Borut,

If this is negligence then I belong on the gallows. A problem that get reported and
and speedily corrected fits my definition of a working model.

Congradulations!!

George


----- Original Message -----
From: "Borut Razem" <borut.razem@siol.net>
To: sdcc-user@lists.sourceforge.net
Sent: Monday, May 4, 2009 1:05:46 PM GMT -05:00 US/Canada Eastern
Subject: Re: [Sdcc-user] I am getting nowhere :(

The initialization of allocated memory was missing. I hope now it is
fixed in svn revision #5448.
Kevin please try the tomorrow's snapshot build and let me know if it
works for you.

Kevin, Raphael and Jesus, thank you for help and I apologize again for
my negligence.

Borut


Jesus Calvino-Fraga wrote:
> Hi Borut,
>
> I am getting an access violation at line 94 of cpp/libcpp/line-map.c:
>
> const struct line_map *
> linemap_add (struct line_maps *set, enum lc_reason reason,
>               unsigned int sysp, const char *to_file, linenum_type to_line)
> {
>    struct line_map *map;
>    source_location start_location = set->highest_location + 1;
>
>    if (set->used && start_location < set->maps[set->used -
> 1].start_location) <-- HERE
>      abort ();
> .
> .
> .
>
> In debug mode using VC6, set->used=0xCDCDCDCD.  That means that the
> variable was never initialized.  This is probably a C runtime library
> issue on Windows when allocating memory. Same goes for set->highest_location.
>
> Jesus
>
> At 10:06 PM 03/05/2009, you wrote:
>
>  
>> Raphael,
>>
>> sorry, I misunderstood where the problem is. So it is my fault!
>>
>> Is this a general problem or it happens only with a specific source code?
>>
>> It is strange that it seems everything OK on other platforms -
>> regression test pass. And everything is OK if I compile sdcc on my win32
>> cygming box too. I didn't try to run tests on the snapshot package...
>>
>> Borut
>>
>>
>> Raphael Neider wrote:
>>    
>>> Dear Borut,
>>>
>>>
>>>      
>>>>> I suppose that the problem was introduced in revision 5439 with
>>>>> "src/pic16/gen.c (genNearPointerGet): fix double indirect access to
>>>>> bitfields (#2782405)", so I can probably remove build 5432. Please
>>>>> confirm if I can remove it.
>>>>>
>>>>>          
>>>> I meant to remove revision 5436, not 5432...
>>>>
>>>>        
>>> That's a definite no! Snapshots r5436 and r5441 work fine,
>>> r5444 breaks in sdcpp (the preprocessor): Its output is empty!!!
>>> Try "sdcpp test.c" with r5441 and compare it to r5444.
>>>
>>> As I see it, the upgrade to gcc cpp 4.4.0 broke things.
>>> This can also be concluded from Kevins initial report,
>>> though on my WinXP-box, windows does not complain about
>>> sdcpp.exe encountering problems as with Kevin. Strange ...
>>>
>>> BTW: Kevin compiled for the default target, which is the 8051.
>>> The pic16 backend is not involved at all (probably no backend is:
>>> sdcc should error out in the frontend due to no tokens seen on end
>>> of file).
>>>
>>>
>>>      
>>>>> Done, snapshots stored in (sdcc-snapshot-20090427) folder at the end of
>>>>> http://sdcc.sourceforge.net/snap.php page.
>>>>>
>>>>>          
>>> Thanks a lot.
>>>
>>>
>>>      
>>>>> I copied both 5436 and 5438 revisions.
>>>>> I suppose that the problem was introduced in revision 5439 with
>>>>> "src/pic16/gen.c (genNearPointerGet): fix double indirect access to
>>>>> bitfields (#2782405)", so I can probably remove build 5432. Please
>>>>> confirm if I can remove it.
>>>>>
>>>>> I thing that it would be better solution to revert the fix of
>>>>> #2782405 until the proper solution is found.
>>>>>
>>>>>          
>>> Nope, the proper solution is to fix the MinGW build of GCC CPP 4.4.0.
>>> As a temporary fix, I would propose to revert r5444.
>>>
>>> I do not want to just put the blame on you -- I really appreciate
>>> your effort in actively maintaining the preprocessor branch. This
>>> time, however, some "interesting" problems seem to have crept in ...
>>>
>>> Best regards,
>>> Raphael
>>>      

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations
Conference from O'Reilly Media. Velocity features a full day of
expert-led, hands-on workshops and two days of sessions from industry
leaders in dedicated Performance & Operations tracks. Use code vel09scf
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Sdcc-user mailing list
Sdcc-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sdcc-user