The plan9port man program doesn't use bzipped manpages, but portages automatically compresses anything in the man directory of a package. I've made a small hack that fixes this, ebuild attached, the relevant lines are the ones in scr_compile: mv man mtamnp and then in pkg_postinst: mv mtamnp man. I know it's ugly, but this is the only way i could find. Also, I am using a newer plan9port package, so the patch won't apply, though it works fine without the patch. I can imagine some weird situation where the patch would be needed, but shouldn't plan9port be installed to /usr/local anyways? Reproducible: Always Steps to Reproduce: 1. install plan9port 2. do an env-update and source /etc/profile 3. 9 man venti Actual Results: man complains about not finding fonts, since they are bzip2ed Expected Results: the manpage of venti gets displayed
Created attachment 154805 [details] This ebuild works The patch is ommited (yeah, i was lazy, and it's not really needed), and a newer version of plan9port is used.
Wouldn't it be better to patch the "plan9port man program" instead?
Wouldn't that mean having to add C code for 9 man to be able to bunzip manpages before displaying them? If that's the case, that's way beyond my knowledge level (for now, i hope). And, for the record, actually the new ebuild doesn't work.
Well, i see that $PLAN9/bin/man is a shellscript, still, the fonts are compressed as well, so i don't really know what good that would do. This is what i had to do (the ugly hack) on archlinux when i made a package from inferno-os aswell, because arch compresses manpages by default too, it didn't mess up the way this ebuild messes up, though, any ideas why it behaves this way?
Moving to maintainer-needed since Anant left Gentoo.
This has been fixed without an ugly hack in the ebuild posted to bug 273890 (which is also in sunrise). So this bug should probably be closed now.
It shouldn't be closed until sunrise ebuild reaches main portage.
Package removed from tree.
