I was looking at this, and with the regex versions of these I read
that you still need to free the memory created by the regcomp. Am I
correct? So shouldn't the following line be added to Browser.C
#if defined(HAVE_REGCOMP) && defined(HAVE_REGEX_H)
regfree(&search_for);
#elif (defined(HAVE_REGCOMP) && (defined(HAVE_REGEX_H) ||
defined(HAVE_REGEXP_H))) || defined(HAVE_REGCMP)
David
Pulled CVS last night and had only one compile error on IRIX and
FreeBSD (both the same). In the regex patches to Browser.C, the conditions
on the free() of the compiled regex were changed. My conditions weren't
general enough; the current are too general as the free() is now compiling
in for platforms where the POSIX regex_t is used; it's only applicable to
the char * versions. The attached patch should fix it.
|If I remember right, CVS write-access for Randall foundered on a firewall
|issue on his end. I'll look into it.
Thanks. At the time, opendx.watson's sshd was terminating the negotiation
for reasons Peter and I never did figure out (ssh worked/works fine for me
to other machines). I've switched to OpenSSH since then, so we can give it
another shot.
Randy
--
Randall Hopper (mailto:[EMAIL PROTECTED])
Lockheed Martin Operation Support
EPA Scientific Visualization Center
US EPA MD/24 ERC-1A; RTP, NC 27711
Attachment converted: Powder1:Browser.C.patch (TEXT/R*ch) (00033248)
--
.............................................................................
David L. Thompson The University of Montana
mailto:[EMAIL PROTECTED] Computer Science Department
http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812
Work Phone : (406)257-8530
