Re: Strange segfault with improper usage --> plus extra fix for internal_trace issue in there

On Wed, Mar 26, 2008 at 06:02:08PM +0100, Ger Hobbelt wrote:
> 
> BTW: bugreport: the first internal_trace dump loop in there has an
> indexing bug. diff attached: apply to your crmregex_tre.c as you have
> it. ;-)

oops, yes, reducing the 2 loops to 1 for final diff I left in the for() on
j instead of i - just do j->i on that line:

 -    j = 0;
-+    i=j=0;
 +    if (internal_trace)
 +    {
 +      fprintf (stderr, "Checking the regex cache for '%s'\n", regex);
-+      for (j=0; j < CRM_REGEX_CACHESIZE; j++) {
++      for (i=0; i < CRM_REGEX_CACHESIZE; j++) {
 +       if (regex_cache[i].regex_len > 0) fprintf (stderr,
 +         "  regex[%d]:'%s' cf=%#x(%#x) len=%d(%d)\n",
 +         i,regex_cache[i].regex,regex_cache[i].cflags,cflags,
 +         regex_cache[i].regex_len,regex_len);
 +      }
  #ifdef REGEX_CACHE_LINEAR_SEARCH
 +      fputs (" [<- rex cache linear]\n", stderr);
 +    }
++    i=j=0;

> - +#error "Must have #define'd ONE of these:
> REGEX_CACHE_RANDOM_ACCESS, REGEX_CACHE_LINEAR_SEARCH"

oh ... #ifdef...RANDOM is redundant, i left it in in case some other algo will
be devised - but so far we could really drop it: deafault is RANDOM, unless
LINEAR is defined.

> and some additional cruft:

I'm fine with them


thanks
-- 
 paolo
 
 GPG/PGP id:0x3A47DE45  - B5F9 AAA0 44BD 2B63 81E0  971F C6C0 0B87 3A47 DE45
 - 9/11: the outrageous deception and ongoing coverup: http://911review.org -

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace