Summary: | nautilus-2.24.2 fails to build with ccache | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Shyam Mani (RETIRED) <fox2mike> |
Component: | [OLD] GNOME | Assignee: | Gentoo Toolchain Maintainers <toolchain> |
Status: | RESOLVED NEEDINFO | ||
Severity: | normal | CC: | gnome, pacho |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: | nautilus-2.24.2 build.log |
Description
Shyam Mani (RETIRED)
2008-12-05 20:40:10 UTC
Created attachment 174352 [details]
nautilus-2.24.2 build.log
Here's the build.log
@toolchain, what say you? This is really weird bug, one I've never seen before... Thanks So after going through a number of things with dang on IRC, If my CFLAGS have a -O2 then it builds fine with gcc-4.3.2, with -O3 it fails. Re-assigning to toolchain, as upstream gnome doesn't support O3. Quite possibly, toolchain is interested in fixing this. build it again with FEATURES=-ccache and -O3 in CFLAGS. that error is not indicative of a failure in gcc. (In reply to comment #5) > build it again with FEATURES=-ccache and -O3 in CFLAGS. that error is not > indicative of a failure in gcc. Okay, it builds fine with ccache disabled and 03. might be related to Bug 196243 ... what needs to go down now is you find the offending ccache object (dont delete it!) and the offending cpp output that ccache is producing but detecting as the old rather than the new ... (In reply to comment #7) > might be related to Bug 196243 ... what needs to go down now is you find the > offending ccache object (dont delete it!) and the offending cpp output that > ccache is producing but detecting as the old rather than the new ... And pray how Mr.SpanKY do I go about doing all this? :D And FWIW, mine is a 32 bit system... in which case, that implies you have cache corruption ... usually this occurs due to fdisk troubles/whatever on files in the ccache dir building lots of stuff -> locks up system -> corrupts ccache -> screws future builds to start with, retrigger the build failure and then find the md5sum of the failing file (.libs/nautilus-file-info.o). make sure that isnt a 0 byte file, and then find the file in your ccache dir that matches. (In reply to comment #10) > to start with, retrigger the build failure and then find the md5sum of the > failing file (.libs/nautilus-file-info.o). make sure that isnt a 0 byte file, > and then find the file in your ccache dir that matches. I can re-trigger the failure, got the md5sum too, but that file isn't in the cache. No matching filename for that md5sum. Suggestions? The difference it, before (the last time I looked) nautilus-file-info.o was a file with some hex data etc, and running file would show just "data". Now, the file is a clump of code and file says ASCII C Program Text. Does that give you any pointers? I could attach nautilus-file-info.o if needed. i dont suppose you feel like running a full fdisk on the disk ? the info i need is the colliding the file and the input ccache uses to get to the colliding file ... but first do the fdisk ... (In reply to comment #12) > i dont suppose you feel like running a full fdisk on the disk ? fdisk? as in fdisk -l? or fsck? > the info i need is the colliding the file and the input ccache uses to get to > the colliding file ... but first do the fdisk ... Sure, as soon as you clarify :) sorry, i was working on another bug where the guy has fdisk problems i of course mean fsck ... (In reply to comment #14) > sorry, i was working on another bug where the guy has fdisk problems Tut Tut! > i of course mean fsck ... Alright. This happens in a few hours..since it's 0400 now. I'll let you know how it goes. (In reply to comment #15) > Alright. This happens in a few hours..since it's 0400 now. I'll let you know > how it goes. So sorry. I'd done this ages ago...and never replied. ########### reiserfsck --check started at Wed Dec 10 23:47:01 2008 ########### Replaying journal.. Reiserfs journal '/dev/sda6' in blocks [18..8211]: 0 transactions replayed Checking internal tree..finished Comparing bitmaps..finished Checking Semantic tree: finished No corruptions found There are on the filesystem: Leaves 258778 Internal nodes 1692 Directories 154990 Other files 1155945 Data block pointers 11575021 (97 of them are zero) Safe links 0 ########### reiserfsck finished at Thu Dec 11 00:11:40 2008 ########### ########### reiserfsck --check started at Thu Dec 11 00:14:28 2008 ########### Replaying journal.. Reiserfs journal '/dev/sda7' in blocks [18..8211]: 0 transactions replayed Checking internal tree..finished Comparing bitmaps..finished Checking Semantic tree: finished No corruptions found There are on the filesystem: Leaves 39635 Internal nodes 268 Directories 77983 Other files 167520 Data block pointers 10749396 (35 of them are zero) Safe links 0 ########### reiserfsck finished at Thu Dec 11 00:17:04 2008 ########### That's both / and /home. It's a new comp/new install so I really didn't expect to see any issues with fsck and if you clear your ccache, does it work ? Is this still a problem you can reproduce? |