🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Thread View: comp.unix.sys5.r3
2 messages
2 total messages Started by peter da silva Fri, 06 Nov 1992 17:06
compiling SAM with non-ANSI compilers
#25
Author: peter da silva
Date: Fri, 06 Nov 1992 17:06
90 lines
2703 bytes
[If you change the followups, please leave one of the groups I've got followups
 to in the Newsgroups line, as I don't follow the other groups]

OK, I've got SAM, but it's for ANSI compilers only. My System V.3.2 and
Sun 4.x systems only have K&R compilers. I can get it all to compile
pretty cleanly with "unproto", but I get a weird memory allocation
failure in erealloc. It looks like this:

	void *erealloc(void *p, unsigned long n)
	{
		p = realloc(p, n);
		if(!p)
			fatal error code
		return p
	}

It barfed when the regular expression code was compiling a character
class. The character class code calls erealloc() with a null pointer.
No problem, just create srealloc that special-cases the null pointer:

	void *srealloc(void *p, unsigned long n)
	{
		if(p)
			p = realloc(p, n);
		else
			p = malloc(n);
		return p;
	}

Now erealloc is failing on the call to srealloc(), but the stack trace
says it's in free(). I can't figure out what's going on here. I'm using
a compiler with 32 bit ints and the standard promotion rules, so I can't
imagine that it's stack bashing.

Is anyone else so demented as to try this? Has anyone had better success?

For reference, the "ansi cc" I'm using looks like this:

:
SAMDIR=/usr1/xds/src/sam
BIN=$SAMDIR/bin

ECHO=:
CPPARGS="-E"
for i
do
  case $i in
    -I*|-D*) CPPARGS="$CPPARGS $i";;
    -v) ECHO=echo
  esac
done

ARGS=
MV=
for i
do
  case $i in
    -I*|-D*) ;;
    *.c)
      $ECHO "- /lib/cpp $CPPARGS $i"
      if /lib/cpp $CPPARGS $i |
	$BIN/unproto > _$i
      then
        ARGS="$ARGS _$i"
        o="`echo $i | sed 's/\.c$/.o/'`"
        MV="$MV mv _$o $o;"
      else exit 2
      fi
      ;;
    *)
      ARGS="$ARGS $i"
      ;;
  esac
done
$ECHO - cc $ARGS
if cc $ARGS
then
  if [ "$MV" ]
  then
    $ECHO "-$MV"
    eval $MV
  else true
  fi
else false
fi
-- 
% Peter da Silva % 77487-5012 % +1 713 274 5180 %
true(<<VV$@\\$'&O 9$O%'$LT$&$"V6"$&$<4$?'&$ #I&&?$=$<<@)24 24 scale 3 21 moveto
{dup 36 eq{pop not}{dup 7 and 4 sub exch 56 and 8 div 4 sub 2 index{rlineto}{
rmoveto}ifelse}ifelse}forall stroke pop showpage % Har du kramat din varg idag?
Re: compiling SAM with non-ANSI compilers
#29
Author: Alex Martelli
Date: Tue, 10 Nov 1992 07:29
30 lines
2069 bytes
peter@ferranti.com (peter da silva) writes:
	...
:OK, I've got SAM, but it's for ANSI compilers only. My System V.3.2 and
:Sun 4.x systems only have K&R compilers. I can get it all to compile
:pretty cleanly with "unproto", but I get a weird memory allocation
:failure in erealloc. It looks like this:
	...
:Now erealloc is failing on the call to srealloc(), but the stack trace
:says it's in free(). I can't figure out what's going on here. I'm using
:a compiler with 32 bit ints and the standard promotion rules, so I can't
:imagine that it's stack bashing.
:
:Is anyone else so demented as to try this? Has anyone had better success?

My suggestion:  get Conor Cahill's precious "dbmalloc" package and
compile and link your code with that.  It will help *mightily* in
diagnosing whatever it is that it's happening to clobber your memory
allocation!  Some pointer is probably being freed twice, or overwritten
accidentally with garbage from elsewhere, and dbmalloc will point it out
fast.  (It might be even better to use the commercial enhanced version
of dbmalloc, "sentinel" - I haven't tried that yet, but it should have
advantahes - however, that costs some $$$ - I think $195 for 386 Unix
and $395 for Sun - so since you're working on a free program I assume
you may not want to invest in that).

dbmalloc has appeared in comp.sources.? in the recent past, so it should
be VERY widely available, from any site that archives that.
-- 
Email: martelli@cadlab.sublink.org                   Phone: ++39 (51) 6130360
CAD.LAB s.p.a., v. Ronzani 7/29, Casalecchio, Italia   Fax: ++39 (51) 6130294 
Thread Navigation

This is a paginated view of messages in the thread with full content displayed inline.

Messages are displayed in chronological order, with the original post highlighted in green.

Use pagination controls to navigate through all messages in large threads.

Back to All Threads