Problem in malloc.c, rev. 1.183 while building phonon-4.2.0
Kyle Allender
kallender at completecomputing.com
Mon Dec 1 18:22:34 PST 2008
Jason Evans wrote:
> Kyle Allender wrote:
>> Stefan Farfeleder wrote:
>>> On Mon, Dec 01, 2008 at 07:34:17AM -0600, Kyle Allender wrote:
>>>> Stefan Farfeleder wrote:
>>>>> On Mon, Dec 01, 2008 at 07:06:32AM -0600, Kyle Allender wrote:
>>>>>> Hello.
>>>>>> Generating mediaobjecttest.moc
>>>>>> Assertion failed: ((ret << BLOCK_COST_2POW) != 0), function
>>>>>> malloc_spin_lock, file /usr/src/lib/libc/stdlib/malloc.c, line 1287.
>>>>>> Abort trap (core dumped)
>>>>> Hi,
>>>>>
>>>>> this assertion has been fixed yesterday (svn r185483).
>>>> Should I rerun build/installworld before re-attempting the kde build
>>>> or just run make clean in the phonon directory?
>>>
>>> You need at least an updated libc. build/installworld will do that but
>>> there are faster ways for this small update. After that you can
>>> continue building kde.
>>>
>> OK. I've kicked off a buildworld as that's about the only method I
>> know that will resolve this issue. The "faster ways" I'm unaware of -
>> if someone could enlighten me, that'd be great. Buildworld will work
>> in the meantime, though.
>
> In this case of libc, it's usually a good idea to do the full
> buildworld, since libc gets statically linked into various toolchain
> binaries. Fortunately, this bug in malloc only affects multi-threaded
> applications; otherwise you'd have an unrecoverable system (though I
> would have noticed that before committing).
>
> Jason
>
I ended up issuing a cd /usr/src/lib/libc; cd /usr/src/lib/libc; make &&
make install which did the trick - the KDE4 installation was able to
proceed after that. I appreciate your help and will make a note of the
better ways to do what I was after this morning.
I would suppose I should go ahead and kick off a buildworld when the kde
build completes later this evening. Will that necessitate having to
rebuild the ports or would that be possible to put off until a later
time? The system should remain stable correct?
K
More information about the freebsd-current
mailing list