Attachment 403074 Details for Bug 549166 – an old irc log

an old irc log

2014-06-17.191938+0200CEST.html (text/html), 8.41 KB, created by Rafał Mużyło on 2015-05-11 22:11:39 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Rafał Mużyło

Created: 2015-05-11 22:11:39 UTC

Size: 8.41 KB

patch

obsolete

><html><head><meta http-equiv="content-type" content="text/html; charset=UTF-8"><title>Conversation with mgorny at wto, 17 cze 2014, 19:19:38 on galtgendo@irc.freenode.net (irc)</title></head><body><h3>Conversation with mgorny at wto, 17 cze 2014, 19:19:38 on galtgendo@irc.freenode.net (irc)</h3>
>(19:19:56) galtgendo: ping 
>(19:20:01) mgorny: pong 
>(19:20:57) galtgendo: would you mind addressing <a href="https://bugs.gentoo.org/show_bug.cgi?id=492544#c11">https://bugs.gentoo.org/show_bug.cgi?id=492544#c11</a> ? 
>(19:21:58) mgorny: sure, i&apos;ll try to in a few minutes 
>(19:22:07) mgorny: after i commit xcb-util-* and libnsfb 
>(19:29:49) galtgendo: also, if you don't mind, even if eventually there'll be a better fix for the main topic of this bug, would you you fix that in glib ebuild too ? 
>(19:30:37) mgorny: i will fix it only in glib ebuild :P 
>(19:31:05) galtgendo: why not in gtk+ too ? 
>(19:33:53) mgorny: oh, sorry, i thought it&apos;s only one of the two :P 
>(19:35:48) galtgendo: well, while most likely there are others, these two (glib and gtk+) (actually, everything between too, to get the most from devhelp) are most important 
>(19:38:27) galtgendo: on that note: I have yet to test multilib gstreamer (and this part is most likely somewhat broken in 1.2, but fixed in 1.3), but did you took steps so this problem wouldn't happen there ? 
>(19:39:07) mgorny: yes 
>(19:39:14) mgorny: every new ebuild i&apos;m committing has proper gtk-doc hacks 
>(19:45:57) mgorny: did you notice any more broken packages? 
>(19:46:03) mgorny: atk has the hack already 
>(19:46:14) galtgendo: odd, I'm looking at both gst-plugins-good ebuild *and* gstreamer eclass and just don't see the workaround implemented anywhere 
>(19:46:35) mgorny: and gtk+ seems to have the hack as well 
>(19:47:14) mgorny: hmm, i guess i missed the fact gst-plugins-good was supposed to install docs 
>(19:47:34) mgorny: so are you sure gtk+ suffers this? 
>(19:48:11) galtgendo: OK, let's slow down a bit - I can't type that fast 
>(19:48:32) galtgendo: first gst-plugins-* 
>(19:48:35) mgorny: no 
>(19:48:38) mgorny: glib goes first :P 
>(19:48:40) mgorny: and i&apos;m rebuilding it 
>(19:49:37) mgorny: glib done 
>(19:49:42) mgorny: now, what&apos;s with gtk+? 
>(19:50:55) galtgendo: this is the part that somewhat broken upstream right now, but fixed in 1.3 - that was actually fixed after I pointed out in a bug that all gst-plugins-* tarballs should uniformly implement a scheme that was already implemented properly in one of them 
>(19:52:26) galtgendo: you see, due to quite odd reasons, upstream took steps so gstreamer and it's plugins could be built *from git* without gtk-doc installed 
>(19:53:50) galtgendo: but initially they've done it in a way that prevented the docs from the tarballs to be installed without '--enable-gtk-doc' (so, a doc rebuild) 
>(19:54:37) galtgendo: eventually, they've implemented that correctly but not in every one of those tarballs 
>(19:56:39) galtgendo: after I've raise this issue (somewhere in March, IIRC), it was fixed in master - meaning 1.3, so once 1.4 arrives (which, according to some posts, should be pretty soon), it should work properly 
>(19:57:03) mgorny: so no hack needed, or should i hack it? 
>(19:57:09) mgorny: [PatchPreferred] 
>(19:58:28) galtgendo: yes, the hack will be necessary, as out-of-tree builds (like multilib eclass) are a different problem 
>(19:59:10) galtgendo: just you won't see it working with most of 1.2 
>(20:00:13) galtgendo: it was gnome bug 725034 
>(20:00:21) mgorny: oh, so they broke our hack? ;P 
>(20:01:07) galtgendo: no, they've broken their own hack, eclasses are our (Gentoo) own issue 
>(20:02:09) mgorny: ok 
>(20:02:12) mgorny: now please make it clear 
>(20:02:19) mgorny: do you want me to fix something else or can i go get some rest? :P 
>(20:03:29) galtgendo: now, as for gtk+, as I've said in that comment, the hack in the ebuild is incorrect, as you need the links to each of html dirs otherwise they're not getting picked up 
>(20:03:50) galtgendo: in short order: 
>(20:04:22) mgorny: ok 
>(20:04:41) galtgendo: glib needs the hack in place, so that the docs get installed 
>(20:05:10) galtgendo: gtk+ needs a fix to the hack , so that it actually works 
>(20:06:10) mgorny: gdk gtk libgail-util 
>(20:06:13) mgorny: did i miss something? 
>(20:06:14) galtgendo: gst-plugins-* need the hack to be future proof, as it won't work for some of gst-plugins-* in 1.2, but it will in 1.4 
>(20:08:06) mgorny: all of them have docs? 
>(20:08:18) galtgendo: gdk gtk libgail-util - about right 
>(20:09:35) galtgendo: yess, even gail-util does - even if very little 
>(20:10:36) mgorny: so now i rebuild gtk+ so patience ;P 
>(20:10:59) galtgendo: as for gst-plugins-*, read that bug - in the initial report I've wrote which are working and which don't 
>(20:13:01) galtgendo: and while I was only interested in three main tarballs, upstream fixed it in -libav too (even if those docs are of somewhat low quality - there's just too much to document) 
>(20:19:35) mgorny: gtk+ fixed 
>(20:22:27) galtgendo: thanks 
></body></html>

<html><head><meta http-equiv="content-type" content="text/html; charset=UTF-8"><title>Conversation with mgorny at wto, 17 cze 2014, 19:19:38 on galtgendo@irc.freenode.net (irc)</title></head><body><h3>Conversation with mgorny at wto, 17 cze 2014, 19:19:38 on galtgendo@irc.freenode.net (irc)</h3>
<font color="#16569E"><font size="2">(19:19:56)</font> <b>galtgendo:</b></font> ping<br/>
<font color="#A82F2F"><font size="2">(19:20:01)</font> <b>mgorny:</b></font> pong<br/>
<font color="#16569E"><font size="2">(19:20:57)</font> <b>galtgendo:</b></font> would you mind addressing <a href="https://bugs.gentoo.org/show_bug.cgi?id=492544#c11">https://bugs.gentoo.org/show_bug.cgi?id=492544#c11</a> ?<br/>
<font color="#A82F2F"><font size="2">(19:21:58)</font> <b>mgorny:</b></font> sure, i&apos;ll try to in a few minutes<br/>
<font color="#A82F2F"><font size="2">(19:22:07)</font> <b>mgorny:</b></font> after i commit xcb-util-* and libnsfb<br/>
<font color="#16569E"><font size="2">(19:29:49)</font> <b>galtgendo:</b></font> also, if you don't mind, even if eventually there'll be a better fix for the main topic of this bug, would you you fix that in glib ebuild too ?<br/>
<font color="#A82F2F"><font size="2">(19:30:37)</font> <b>mgorny:</b></font> i will fix it only in glib ebuild :P<br/>
<font color="#16569E"><font size="2">(19:31:05)</font> <b>galtgendo:</b></font> why not in gtk+ too ?<br/>
<font color="#A82F2F"><font size="2">(19:33:53)</font> <b>mgorny:</b></font> oh, sorry, i thought it&apos;s only one of the two :P<br/>
<font color="#16569E"><font size="2">(19:35:48)</font> <b>galtgendo:</b></font> well, while most likely there are others, these two (glib and gtk+) (actually, everything between too, to get the most from devhelp) are most important<br/>
<font color="#16569E"><font size="2">(19:38:27)</font> <b>galtgendo:</b></font> on that note: I have yet to test multilib gstreamer (and this part is most likely somewhat broken in 1.2, but fixed in 1.3), but did you took steps so this problem wouldn't happen there ?<br/>
<font color="#A82F2F"><font size="2">(19:39:07)</font> <b>mgorny:</b></font> yes<br/>
<font color="#A82F2F"><font size="2">(19:39:14)</font> <b>mgorny:</b></font> every new ebuild i&apos;m committing has proper gtk-doc hacks<br/>
<font color="#A82F2F"><font size="2">(19:45:57)</font> <b>mgorny:</b></font> did you notice any more broken packages?<br/>
<font color="#A82F2F"><font size="2">(19:46:03)</font> <b>mgorny:</b></font> atk has the hack already<br/>
<font color="#16569E"><font size="2">(19:46:14)</font> <b>galtgendo:</b></font> odd, I'm looking at both gst-plugins-good ebuild *and* gstreamer eclass and just don't see the workaround implemented anywhere<br/>
<font color="#A82F2F"><font size="2">(19:46:35)</font> <b>mgorny:</b></font> and gtk+ seems to have the hack as well<br/>
<font color="#A82F2F"><font size="2">(19:47:14)</font> <b>mgorny:</b></font> hmm, i guess i missed the fact gst-plugins-good was supposed to install docs<br/>
<font color="#A82F2F"><font size="2">(19:47:34)</font> <b>mgorny:</b></font> so are you sure gtk+ suffers this?<br/>
<font color="#16569E"><font size="2">(19:48:11)</font> <b>galtgendo:</b></font> OK, let's slow down a bit - I can't type that fast<br/>
<font color="#16569E"><font size="2">(19:48:32)</font> <b>galtgendo:</b></font> first gst-plugins-*<br/>
<font color="#A82F2F"><font size="2">(19:48:35)</font> <b>mgorny:</b></font> no<br/>
<font color="#A82F2F"><font size="2">(19:48:38)</font> <b>mgorny:</b></font> glib goes first :P<br/>
<font color="#A82F2F"><font size="2">(19:48:40)</font> <b>mgorny:</b></font> and i&apos;m rebuilding it<br/>
<font color="#A82F2F"><font size="2">(19:49:37)</font> <b>mgorny:</b></font> glib done<br/>
<font color="#A82F2F"><font size="2">(19:49:42)</font> <b>mgorny:</b></font> now, what&apos;s with gtk+?<br/>
<font color="#16569E"><font size="2">(19:50:55)</font> <b>galtgendo:</b></font> this is the part that somewhat broken upstream right now, but fixed in 1.3 - that was actually fixed after I pointed out in a bug that all gst-plugins-* tarballs should uniformly implement a scheme that was already implemented properly in one of them<br/>
<font color="#16569E"><font size="2">(19:52:26)</font> <b>galtgendo:</b></font> you see, due to quite odd reasons, upstream took steps so gstreamer and it's plugins could be built *from git* without gtk-doc installed<br/>
<font color="#16569E"><font size="2">(19:53:50)</font> <b>galtgendo:</b></font> but initially they've done it in a way that prevented the docs from the tarballs to be installed without '--enable-gtk-doc' (so, a doc rebuild)<br/>
<font color="#16569E"><font size="2">(19:54:37)</font> <b>galtgendo:</b></font> eventually, they've implemented that correctly but not in every one of those tarballs<br/>
<font color="#16569E"><font size="2">(19:56:39)</font> <b>galtgendo:</b></font> after I've raise this issue (somewhere in March, IIRC), it was fixed in master - meaning 1.3, so once 1.4 arrives (which, according to some posts, should be pretty soon), it should work properly<br/>
<font color="#A82F2F"><font size="2">(19:57:03)</font> <b>mgorny:</b></font> so no hack needed, or should i hack it?<br/>
<font color="#A82F2F"><font size="2">(19:57:09)</font> <b>mgorny:</b></font> [PatchPreferred]<br/>
<font color="#16569E"><font size="2">(19:58:28)</font> <b>galtgendo:</b></font> yes, the hack will be necessary, as out-of-tree builds (like multilib eclass) are a different problem<br/>
<font color="#16569E"><font size="2">(19:59:10)</font> <b>galtgendo:</b></font> just you won't see it working with most of 1.2<br/>
<font color="#16569E"><font size="2">(20:00:13)</font> <b>galtgendo:</b></font> it was gnome bug 725034<br/>
<font color="#A82F2F"><font size="2">(20:00:21)</font> <b>mgorny:</b></font> oh, so they broke our hack? ;P<br/>
<font color="#16569E"><font size="2">(20:01:07)</font> <b>galtgendo:</b></font> no, they've broken their own hack, eclasses are our (Gentoo) own issue<br/>
<font color="#A82F2F"><font size="2">(20:02:09)</font> <b>mgorny:</b></font> ok<br/>
<font color="#A82F2F"><font size="2">(20:02:12)</font> <b>mgorny:</b></font> now please make it clear<br/>
<font color="#A82F2F"><font size="2">(20:02:19)</font> <b>mgorny:</b></font> do you want me to fix something else or can i go get some rest? :P<br/>
<font color="#16569E"><font size="2">(20:03:29)</font> <b>galtgendo:</b></font> now, as for gtk+, as I've said in that comment, the hack in the ebuild is incorrect, as you need the links to each of html dirs otherwise they're not getting picked up<br/>
<font color="#16569E"><font size="2">(20:03:50)</font> <b>galtgendo:</b></font> in short order:<br/>
<font color="#A82F2F"><font size="2">(20:04:22)</font> <b>mgorny:</b></font> ok<br/>
<font color="#16569E"><font size="2">(20:04:41)</font> <b>galtgendo:</b></font> glib needs the hack in place, so that the docs get installed<br/>
<font color="#16569E"><font size="2">(20:05:10)</font> <b>galtgendo:</b></font> gtk+ needs a fix to the hack , so that it actually works<br/>
<font color="#A82F2F"><font size="2">(20:06:10)</font> <b>mgorny:</b></font> gdk gtk libgail-util<br/>
<font color="#A82F2F"><font size="2">(20:06:13)</font> <b>mgorny:</b></font> did i miss something?<br/>
<font color="#16569E"><font size="2">(20:06:14)</font> <b>galtgendo:</b></font> gst-plugins-* need the hack to be future proof, as it won't work for some of gst-plugins-* in 1.2, but it will in 1.4<br/>
<font color="#A82F2F"><font size="2">(20:08:06)</font> <b>mgorny:</b></font> all of them have docs?<br/>
<font color="#16569E"><font size="2">(20:08:18)</font> <b>galtgendo:</b></font> gdk gtk libgail-util - about right<br/>
<font color="#16569E"><font size="2">(20:09:35)</font> <b>galtgendo:</b></font> yess, even gail-util does - even if very little<br/>
<font color="#A82F2F"><font size="2">(20:10:36)</font> <b>mgorny:</b></font> so now i rebuild gtk+ so patience ;P<br/>
<font color="#16569E"><font size="2">(20:10:59)</font> <b>galtgendo:</b></font> as for gst-plugins-*, read that bug - in the initial report I've wrote which are working and which don't<br/>
<font color="#16569E"><font size="2">(20:13:01)</font> <b>galtgendo:</b></font> and while I was only interested in three main tarballs, upstream fixed it in -libav too (even if those docs are of somewhat low quality - there's just too much to document)<br/>
<font color="#A82F2F"><font size="2">(20:19:35)</font> <b>mgorny:</b></font> gtk+ fixed<br/>
<font color="#16569E"><font size="2">(20:22:27)</font> <b>galtgendo:</b></font> thanks<br/>
</body></html>

Actions: View

Attachments on bug 549166: 403074