Bug 198111
| Summary: | directfb requires libsysfs.so.1 no longer in -development | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Matt Domsch <matt_domsch> | ||||||||
| Component: | directfb | Assignee: | Thomas Vander Stichele <thomas> | ||||||||
| Status: | CLOSED RAWHIDE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||||||
| Severity: | medium | Docs Contact: | |||||||||
| Priority: | medium | ||||||||||
| Version: | rawhide | CC: | che666, cragel, extras-qa, hdegoede, jsacco, jukka, jwboyer, matthias, mtasaka, nicolas.mailhot, redhat, wtogami | ||||||||
| Target Milestone: | --- | ||||||||||
| Target Release: | --- | ||||||||||
| Hardware: | All | ||||||||||
| OS: | Linux | ||||||||||
| Whiteboard: | |||||||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||||||
| Doc Text: | Story Points: | --- | |||||||||
| Clone Of: | Environment: | ||||||||||
| Last Closed: | 2006-09-15 14:22:29 UTC | Type: | --- | ||||||||
| Regression: | --- | Mount Type: | --- | ||||||||
| Documentation: | --- | CRM: | |||||||||
| Verified Versions: | Category: | --- | |||||||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||||
| Embargoed: | |||||||||||
| Attachments: |
|
||||||||||
|
Description
Matt Domsch
2006-07-09 19:47:09 UTC
http://buildsys.fedoraproject.org/logs/fedora-development-extras/12340-directfb-0.9.25.1-2.fc6/x86_64/build.log Build Failure on x86_64 in FE6. Hmm, appearantly 0.9.25 breaks with x86_64, which is actually lucky though since currently in the repo is 0.9.24 and directfb changes its soname each upstream release! So what needs to be done is co 0.9.24-5 and make that 0.9.25-6 with the BR fix and build that one! Created attachment 132277 [details]
DirectFB-0.9.24-no-linux-compiler.h.patch
More is needed.
configure --enable-gettid=no
and a patch to not try including linux/compiler.h in one file (attached)
After doing this, I was able to successfully build the package on i386.
Since as sorta usual when this happens the directfb owner himself is not responding in a timely matter and since I've done a rogue takeover of it once and he then all of a sudden resurface and wanted (and got) it back, maybe its time to just kick directfb from the -devel branch. AFAIK even the directfb owner himself has said as much as that he has sorta lost interest. With directfb kicked, packages now using it can be easily rebuild to live without it, since afaik there are no packages which cannot live without it, mplayer from that-other-repo uses directfb (In reply to comment #6) > mplayer from that-other-repo uses directfb I know but it can be build without it just fine, actually uses directfb is a big word, mplayer is compiled so that it can use directfb when the user instructs it to use directfb, but by default it doesn't. please someone mark bug #198531 as a dupe of this one *** Bug 198531 has been marked as a duplicate of this bug. *** Maybe for x86_64, adding --enable-fbdev=no to configure option will be sufficient?
(this option suppresses building directfb-0.9.25/systems/libdirectfb_fbdev.so)
Note: directfb-0.9.25.1-2.fc6 also fails rebuilding on i386.
For this issue,
perl -pi -e's,#include <linux/compiler.h>,,' \
interfaces/IDirectFBVideoProvider/idirectfbvideoprovider_v4l.c
is needed.
0.9.25 has a different soname, so building that while fixing the libsysfs problem will only introduce new problems further down (up?) the dep chain. If you want to help please focus on fixing / testing 0.9.24 Thanks, Hans Created attachment 132740 [details]
Patch to compile 0.9.24 on current rawhide
As for directfb-0.9.24-5.fc5, changing BR from sysfsutils-devel to
libsysfs-devel and applying the patch attached will be sufficient for
i386 ( I cannot try x86_64 or ppc ).
This patch is almost the same as Matt's patch.
Created attachment 132742 [details]
compressed tarball containing a spec file + 3 patch files
I patched version 0.9.24 to build and run on a 32-bit PPC under the current
rawhide environment. See attachment.
-Joseph
Any chance of this getting fixed any time soon? It's been almost a month now. (26 days according to the broken dependencies report) (In reply to comment #14) > Any chance of this getting fixed any time soon? It's been almost a month now. > (26 days according to the broken dependencies report) Well the impact of this bug has been severely limited by a new version of mplayer from that other repo which simply fixes this by no longer linking / using directfb (as I suggested in commnet 7). So AFAIK nothing is using directfb, so I think we should remove it altogether from the repo. But thats up to Thomas. Hello Thomas, ping? gstreamer-plugins-bad from freshrpms still needs it. *** Bug 203841 has been marked as a duplicate of this bug. *** *** Bug 201666 has been marked as a duplicate of this bug. *** Still coming up as a daily exception on yum. Is this a more complex issue or perhaps just put off until 'final'? If poked at this over various weekends, and not having an x86_64 box I have directfb-using cards in I can actually test anything with is not helping. Hans is obviously bitter at me, and directfb has already given me a lot more grief for maintaining it than it has given me any enjoyment at all, so if Hans wants to turn his bitterness into positive energy he is more than welcome to update directfb, break everyone's repository a few times, and get lots of crap by taking over the package. So unless someone objects I am going to put it up for adoption. (In reply to comment #20) > Hans is obviously bitter at me, and directfb has already given me a lot more > grief for maintaining it than it has given me any enjoyment at all, so if Hans > wants to turn his bitterness into positive energy he is more than welcome to > update directfb, break everyone's repository a few times, and get lots of crap > by taking over the package. > I'm not bitter with you at all, to quote myself from comment 5: "Since as sorta usual when this happens the directfb owner himself is not responding in a timely matter" So I'm not bitter, but I'm worried about you not responding in a timely manner the last time directfb broke and not responding in a timely manner this time too. So I see a pattern here. > So unless someone objects I am going to put it up for adoption. Yes orphaning it is a good idea, then it will get removed from the repo for FC-6, as I already suggested in comment 15, where I wrote: "So AFAIK nothing is using directfb, so I think we should remove it altogether from the repo. But thats up to Thomas." Appologies if I sounded bitter somewhere and keep up the good work with your other pakcages! Hans mentions that 'nothing is using directfb' so I've removed it, but now Firefox crashes on some pages, unfortunately without giving feedback in the terminal. Could this be connected, as Firefox has been completely stable prior to removing directfb? If so, any possible correction, as it can't be reinstalled with the current sysfsutils problem. I'm about to push 0.9.25.1 for FC6, any objections? Only two trivial patches are required to get it to build, one general (the linux include) and one for x86_64 (asm types). Matthias does that mean that you are taking over as maintainer? Have you discussed this with Thomas? Firing of a one time build is not going to help anyone in that case it would actually be better to let directfb get removed for FC6. Yup, spoke to Thomas quickly about it. If no one wants to take over the package, I will... and I've got it ready for an FC6 rebuild. So either it's a one time build just to help the next maintainer, or it's taking care of the issue as the new maintainer. |