowcs(), namely that if you give it a string which exactly fits in the output length _without_ the trailing NUL, then it will write n-1 characters of the string plus a NUL where I believe it should write n characters of the string and no NUL. [originally from svn r7622]
* Add rule to compile emacsenc.c. Noticed by David Leonard.Ben Harris2007-04-30
| | | | | [originally from svn r7495] [this svn revision also touched charset,filter,timber]
* Add a mechanism for translating to and from the coding system symbolsBen Harris2007-04-09
| | | | | | | | used by GNU Emacs. This is likely to be useful for generating or interpreting "coding:" entries in file local variables. [originally from svn r7455] [this svn revision also touched charset,filter,timber]
* List some documentation sources, and another todo item.Ben Harris2007-04-08
| | | | [originally from svn r7454]
generated by cgit v1.1 at 2026-04-25 06:55:18 +0000