Hi Matthew,
many thanks again for your investigation.
On Sat, Apr 18, 2020 at 01:15:49PM -0700, Matthew Fernandez wrote:
> > Upstream is in the row of this investigation. Its quite interesting
> > that the issue could also observed on amd64. So probably this is a real
> > issue which is just exposed on mipsel. I think just bumping the array
> > boundaries is cheap. If there will be no response from upstream (or
> > somebody else who is comfortable with the algorithm which I'm not) I'll
> > check again on mipsel and in case it will work I'll upload.
>
> Fair enough. While we’re discussing this, here’s another patch for a memory
> leak that fixes a typoed ifdef and a missing free():
>
> diff --git a/src/squid/clustal.c b/src/squid/clustal.c
> index 650004a..bb1fec8 100644
> --- a/src/squid/clustal.c
> +++ b/src/squid/clustal.c
> @@ -207,7 +207,7 @@ WriteClustal(FILE *fp, MSA *msa)
> int namelen; /* maximum name length used */
> int pos; /* position counter */
> #ifdef CLUSTALO
> - char *buf; /* buffer for writing seq */
> + char *buf = NULL; /* buffer for writing seq
> */
> int cpl = msa->alen<iWrap ? msa->alen+10 : (iWrap > 0 ? iWrap :
> 60); /* char per line (< 64) */
> #else
> char buf[80]; /* buffer for writing seq */
> @@ -410,8 +410,9 @@ WriteClustal(FILE *fp, MSA *msa)
> #endif
> }
>
> -#ifdef CLUSTAL
> +#ifdef CLUSTALO
> free(piResCnt); piResCnt = NULL;
> + free(buf); buf = NULL;
> #endif
>
> return;
I tried the suggested patches again step by step on mipsel but the bus
error issue remains. :-(
Kind regards
Andreas.
--
http://fam-tille.de
