[Gambas-user] Segfault with rev 7562

[adamnt42 at ...626...]

Wow. I finally found it! Somehow the source (here) for gbx_class.c got corrupted with a bad character in the file. Even though svn said it was OK and updated, it was still the bad version that was being compiled.
(There was an #07 character/byte in the middle of the source file - it was only visible in hexedit .... ?)
C'est la vie!
Sorry for the noise.

 best regards
bruce
 
On Thu, 7 Jan 2016 10:31:24 +0100
Benoît Minisini <gambas at ...1...> wrote:

> Le 07/01/2016 05:26, adamnt42 at ...626... a écrit :
> > I getting a segfault with gbr3.
> >
> > It happens when running any executable archive including the IDE.
> >
> > The gdb output is:
> > Program received signal SIGSEGV, Segmentation fault.
> > EXEC_special (special=11, class=0x80ac13c, object=0x813e68c, nparam=0, drop=1 '\001') at gbx_exec.c:1649
> > 1649		if (CLASS_DESC_get_type(desc) == CD_STATIC_METHOD)
> > 2: index = 0
> > 1: desc = (CLASS_DESC *) 0x0
> >
> > The routine in gbx_exec.c is : bool EXEC_special(int special, CLASS *class, void *object, int nparam, bool drop).
> > The introductory code is:
> > 	CLASS_DESC *desc;
> > 	short index = class->special[special];
> >
> > 	if (index == NO_SYMBOL)
> > 		return TRUE;
> >
> > 	desc = CLASS_get_desc(class, index);
> >
> > 	if (CLASS_DESC_get_type(desc) == CD_STATIC_METHOD)
> >
> > The problem is that index is 0, this is occurring invariably when an Object[] class is being initialised. The backtrace is:
> > (gdb) backtrace
> > #0  0x08078d4e in EXEC_special (special=11, class=0x80ac13c, object=0x813e68c, nparam=0, drop=1 '\001')
> >      at gbx_exec.c:1649
> > #1  0x08079acb in EXEC_new () at gbx_exec.c:1947
> > #2  0x08082ac4 in EXEC_loop () at gbx_exec_loop.c:916
> > #3  0x08076da5 in EXEC_function_loop () at gbx_exec.c:931
> > #4  0x0807763f in EXEC_function_real () at gbx_exec.c:895
> > #5  0x08051168 in CLASS_run_inits (class=0x813a994) at gbx_class_load.c:1287
> > #6  0x0805120e in CLASS_load_real (class=0x813a994) at gbx_class_load.c:1323
> > #7  0x08062f15 in ARCHIVE_load_exported_class (arch=0x80c8584, pass=1) at gbx_archive.c:192
> > #8  0x0807035b in COMPONENT_load_all_finish () at gbx_component.c:147
> > #9  0x08055da3 in PROJECT_load_finish () at gbx_project.c:486
> > #10 0x0806530c in init (file=file at ...2861...=0xbfffeeeb "./gambas3.gambas", argc=1, argv=0xbfffecc4) at gbx.c:103
> > #11 0x0804b390 in main (argc=1, argv=0xbfffecc4) at gbx.c:372
> >
> > A bit more info:
> > (gdb) disp class->special
> > 3: class->special = {12, -1, 21, 15, -1, 22, -1, -1, -1, -1, -1, 0}
> > (gdb) disp special
> > 4: special = 11
> >
> > i.e.  index = class->special[special] is the last entry in the class->special array.
> > and then
> > desc = CLASS_get_desc(class, index);
> >
> > (gdb) disp desc
> > 5: desc = (CLASS_DESC *) 0x0
> >
> > Finally, It happens the second time that routine is called. The first call executes properly (the value of the "special" param is 0).
> >
> > N.B. this is on a clean copy of the gambas source, not our local version.
> >
> > Any other info needed?
> >
> > regards
> > bruce
> >
> 
> No problem there... The 11th special method ("_ready") has been added in 
> revision r7541.
> 
> Did you try a recompilation and reinstallation from scratch?
> 
> Are you sure that you didn't have binary packages of previous gambas 
> versions installed that you didn't uninstall?
> 
> -- 
> Benoît Minisini
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> Gambas-user mailing list
> Gambas-user at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/gambas-user


-- 
B Bruen <adamnt42 at ...3379... (sort of)>