[Gcl-devel] Re: 2.6.5 errata page updates

[Camm Maguire]

Greetings, and thanks!

"Mike Thomas" <address@hidden> writes:

> Hi Camm.
> 
> | 3)  Any additions to the errata page are welcome.  Mike, don't you have
> |     one?  We can keep track of essential items which must go into a
> |     2.6.6 here in case we ever release one.  Thoughts here are most
> |     welcome as well.
> 
> Putting aside updating the Windows object linker (despite pondering this
> question for the past 24 hours I can't recall if that is necessary or not -
> I believe not), the following change is required to make 2.6.5 work with
> MinGW32 gcc 3.4.2 which seems to want to call malloc() in it's pre-main
> initialisation phase.
> 
> diff -r /c/lang/source/gcl/gcl-2.6.5/o/alloc.c
> /c/lang/source/gcl/gcl-2.6.5-ansi
> /o/alloc.c
> 1487,1489d1486
> < #ifdef  SET_REAL_MAXPAGE
> <                 SET_REAL_MAXPAGE
> < #endif
> 

Here is what we have now in malloc:

#ifdef       __MINGW32__
                /* If malloc() gets called by the C runtime before
                 * main starts and the shared memory is not yet
                 * initialised causing boofo. 
                 * SET_REAL_MAXPAGE calls init_shared_memory().
                 * This problem arose with gcc 3.4.2 and new libs.
                 */
                SET_REAL_MAXPAGE
                ;
#endif

Why does not the call to set_maxpage in gcl_init_alloc work for you?
This would appear the logical place for this.

Separately, was there not a .rdata section issue with newer gcc?  Is
that to what you refer above as a linker mod, or do you mean the
separate pecoff files you've so wonderfully supplied?  A minimal
.rdata change to the 2.6.5 code might be a good errata candidate.

Take care.



> Cheers
> 
> Mike Thomas
> 
> 
> 
> 
> 

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