Re: The math crash

2002-11-30 Thread Darren Freeman
On Sat, 2002-11-30 at 15:03, John Levon wrote: > On Sat, Nov 30, 2002 at 02:59:36PM +1030, Darren Freeman wrote: > > > I just did a cvs update, make, and ran lyxdbg0.2b on the problem. > > > > Did you fix it in the cvs or on your local copy? I will try Angus' > > patches sometime soon for myself.

Re: The math crash

2002-11-29 Thread John Levon
On Sat, Nov 30, 2002 at 02:59:36PM +1030, Darren Freeman wrote: > I just did a cvs update, make, and ran lyxdbg0.2b on the problem. > > Did you fix it in the cvs or on your local copy? I will try Angus' > patches sometime soon for myself. I've just committed it, anoncvs has a delay. Check the ch

Re: The math crash

2002-11-29 Thread Darren Freeman
On Sat, 2002-11-30 at 14:38, John Levon wrote: > On Fri, Nov 29, 2002 at 11:25:01AM +, Angus Leeming wrote: > > > John, Darren, could you see if this patch stops the crash, at the expense > > that the Close button label does not change to Cancel. > > I've fixed it properly, please double che

Re: The math crash

2002-11-29 Thread John Levon
On Fri, Nov 29, 2002 at 11:25:01AM +, Angus Leeming wrote: > John, Darren, could you see if this patch stops the crash, at the expense > that the Close button label does not change to Cancel. I've fixed it properly, please double check regards john -- "It is very difficult to prophesy, es

Re: The math crash

2002-11-29 Thread Angus Leeming
On Wednesday 27 November 2002 8:14 pm, John Levon wrote: > #6 0x40414114 in chunk_realloc (ar_ptr=0x404b4680, oldp=0x8418c04, > oldsize=4294967292, nb=16) at malloc.c:3390 #7 0x40413dc1 in > __libc_realloc (oldmem=0x8418c0c, bytes=7) at malloc.c:3286 #8 0x400d400d > in fl_set_object_label (ob=0x

Re: The math crash

2002-11-28 Thread Angus Leeming
On Thursday 28 November 2002 9:43 am, Lars Gullik Bjønnes wrote: > What string literal? > > fl_realloc is trying to free ob->label not label. > > my Q is what is ob->label set to? is it initialized? Yes. All this suggests that the "const char *label" passed to xforms is rubbish. Ie, the problem

Re: The math crash

2002-11-28 Thread Lars Gullik Bjønnes
John Levon <[EMAIL PROTECTED]> writes: | On Wed, Nov 27, 2002 at 10:08:14PM +, Angus Leeming wrote: | | > Can't see nothing wrong with this... | > | > void fl_set_object_label(FL_OBJECT * ob, const char *label) { | > ... | > if (!label) | > label = ""; | > ... | >

Re: The math crash

2002-11-27 Thread Darren Freeman
Just an update. I don't know if anything changed in the CVS or not, but here's another trace obtained in the same manner as before, cvs update, make, lyxdbg.. etc. Darren LyX 1.3.0cvs of Fri, May 3, 2002 Built on Nov 28 2002, 14:27:15 Configuration Host type: i686-pc-linux

Re: The math crash

2002-11-27 Thread John Levon
On Wed, Nov 27, 2002 at 10:08:14PM +, Angus Leeming wrote: > Can't see nothing wrong with this... > > void fl_set_object_label(FL_OBJECT * ob, const char *label) { > ... > if (!label) > label = ""; > ... > ob->label = fl_realloc(ob->label, strlen(label) +

Re: The math crash

2002-11-27 Thread Angus Leeming
On Wednesday 27 November 2002 8:14 pm, John Levon wrote: > #6 0x40414114 in chunk_realloc (ar_ptr=0x404b4680, oldp=0x8418c04, > oldsize=4294967292, nb=16) at malloc.c:3390 #7 0x40413dc1 in > __libc_realloc (oldmem=0x8418c0c, bytes=7) at malloc.c:3286 #8 0x400d400d > in fl_set_object_label (ob=0x