httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Trawick <>
Subject Re: cvs commit: apache-2.0/src/lib/apr/strings apr_snprintf.c
Date Wed, 13 Sep 2000 23:12:37 GMT
Jim Jagielski <> writes:

> Is that always true? That for every platform widest_int is apr_int64_t?
> Does APR's configure test that long long is supported?

apr_int64_t is always defined to *something* (see, so it
seemed pretty safe.

> We use it elsewhere, so I guess it's pretty safe.
> Good catch... we don't check APR_LONGEST_LONG anymore in
> 2.0, so it was always being defined as long.

Hey, it was an easy catch.  Greg S. changed the chunk filter to use
%qx recently... today I went to update an OS/390 patch which gets the
chunk header bucket translated from EBCDIC->ASCII, but in testing it I
noticed that the chunk len was zero.  It turned out that the
apr_snprintf() helper routine was only grabbing a sizeof(long)-byte
var arg instead of a sizeof(long long) one.
Jeff Trawick | | PGP public key at web site:
          Born in Roswell... married an alien...

View raw message