On 03/12/2016 18:45, Pierre Labastie wrote: > On 03/12/2016 17:29, Ken Moffat wrote: >> On Sat, Dec 03, 2016 at 08:49:47AM -0700, Roger Koehler wrote: >>> On Fri, Dec 2, 2016 at 8:30 PM, Ken Moffat <[email protected]> wrote: >>>> If I can't make any progress in the next few days, I guess I'd >>>> better raise a BLFS ticket asserting that this is broken. >>> Could be the kernel. I just rebuilt my system with all the latest >>> (SVN) packages except for the kernel. I built with linux 3.16.39, and >>> lxdm works fine. I am on an intel x86_64 laptop. >> Could be, at this point all I know is that it's broken for me, and I >> don't know when it last worked (i.e. I forget why I gave up on it >> after BLFS-7.6) but fedora-25 will not be using anything that old. >> >> ĸen > Hi, > I know how "worksforme" is frustrating. But I have a working version of lxdm > with kernel 4.8.6 (needs to update, I know). > I am building the recent versions of the gtk stack (was still with the one > from the end of october or so). Will let you know whether it still runs after > that. > > Pierre > > > Everything still works with (sed'ed) kernel 4.8.12, and up to date BLFS. OTOH, LFS is 7.10, except that glibc is 2.24 (I recompiled it to get debug symbols when debugging the firefoc/ICU segfault).
I have looked more closely at your gdb backtrace. Glib asks iconv to convert a message from UTF-8 to ANSI_X3.4-1968 (a.k.a ASCII). I do not see how it could be related to anything about X (fonts or whatever). It'd be interesting to understand why glib wants the conversion, since UTF-8 seems to be the default encoding used by glib... Anyway, the bug seems that _int_malloc segfaults. The gdb trace does not have details on what is passed to malloc, but I'd say that it should be guarded against wrong parameters and return an error rather than segfaulting. That's all I can say for now. Sincerely, Pierre -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
