[Gcl-devel] Re: gcl failure

[Camm Maguire]

Greetings!

Dan Stanger <address@hidden> writes:

> MAXPAGE 8000 DBEGIN, 10100000 PAGEWIDTH b

Great!  Your problem certainly seems to be PAGEWIDTH.  It is defined to
'b'?  In which file?  Can you please include the whole file?  It
should be something like 11,12 or 13.

Take care,

> 
> 
> "/cygdrive/c/WINNT/System32/advapi32.dll": error reading line numbers
> 
> 
> Program received signal SIGSEGV, Segmentation fault.
> 0x0040261d in add_page_to_freelist (p=0x1a100000 "", tm=0x50a598)
>     at alloc.c:162
> 162      type_map[page(p)]= t;
> (gdb) bt
> #0  0x0040261d in add_page_to_freelist (p=0x1a100000 "", tm=0x50a598)
>     at alloc.c:162
> #1  0x004027c4 in alloc_object (t=t_string) at alloc.c:238
> #2  0x0042c5bc in alloc_simple_string (l=16) at string.d:41
> #3  0x00404106 in malloc (size=16) at alloc.c:1086
> #4  0x6102d2c4 in _libkernel32_a_iname ()
> #5  0x6102d712 in _libkernel32_a_iname ()
> #6  0x6102a997 in _libkernel32_a_iname ()
> #7  0x6105219c in _libkernel32_a_iname ()
> #8  0x61003ba1 in _libkernel32_a_iname ()
> #9  0x610041b9 in _libkernel32_a_iname ()
> #10 0x610041f9 in _libkernel32_a_iname ()
> #11 0x004f5107 in cygwin_crt0 ()
> #12 0x0040103d in mainCRTStartup ()
> #13 0x77f1b9ea in ?? ()
> (gdb)
> Camm Maguire wrote:
> 
> > Greetings, and thanks for your detailed report!
> >
> > OK, I think we need to know:
> >
> > 1) Your page macro is returning some bogus offset.  Here is the
> > definition in page.h:
> >
> > #define page(p) (((int)(((char *)(p))-DBEGIN)>>PAGEWIDTH))
> >
> > Likewise:
> > char type_map[MAXPAGE];
> >
> > So we need to know MAXPAGE, DBEGIN, and PAGEWIDTH.  You've sent me p
> > as
> > 0x1a100000.
> >
> > So I'd also like to know where this routine was called with this
> > value.  Under gdb at the point you were at, please send me the output
> > of 'bt' (backtrace).  I'm hoping its just a bad setting of the three
> > macros above in the configure process.  (If you can figure out any way
> > to get text from your nt box, that would be very helpful too.  The bmp
> > will work if you can't.)
> >
> > Also, have you successfully gotten any earlier version of gcl to work?
> > If so, what is the latest such version, and can you reproduce it so we
> > can make some comparisons?
> >
> > Finally, are you just running gcl with no command line args and stdin
> > attached to a terminal?
> >
> > Take care,
> >
> > Dan Stanger <address@hidden> writes:
> >
> > > This is a multi-part message in MIME format.
> > > --------------AF39F38253E6B3ABEF916593
> > > Content-Type: text/plain; charset=us-ascii
> > > Content-Transfer-Encoding: 7bit
> > >
> > > I built gcl cvs with -g, ran with gdb, it segfaulted at line 161.
> > > void
> > >   148 add_page_to_freelist(p,tm)
> > >   149      char *p;
> > >   150      struct typemanager *tm;
> > > - 151 {short t,size;
> > > - 152  int i=tm->tm_nppage,fw;
> > >   153  int nn;
> > >   154  object x,f;
> > > - 155  t=tm->tm_type;
> > >   156 #ifdef SGC
> > >   157  nn=page(p);
> > >   158  if (sgc_enabled)
> > >   159    { if (!WRITABLE_PAGE_P(nn)) make_writable(nn,nn+1);}
> > >   160 #endif
> > > - 161  type_map[page(p)]= t;
> > >
> > > The locals are attached in a bitmap file, as I could not copy the gdb 
> > > window.
> > > If you cant read this,
> > > I will type them in.
> > > Dan
> > >
> > >
> > >
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah