Re: [PATCH] etc: import changes to texi2pod.pl from gcc
Andrew Burgess <[email protected]>
| Newsgroups | gmane.comp.gdb.patches,gmane.comp.gnu.binutils |
|---|---|
| Message-ID | <[email protected]> |
Andrew Burgess <[email protected]> writes: > > commit 1dd632a1865281fbc054e6af55dc1d4a371c4b61 > Author: Andrew Burgess <[email protected]> > Date: Fri Jul 24 17:49:19 2026 +0100 > > etc: import changes to texi2pod.pl from gcc > > I was looking at making some cleanup in the GDB documentation and I > ran into a limitation of the texi2pod.pl script, it is not able to > correctly format: '@option{--text}' or '@table @option'. > > Patching our copy of texi2pod.pl is easy enough, but it is probably > better to just sync with gcc's version of the script, which already > has some @option support in place. > > I took the version from gcc commit 3f337c3d7e2 -- that's is reasonably > close to gcc's current HEAD. > > I then checked that the man pages for gcc, gas, binutils, ld, gprof, > and gprofng all regenerated without any changes. Unfortunately that > wasn't the case, so I had to add back some local changes: > > 1. Support for the --no-split option, see binutils-gdb commit > 96a7037cd8573cf065aa6b12baca68696f96d9ca. > > 2. Add support for TARGET as a man page section, see binutils-gdb > commit 6387924ae329fe2a89abb48b9868a340e23a9630. > > 3. Revert gcc commit 386c402747d2bc6bad3e1bc2f9383d4d669576e0 to > allow our lists to be correctly formatted again. > > I suspect item (3) could be better fixed by understanding the original > gcc issue and crafting a smarter solution that both solves gcc's > original issue while also allowing our lists to be formatted > correctly. However, given we've managed without (3) for years, and I > need to make some local fixes anyway, I propose just removing that > small part. At least for now. > > One final note, this import from gcc changes the licence on the > texi2pod.pl script from GPLv2 to GPLv3. As I understand it, GPL > allows for this roll forward, and most of binutils-gdb is under GPLv3, > so I don't see this being a problem. > I've gone ahead and checked this in. If anyone sees any problems then let me know and I'll get it sorted. Thanks, Andrew