This is the mail archive of the
cygwin
mailing list for the Cygwin project.
Re: pcre, C++
Dave said:
> This is most likely the passing-empty-strings-across-dll-boundaries
> problem aka PR24196; would you care to install the experimental
> gcc-3.4.4-2 and see if it fixes the problem?
Why, yes, it does; thank you. (I shoulda remembered that...)
Unfortunately, this is not a solution for me (yet) because:
RE: Problems with GCC-3.4.4-2 (exp) ?
http://www.cygwin.com/ml/cygwin/2006-11/msg00512.html
(debugging symbols/specs file?)
and
"RE: Status of gcc 3.4.4-2"
http://www.cygwin.com/ml/cygwin/2006-09/msg00269.html
(specifically: "...but I'll have to roll a fresh package to fix a couple
of bugs
that showed up")
and
"RE: Requesting an updated gcc-mingw-g++ package to match
gcc-g++-3.4.4-2 (PR24196)"
http://www.cygwin.com/ml/cygwin/2006-10/msg00220.html
(OT: it appears that the mingw project does not have ANY releases that
incorporate this fix; looks like I'll have to roll my own for THAT
environment.)
--
Chuck
P.S. It's my understanding that this change (to std::string) would
constitute an ABI change of the underlying C++/STL runtime library on
cygwin. Thus, any C++ packages -- or at least, those that are NOT
compiled with -nostdlib -- would have to be recompiled (and version
bumped?) once a g++ with this change was promoted to curr: Right?
Fortunately, there aren't many of these: libncurses++, xerces, groff,
and the various octave stuff is all that comes to mind -- and octave
already requires binary clients to use g++-3.4.4-2 IIRC. Oh, and pcre's
libpcrepp. <g>
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/