#linuxcnc-devel Logs

Nov 09 2021

01:03 AM CaptHindsight[m]: 🎆 thanks for the big cleanup
01:30 AM pere: CaptHindsight[m]: do 'because volunteers' mean there is no plan? What is the preferred format to maintain manual pages here?
01:31 AM CaptHindsight[m]: pere: ask again in about 8-9 hours
01:31 AM CaptHindsight[m]: likely some traditional guideline that is not policed
01:32 AM pere: personally I recommend using as few formats as possible, and given that the rest of the documentation is in asciidoc, I would suggest to migrate manual pages to asciidoc too.
01:40 AM CaptHindsight[m]: asciidoc https://github.com/LinuxCNC/linuxcnc/commit/f6ddca32e5f32a31d56602393e3ef7b1e2081f85
01:43 AM pere: yeah, that was the code fragment I found which let me know both roff and adoc files were used as manual page sources.
10:12 AM seb_kuzminsky: rene-dev: looks like commit 9c05629c9c303 broke the build, can you look at it?
10:14 AM seb_kuzminsky: pere: it's true there's not a formal plan for consolidating our manpages into one source format. We havea bunch of old manpages that were written in roff before asciidoc was a thing, and some that we wrote in asciidoc once that became an option. I'm not opposed to moving all manpages to asciidoc.
10:15 AM seb_kuzminsky: oh, i should note we also have a tool called halcompile that helps simplify(?) the creation of new hal components, including their manpages, from source files like this: https://github.com/LinuxCNC/linuxcnc/blob/master/src/hal/components/abs.comp
10:19 AM seb_kuzminsky: halcompile is described here: http://linuxcnc.org/docs/2.8/html/hal/comp.html
10:21 AM rene-dev: Ok, let me check
10:23 AM pere: seb_kuzminsky: but man->asciidoc is trivially automatable?
10:25 AM pere: and asciidoc->man too, so why mix and match when it is trivial to convert them all over in 20 seconds?
10:47 AM seb_kuzminsky: will having all the manpages in asciidoc source format help with po4a?
11:42 AM pere: seb_kuzminsky: perhaps a bit. either that or the man pages need to limit themselves to the features po4a understand, ie no .DE
11:45 AM pere: seb_kuzminsky: see d257657167bd2b0f8533077f9c8c8382d86efe8b for the ones I had to avoid because po4a do not understand them.
12:03 PM pere: seb_kuzminsky: the reason for my proposal to migrate all roff manual pages to asciidoc is that I believe it is an advantage to the linuxcnc community to only have to work with one documentation format. po4a can handle this just fine, modulo the unsupported nroff features.
12:18 PM seb_kuzminsky: yeah, sounds good too me. I sure prefer editing asciidoc over editing roff :-)
01:29 PM jepler[m]: halcompile will need a new asciidoc mode as well, currently it generates manpages from the docstrings..
01:30 PM jepler[m]: no need for it to support both, one output mode is plenty
01:31 PM jepler[m]: I see seb had already mentioned halcompile
01:34 PM pere: I was not really planning to rewrite the entire documentation handling, just make life easier for translators...
01:37 PM pere: can someone explain to me why <URL: https://github.com/LinuxCNC/linuxcnc/pull/1356 > seem to introduce new dependencies, and fail to build because they are missing. why did this not fail before?
02:05 PM seb_kuzminsky: pere: it did fail before - someone pushed a change just after smoe uploaded the package, and just before you forked the translate-po4a branch, that broke the build in master
02:05 PM seb_kuzminsky: i think rene is working on fixing it
02:06 PM seb_kuzminsky: our github repo allows members of the project to push to master, without going through the pr & code-review process, it make things quicker but sometimes bugs slip through :-/
02:10 PM pere: ah
02:11 PM pere: very good. it confused the heck out of me. :)
02:13 PM seb_kuzminsky: possibly consider rebasing your branch back in time, i think 2b2200a7596987e724365e9e18ec4abeb954a3d4 is good
02:22 PM pere: I'll wait for master to be fixed, I think.
03:50 PM rene-dev: seb_kuzminsky looks more like https://github.com/LinuxCNC/linuxcnc/commit/1fe3119b1f07c7a66ffdb033a46d3888842ceb30 broke it
03:50 PM rene-dev: the error refers to this path being invalid
04:27 PM pere: smoe: hi. :)
04:27 PM smoe: hi#
06:00 PM seb_kuzminsky: rene-dev: you're right, i'm wrong, sorry for pointing my finger at your commit incorrectly
06:01 PM seb_kuzminsky: phill fixed it, master builds again now
11:15 PM pere: seb_kuzminsky: which commit is this? I see nothing new in master.
11:44 PM seb_kuzminsky: pere: 1fe3119b1 broke the build and 2c2902006 fixed it
11:46 PM seb_kuzminsky: hmm, though i see the buildbot still didn't like 2c2902006bb...
11:47 PM pere: well, I do not see it either in master.
11:47 PM seb_kuzminsky: what don't you see in master?
11:47 PM pere: at least not as a fix.
11:50 PM pere: my translate-po4a was already based on 2c2902006bba36972a976c0f9c8032f61360267d.
11:50 PM seb_kuzminsky: are you getting errors about checklink not found?
11:52 PM pere: yeah, see <URL: https://github.com/LinuxCNC/linuxcnc/pull/1356 >
11:56 PM seb_kuzminsky: i see the error you're getting there, it's the same as the error the buildbot is getting
11:57 PM seb_kuzminsky: i have a fix here, i'll do one more test and if that's good then i'll push it
11:57 PM seb_kuzminsky: so there were two bugs keeping master from building, that's funny
11:58 PM pere: right. any idea how it crept in there?
11:58 PM seb_kuzminsky: i guess two people pushed and didn't watch the CI error reports
11:59 PM pere: I do not understand how the changes could cause missing dependencies.
11:59 PM seb_kuzminsky: that's a misleading error message :-(