<farkuhar>
my earlier attempt at refactoring your pkgmk patch was missing a "basename" command. Here's the fixed version: http://sprunge.us/KfTW9c?diff
<farkuhar>
I started hacking at docbook-x{m,s}l because I wanted to figure out why the manpages for xmlto and xdg-mime were getting generated incorrectly (lots of unnecessary whitespace, indentation; and visible troff formatting codes). So I borrowed some ideas from Arch and Void to revise your port. After applying the above patches and rebuilding, I finally got clean manpages for xmlto and xdg-mime.
<beerman>
nice, i have noticed the whitespaces too but never cared enough to look at it
<jaeger>
Same. I admit I have zero attachment to docbook, only took it over to keep it alive
<jaeger>
Thanks, I'll try to check them out after work.
<jaeger>
Anyone else tested nounpack yet?
<beerman>
never heard of it actually
<jaeger>
The intent is to remove the need for overriding unpack_source()
<beerman>
oh, lol. i didn't test it yet
<farkuhar>
beerman, try http://sprunge.us/pe9JAx?diff if you want to rebuild docbook-xml without applying any new patches to pkgmk. In this case the renames array is just as tidy as jaeger's nounpack array.
<beerman>
i'll try it out later
<beerman>
i think i saw it just kinda forgot about it too, sorry
<farkuhar>
btw, line 282 in the latest diff is a small correction to my previous loop-breaking condition (which was selecting the wrong combinations in the doubly-nested loop).
<farkuhar>
better idea, just generate the catalog entries by recursing through the filesystem itself (and while we're at it, fold the unpacking command into the xmlcatalog loop): http://sprunge.us/6EVP3z?diff
pitillo has quit [Ping timeout: 260 seconds]
pitillo has joined #crux-devel
<jaeger>
Do you want to take over docbook? I'd gladly give it up, heh
<beerman>
i hope you didn't mean me 😄
<farkuhar>
I'd be happy to take over docbook, considering how infrequent the release cycle seems to be. But I should probably start by submitting a formal application to join contrib.