libsoldout

manual pages cleanup Leaf check-in: 9e572998d0 user: github tags: trunk

renderers: fix a typo check-in: a520e7aa21 user: github tags: trunk

markdown: fix typos check-in: 8fcd613f7e user: github tags: trunk

fix description of MKDA_NORMAL check-in: 559d152cdc user: natgh tags: trunk

markdown: fix a typo in a comment check-in: 4d25fef47c user: natgh tags: trunk

explit => explicit check-in: d0429906cf user: frgm tags: trunk

markdown: fix potential leak of memory pointed to by 'text' check-in: c70ae18d2a user: natgh tags: trunk

see also: https://github.com/faelys/libsoldout/issues/32 check-in: c43896a0c4 user: frgm tags: trunk

soldout_*.3: put in the section NAME all functions names check-in: a62bf1ebc1 user: natgh tags: trunk

makefiles: remove unneeded '.m' rule check-in: eb39494f8a user: natgh tags: trunk

remove 'BUFFER_STDARG' macro check-in: a5ce1a3abd user: natgh tags: trunk

for some reasons '\t' does not work as expected; solution: use literal tab check-in: 413236099e user: frgm tags: trunk

as there is no more need for it: https://github.com/faelys/libsoldout/issues/26 check-in: 95a8c77643 user: frgm tags: trunk

see: https://github.com/faelys/libsoldout/issues/20 check-in: a222d618aa user: frgm tags: trunk

fix two warnings check-in: 13030d946c user: natgh tags: trunk

no previous extern declarations for non-static variables 'to_latex' and 'to_man' [-Wmissing-variable-declarations] check-in: 06bca437f0 user: frgm tags: trunk

no previous prototype for functions 'latex_text_escape' and 'usage' [-Wmissing-prototypes] check-in: 9158c8be45 user: frgm tags: trunk

add amalgamation build check-in: 7b4fb0eefe user: natgh tags: trunk

as their `sed` does not have an argumentless variant of `-i` option check-in: 531a42cf16 user: frgm tags: trunk

mkd2man.1: s/should/must/ check-in: f2017bc29a user: natgh tags: trunk

*.3: remove a LIBRARY section (as libsoldout is a portable library) check-in: aee5703bc0 user: natgh tags: trunk

see: http://manpages.bsd.lv/part1-3.html and http://thread.gmane.org/gmane.comp.tools.mdocml.user/793/focus=794 check-in: 6b7819e896 user: frgm tags: trunk

adjust makefiles check-in: 4738b3b542 user: natgh tags: trunk

* add configurable CC, CFLAGS and LDFLAGS * remove useless flags for LDFLAGS * add a static library build check-in: 9fb854c77b user: frgm tags: trunk

Previously, `main` function in mkd2man returned a literal integer or one of POSIX `EXIT_*` constants. Since the exit code doesn't seem to encoding anything other than success or failure, `EXIT_SUCCESS` and `EXIT_FAILURE` where chosen. check-in: 2d59c88528 user: nat tags: trunk

convert manual pages into mdoc format check-in: 0e0a114737 user: natgh tags: trunk

imply => implies check-in: f3cca3229f user: frgm tags: trunk

first without argument, then other; + fix a typo check-in: 1ff5181da5 user: frgm tags: trunk

The point of this check was to do nothing when the first byte after used data exists and is already 0. So it saves a memory write (and possible cache invalidation) at the cost of a comparison. The trade-off is far from obvious.

On top of that, it is counted as a conditional that depends on uninitialized memory by valgrind (as reported in #15 ), which is technically true, even though that particular instance is harmless.

So I decided to remove this check and unconditionally write 0 after the end of the current data (after making for it if needed). check-in: e571a14858 user: nat tags: trunk

do not define BUFFER_STATS and fix a warning check-in: 0d175d62ef user: natgh tags: trunk

no previous extern declaration for non-static variables 'buffer_stat_nb' and 'buffer_stat_alloc_bytes' check-in: 458d410f51 user: frgm tags: trunk

renderers: add _const_ attribute to argument 'src' in lus_*_escape() functions check-in: 94fbccaed2 user: natgh tags: trunk

The second `if` should be inside the first, as denoted by the indentation. Fortunately, this doesn't result in a visible bug, because `j` is initialized to zero and not touched otherwise. However, it does trigger a warning in gcc-6, which causes a build failure because of `-Werror`, reported in https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=811587

Thanks to Massimo Manghi for letting me know. check-in: d971cdd23d user: nat tags: trunk

since they do not modify 'src' check-in: 05afc27f69 user: frgm tags: trunk

mkd2man: wrap long lines check-in: a4ffc9524f user: natgh tags: trunk

mkd2man: fix a warning check-in: 178e13ac54 user: natgh tags: trunk

comparison between signed and unsigned integer expressions [-Wsign-compare] check-in: c66a6b5693 user: frgm tags: trunk

buffer: fix a warning check-in: 05564c9b36 user: natgh tags: trunk

ISO C does not allow extra ';' outside of a function [-Wpedantic] check-in: cb8ef52efc user: frgm tags: trunk

mkd2man: fix 'longopts' check-in: ab254af2cd user: natgh tags: trunk

README: Fix mistake check-in: 2a9a68f6f3 user: natgh tags: trunk

I think you confused 2 words there. check-in: 2652c0beaf user: nbraud tags: trunk

typos check-in: 1adac1ae44 user: natgh tags: trunk

fix trailing whitespace check-in: 5d4e1bc1e0 user: natgh tags: trunk

typos check-in: 2d9f1e2a40 user: natgh tags: trunk

Thanks to @frgm on github for reporting it. I still don't understand why it wasn't caught in my FreeBSD hosts. check-in: 214f69dc62 user: nat tags: trunk

Add .gitignore file check-in: 21119eef93 user: natgh tags: trunk

Special symbols to not need escape in verbatim blocks, since there only the special sequence "\end{verbatim}" is recognized. However there is currently no espcial handling of this sequence, so putting one in the source code triggers a LaTeX-injection vulnerability. check-in: 2d1ea0da9c user: nat tags: trunk

The current conversion table is a minimalistic placeholder for the Markdown test suite. check-in: 056b90968c user: nat tags: trunk

Written by fernozzle as part of Google Code-In 2012: http://www.google-melange.com/gci/task/view/google/gci2012/7981207 check-in: b56833da08 user: nat tags: trunk

Thanks a lot to Jeremy Cowgar for pointing out and providing a patch. check-in: e3f435e9b2 user: nat tags: trunk

The check was supposed to prevent releasing volatile buffers, but it also prevent releasing buffers that have never been used. Moreover, volatile buffers must have unit == 0 anyway. check-in: 61475d352e user: nat tags: trunk

See README for details and proof. check-in: ef49755411 user: nat tags: trunk

Lace being what you see when you upskirt. check-in: fa08928de3 user: nat tags: trunk

It should be more efficient to insert links at the end of the array and then qsort(3) the final array rather than performing sorted insertion. check-in: 6e30d08540 user: nat tags: 1.0-rc1, trunk

Big thanks to bapt for this renderer. check-in: 50c1c1ea23 user: nat tags: trunk

The "entity" renderer callback is no longer called when a lone '&' is encountered, only when a valid entity ("&#?[A-Za-z0-9]+;") is recognized. Otherwise '&' is considered as an inactive character and handed over to "normal_text" callback along with other inactive charaters.

In the example renderers rndr_entity has therefore been replaced by NULL, because the only treatment of entities is to copy them verbatim into the output (without the escaping rndr_normal_text does). check-in: e5946f097a user: nat tags: trunk

- Entities are now handled by a dedicated callback - A new callback for normal text has been added check-in: 89c8399043 user: nat tags: trunk

The text argument to blockcode() and codespan() is no longer entity-escaped by the parser. Now the renderer has to do whatever it takes to have the input correctly escaped into the output. check-in: ee37e3fa83 user: nat tags: trunk

The following arguments to callbacks are no longer escaped by the parser: - link in autolink() - link and title in link() - link, title and alt in image()

To help custom renderers, lus_attr_escape() is exported, which escapes '<', '>', '&' and '"' from the input, replacing them by entities. check-in: 14e0e4e038 user: nat tags: trunk

In addition to the extensions I took from discount (image size, pseudo-protocols and class blocks), here is a list of markdown extensions for my own needs: - id attribute for headers, with the syntax: id#Header text - class attribute for paragraph, between parenthesis at the very beginning of the paragraph - and spans, using respectively "++" and "--" as delimiters - plain for text between '|' check-in: 6f4a60475a user: nat tags: trunk

Instead of having everything in markdown.*, now the example renderers are in renderers.*, and markdown.* only contains parsing code. check-in: 7f5412c1c8 user: nat tags: trunk