This is the mail archive of the
mailing list for the Cygwin project.
RE: compile problems on cygwin
- From: "Dave Korn" <dave dot korn at artimi dot com>
- To: <cygwin at cygwin dot com>
- Date: Sat, 17 May 2008 22:07:03 +0100
- Subject: RE: compile problems on cygwin
- References: <email@example.com> <200805161735.m4GHZSHi005168@pippin.first.lan> <alpine.LRH.firstname.lastname@example.org>
R P Herrold wrote on 16 May 2008 20:44:
> The tarball at the perma-link:
> I get (in ./src. on an unpacked tarball):
> $ g++ rph_test.c
> In file included from glossary.h:48,
> from rph-test.c:1:
> functional.h:213: error: expected init-declarator before '&' token
> functional.h:213: error: expected `,' or `;' before '&' token
> And so something in the Cygwin macro def's is not
> #undef'ing after itself. I assume the Cygwin folks know how
> they want a bug report and to track this down?
It would have been handier if you had showed us what was around line 213
of functional.h, or made available preprocessed sources from the failed
compilation rather than a 5.6 meg tarball. "Minimal testcases", we call
> I am used to seeing Corinna Vinschen <vinschen>
Please avoid quoting email addresses in body text, they get harvested by
> post update release notices on the cygwin-announcement mailing
> list. In his announcements, he
> requests reports to:
> So I have sought to subscribe there, and file this there, per
> the request of those announcemnts.
I downloaded your tarball.
It does not contain any "rph-test.c". Still, I could reproduce the
problem with just a single line #include "glossary.h". Line 213 of
213 typedef const sockaddr & sa_ref;
After pre-processing, this comes out as
typedef const sockaddr & sa_ref;
Therefore there is no #define involved.
You forgot to #include the definition of sockaddr. It lives in
If you had done even the most trivial investigation into this bug, you
would have found this out. Perhaps you've encountered the same error
message before in a case where it actually /was/ caused by a #define, and
you jumped to a false conclusion that the same thing was happening again?
That's not a valid step of reasoning, I'm afraid: there's more than one
thing that can cause any given error message, particularly where it
indicates some kind of generalised syntax or parsing error, as is the case
Can't think of a witty .sigline today....
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html