Bug 508124

Summary:	gnome2.eclass: append GST_INSPECT=$(type -P true) to configure
Product:	Gentoo Linux	Reporter:	Pacho Ramos <pacho>
Component:	[OLD] GNOME	Assignee:	Gentoo Linux Gnome Desktop Team <gnome>
Status:	RESOLVED FIXED
Severity:	normal	CC:	esigra, gstreamer
Priority:	Normal
Version:	unspecified
Hardware:	All
OS:	Linux
URL:	https://bugzilla.gnome.org/show_bug.cgi?id=744135
See Also:	https://bugzilla.gnome.org/show_bug.cgi?id=744134 https://bugs.gentoo.org/show_bug.cgi?id=570624 https://bugs.gentoo.org/show_bug.cgi?id=508096
Whiteboard:
Package list:		Runtime testing required:	---
Bug Depends on:
Bug Blocks:	262490

Description Pacho Ramos gentoo-dev

2014-04-19 17:05:44 UTC

We already need it for cheese, sound-juicer and rhythmbox. Is there any case when setting GST_INSPECT to true is harmful? Adding this to the eclass with help to prevent users from getting random sandbox failures when a package calls gst-inspect (or starts to do that)

What do you think?

Thanks

Comment 1 Gilles Dartiguelongue (RETIRED) gentoo-dev

2014-04-19 22:36:19 UTC

not sure its place should be in gnome2.eclass.
Also, not all packages respect this variable, a few I had to actively patch for them to respect it and leave the PM do the job.

Comment 2 Pacho Ramos gentoo-dev

2014-04-20 08:38:22 UTC

Well, the idea would be to at least try to cover the packages honoring that variable. Regarding the place, I always doubt if PMs should pass some variables/options for "cleaning" purposes (like exporting HOME and so), but that suggestions always get stalled (like bug 444568)

Comment 3 Gilles Dartiguelongue (RETIRED) gentoo-dev

2014-04-20 11:33:39 UTC

Just saw bug #508096, imho this is not the solution for it (and iirc it does not look that variable up). The solution is to report upstream what plugin tries to open this device while simply scanning the available plugins.

When I last spoke to upstream about this, they said it was a bug and needed to be reported.

Comment 4 Pacho Ramos gentoo-dev

2014-04-20 11:51:20 UTC

Will try to report then if I figure out how to find that offending plugins ;)

Comment 5 Pacho Ramos gentoo-dev

2015-02-13 11:38:35 UTC

I would still export it to at least cover the packages that respect that variable instead of waiting them to randomly break and come after fixing/workarounding them one by one

Comment 6 Pacho Ramos gentoo-dev

2016-04-03 13:01:13 UTC

Done, for the packages not respecting it we will still need to figure out about how to handle them in bug 570624 (that will likely be harder as upstream think we are wrong preventing that accesses... please note that using addpretend for them and, then, still forbidding the access but not dying on our side, makes the packages happy)