Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm
Sender: cygwin-owner AT sources DOT redhat DOT com
Message-Id: <200107271536.f6RFarQ28343@mailsrv1.sanger.ac.uk>
Date: Fri, 27 Jul 2001 17:56:11 +0100
From: srk AT sanger DOT ac DOT uk
To: cygwin AT cygwin DOT com
Subject: Strange bug in C pre-processor - cygwin 1.0
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII

I've just encountered a very strange bug in the C pre-processor in cygwin 1.0 (the
cdrom version.)

The bug manifests itself as the pre-processor just ignoring a subset of the
#include statements in the source file. Exactly which #includes will be ignored
is variable - two successive compiles of the same source file will cause different
include files to be ignored. Obviously, dropping random include files leads to
very odd diagnostics from the compiler but it's easy to see what is happening using the
-E flag to just run the precprocessor. That is enough to show the bug.

What seems to provoke this bug is a very long (>100 lines) single C comment at the
start of the file. The presence of that is what distinguished the broken source file from
all the others which compile OK. When I split the comment into three, the problem
went away.

Environment:

cygwin 1.0
NT 4.0 SP3

I'm happy to send someone enough stuff to re-produce this if needed. (The code
I'm trying to compile is all GPL).

Simon Kelley.


--
-------------------------------------------------------------------------------
Simon Kelley.  srk AT sanger DOT ac DOT uk  simon AT thekelleys DOT org DOT uk  

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/