Bug 152622 – x11-plugins/gaim-encryption-3.0_beta6 install errors

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Aaron Stone 2006-10-23 22:33:28 UTC]

I know, I unmasked it, so I should be debugging it, but here's the error first. I'll see about providing a patch in a few minutes.

>>> Install gaim-encryption-3.0_beta6 into /var/tmp/portage/gaim-encryption-3.0_beta6/image/ category x11-plugins
make  install-recursive
make[1]: Entering directory `/var/tmp/portage/gaim-encryption-3.0_beta6/work/gaim-encryption-3.0beta6'
Making install in po
make[2]: Entering directory `/var/tmp/portage/gaim-encryption-3.0_beta6/work/gaim-encryption-3.0beta6/po'
/bin/sh `case "@MKINSTALLDIRS@" in /*) echo "@MKINSTALLDIRS@" ;; *) echo "../@MKINSTALLDIRS@" ;; esac` /var/tmp/portage/gaim-encryption-3.0_beta6/image//usr/share
/bin/sh: ../@MKINSTALLDIRS@: No such file or directory
make[2]: *** [install-data-yes] Error 127
make[2]: Leaving directory `/var/tmp/portage/gaim-encryption-3.0_beta6/work/gaim-encryption-3.0beta6/po'
make[1]: *** [install-recursive] Error 1
make[1]: Leaving directory `/var/tmp/portage/gaim-encryption-3.0_beta6/work/gaim-encryption-3.0beta6'
make: *** [install] Error 2

[Comment 1 Aaron Stone 2006-10-23 22:55:25 UTC]

Nope, no clue. I have a hard enough time debugging my own's software automake craziness, digging into someone else's is nearly impossible. Help needed. Thanks!

[Comment 2 dongxu li 2006-10-24 00:49:40 UTC]

Please apply the following patch to the ebuild. It should install with MKINSTALLDIRS set.


diff -purN gaim-encryption-old/gaim-encryption-3.0_beta6.ebuild gaim-encryption/gaim-encryption-3.0_beta6.ebuild
--- gaim-encryption-old/gaim-encryption-3.0_beta6.ebuild        2006-10-23 19:05:05.000000000 -0500
+++ gaim-encryption/gaim-encryption-3.0_beta6.ebuild    2006-10-24 02:28:10.000000000 -0500
@@ -30,6 +30,6 @@ src_compile() {
 }
 
 src_install() {
-       make install DESTDIR="${D}" || die "Install failed"
+       make install DESTDIR="${D}" MKINSTALLDIRS="mkinstalldirs" || die "Install failed"
        dodoc CHANGELOG INSTALL NOTES README TODO VERSION WISHLIST
 }

[Comment 3 Laurento Frittella (mrfree) 2006-10-24 01:17:43 UTC]

(In reply to comment #2)
> Please apply the following patch to the ebuild. It should install with
> MKINSTALLDIRS set.

Ok, works here ;)
Thanks

[Comment 4 junhao 2006-10-24 06:04:48 UTC]

patch worked for me too. Thanks! =)

[Comment 5 Brandon Grace 2006-10-24 07:45:21 UTC]

works for me.

[Comment 6 Mihai Moldovan 2006-10-24 08:07:34 UTC]

Next time attach the patch as a file please, else there are complications due to the automagic newline setting.

Confirming, patch is working.

[Comment 7 Sergio Bevilacqua 2006-10-24 09:00:01 UTC]

i have the same problem.

the patch fixes the problem

[Comment 8 Jory A. Pratt 2006-10-24 11:57:05 UTC]

this is not the proper fix and will not be added. I am talking with Bill upstream for a proper solution.

[Comment 9 Martin von Gagern 2006-10-24 13:59:19 UTC]

(In reply to comment #8)
> this is not the proper fix and will not be added.

Why not? OK, the build system should handle this, but until it does, this ad hoc "solution" does well enough for most people including me, so why not have it in ~arch unless it causes some kind of problem? As long as the issue is still addressed upstream, there is not much point in waiting, is there?

Imho for ~arch a hacky but working ebuild is better than a masked ebuild is better than an ebuild that fails to compile on a lot of systems.

[Comment 10 Jory A. Pratt 2006-10-24 14:03:24 UTC]

I will have gothgirl add it when she gets in tonight from work :)

I just rolled a new tarball with the final fixes from upstream that myself and Bill have discussed.

[Comment 11 Jory A. Pratt 2006-10-24 14:04:43 UTC]

(In reply to comment #9)
> (In reply to comment #8)
> > this is not the proper fix and will not be added.
> 
> Why not? OK, the build system should handle this, but until it does, this ad
> hoc "solution" does well enough for most people including me, so why not have
> it in ~arch unless it causes some kind of problem? As long as the issue is
> still addressed upstream, there is not much point in waiting, is there?
> 
> Imho for ~arch a hacky but working ebuild is better than a masked ebuild is
> better than an ebuild that fails to compile on a lot of systems.
> 

I understand your frustration but tonight gothgirl will commit a revision with the new tarball which has already addressed the issue.

[Comment 12 Martin von Gagern 2006-10-24 14:12:20 UTC]

(In reply to comment #11)
> > [...] this ad hoc "solution" does well enough for most people including me,
> I understand your frustration

Not really frustrated, I modified the ebuild locally. :)
Just didn't understand your motivation, and I wanted to save others the trouble of looking at bugzilla before they can upgrade their system.

> but tonight gothgirl will commit a revision with
> the new tarball which has already addressed the issue.

OK, that was fast enough. The keywords "upsteam" and "proper solution" sounded like we'd have to wait much longer than adding this variable would have taken.
Thanks for your effort, btw. :)

[Comment 13 Jory A. Pratt 2006-10-24 14:13:35 UTC]

Created attachment 100414 [details]
re-roll of gaim-encrytpion source

Here is the new source tarball I will add a diff for it in a bit

[Comment 14 Jory A. Pratt 2006-10-24 14:19:14 UTC]

Created attachment 100415 [details]
changes for -r1 ebuild

Here is the patch for what will be a revision so gaim-encryption-3.0_beta6-r1

[Comment 15 Olivier Crete (RETIRED) 2006-10-24 19:43:54 UTC]

its fixed

[Comment 16 Randall Wald 2006-10-24 20:58:08 UTC]

*** Bug 152740 has been marked as a duplicate of this bug. ***