www.digitalmars.com         C & C++   DMDScript  

c++ - dmc internal error cgcod 2347

reply Laurentiu Pancescu <plaur crosswinds.net> writes:
I got a cgcod 2347 internal error during compilation of rtpacket.c, from 
SpeakFreely source (www.speakfreely.org).  The problem seems to be 
caused by:

int i, j, k;
for(i=j=k = 0; i < 100; i++)
{
	if ((k & 3) != 2  && ((i % 580) != 579)) {
	        /* if j++ is replaced with k++, error goes away */
		sb.buffer.buffer_val[j++] = 1;
	}
	k = (k + 1) % 11;
}

Unfortunately, when I make a standalone program, using similar loop, the 
  internal error won't show up!  :(((  If you want (you==Walter Bright 
:), I can put together a small archive with only the necessary files to 
compile rtpacket.c.

Regards,
   Laurentiu
Sep 29 2002
parent reply Laurentiu Pancescu <plaur crosswinds.net> writes:
I forgot to mention this happens only with "-o+all"...

Laurentiu

Laurentiu Pancescu wrote:

 I got a cgcod 2347 internal error during compilation of rtpacket.c, from 
 SpeakFreely source (www.speakfreely.org).  The problem seems to be 
 caused by:
 
 int i, j, k;
 for(i=j=k = 0; i < 100; i++)
 {
     if ((k & 3) != 2  && ((i % 580) != 579)) {
             /* if j++ is replaced with k++, error goes away */
         sb.buffer.buffer_val[j++] = 1;
     }
     k = (k + 1) % 11;
 }
 
 Unfortunately, when I make a standalone program, using similar loop, the 
  internal error won't show up!  :(((  If you want (you==Walter Bright 
 :), I can put together a small archive with only the necessary files to 
 compile rtpacket.c.
 
 Regards,
   Laurentiu
Sep 29 2002
parent reply "Walter" <walter digitalmars.com> writes:
Yes, please, I need the reproducible archive. -Walter

"Laurentiu Pancescu" <plaur crosswinds.net> wrote in message
news:3D977264.1050202 crosswinds.net...

 I forgot to mention this happens only with "-o+all"...

 Laurentiu

 Laurentiu Pancescu wrote:

 I got a cgcod 2347 internal error during compilation of rtpacket.c, from
 SpeakFreely source (www.speakfreely.org).  The problem seems to be
 caused by:

 int i, j, k;
 for(i=j=k = 0; i < 100; i++)
 {
     if ((k & 3) != 2  && ((i % 580) != 579)) {
             /* if j++ is replaced with k++, error goes away */
         sb.buffer.buffer_val[j++] = 1;
     }
     k = (k + 1) % 11;
 }

 Unfortunately, when I make a standalone program, using similar loop, the
  internal error won't show up!  :(((  If you want (you==Walter Bright
 :), I can put together a small archive with only the necessary files to
 compile rtpacket.c.

 Regards,
   Laurentiu
Sep 29 2002
parent reply Laurentiu Pancescu <plaur crosswinds.net> writes:
Okay, I attached it.  The cgcod error only appears when compiled with 
"dmc -c -o -5 rtpacket.c".  Using different processor spec is okay (I 
tried with -3, -4, -6).

Thanks for your fast answer (DigitalMars has really good support!). 
Sorry for not being able to shrink the source, I'm really too tired - 
just got back from work...  :(

Regards,
   Laurentiu


Walter wrote:

 Yes, please, I need the reproducible archive. -Walter
Sep 30 2002
next sibling parent reply Laurentiu Pancescu <plaur crosswinds.net> writes:
I could use -6 for the version I use.  Will code compiled with -6 also 
run on Pentium machines?  I mean PentiumMMX, not PentiumII or newer.

Laurentiu Pancescu wrote:

 Okay, I attached it.  The cgcod error only appears when compiled with 
 "dmc -c -o -5 rtpacket.c".  Using different processor spec is okay (I 
 tried with -3, -4, -6).
Oct 01 2002
parent "Walter" <walter digitalmars.com> writes:
Yes, it will work.

"Laurentiu Pancescu" <plaur crosswinds.net> wrote in message
news:3D99F0F7.5080804 crosswinds.net...

 I could use -6 for the version I use.  Will code compiled with -6 also
 run on Pentium machines?  I mean PentiumMMX, not PentiumII or newer.

 Laurentiu Pancescu wrote:

 Okay, I attached it.  The cgcod error only appears when compiled with
 "dmc -c -o -5 rtpacket.c".  Using different processor spec is okay (I
 tried with -3, -4, -6).
Oct 01 2002
prev sibling parent reply "Walter" <walter digitalmars.com> writes:
Ok, it's been fixed and a new beta posted. -Walter
Oct 03 2002
next sibling parent Laurentiu Pancescu <plaur crosswinds.net> writes:
I just tried, it works!  Thanks a lot!!!

Laurentiu

Walter wrote:

 Ok, it's been fixed and a new beta posted. -Walter
Oct 03 2002
prev sibling parent Mark Evans <Mark_member pathlink.com> writes:
Walter when you post a new beta please post a note with some standard subject
line and the version number.  New 8.30.3 beta say.

And then if you would, consolidate the changes/fixes in one place like the
compiler release page.

I think it's good marketing even to do this for the most minor beta releases
because it shows that the compiler is under very active development and the
author tracks down minor bugs.  For which we all thank you,

Mark


In article <angspg$369$1 digitaldaemon.com>, Walter says...

Ok, it's been fixed and a new beta posted. -Walter
Oct 05 2002