X-Recipient: archive-cygwin@delorie.com X-Spam-Check-By: sourceware.org From: "Dave Korn" To: References: <47F1B87A.3050709@gmail.com> <47F538F3.CAFABB19@dessent.net> <47F71B44.8030404@gmail.com> <47F72A42.C56BC03F@dessent.net> <47F7F17E.1090202@gmail.com> <003901c89972$d08f41b0$2708a8c0@CAM.ARTIMI.COM> <47FBA00A.2090201@gmail.com> Subject: RE: Compiling Cygwin on Linux Date: Tue, 8 Apr 2008 19:47:26 +0100 Message-ID: <00c801c899a8$fd7dba90$2708a8c0@CAM.ARTIMI.COM> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Mailer: Microsoft Office Outlook 11 In-Reply-To: <47FBA00A.2090201@gmail.com> Mailing-List: contact cygwin-help@cygwin.com; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner@cygwin.com Mail-Followup-To: cygwin@cygwin.com Delivered-To: mailing list cygwin@cygwin.com Robert Eckhoff wrote on 08 April 2008 17:41: > Below are the errors I received before I made their > associated changes. > Dave Korn wrote: >> Robert Eckhoff wrote on 05 April 2008 22:39: >> >> >>> directories. Winsup also had some source problems that I corrected. >>> >>> ~/cygwin-1.5.25-11/winsup/cygwin/winsup.h:276 >>> -extern bool wsock_started; >>> +extern "C" bool wsock_started; >>> >>> > Before I made the change, Error: > /home/rpeckhoff/cygwin-1.5.25-11/winsup/cygwin/winsup.h:276: error: > previous declaration of 'bool wsock_started' with 'C++' linkage > /home/rpeckhoff/cygwin-1.5.25-11/winsup/cygwin/autoload.cc:246 >> error: > conflicts with new declaration with 'C' linkage How bizarre. You said you were using a series 4 GCC, didn't you? This could be a bug or oversight in the code that generates the error - maybe it should be downgraded to a warning. Or maybe there's something coming up in C++0x that applies linkage to variables. I'll have to raise it on the gcc list, thanks for the details. >>> ~/cygwin-1.5.25-11/winsup/cygwin/syscall.cc:3238 >>> Uses goto's which breaks the build with gcc4.3. I rewrote the function >>> to use if statements. >>> >> >> What kind of problem does gcc4.3 have with gotos? There's nothing wrong >> with them in general and lots of other functions in that file have them, >> so maybe there's some underlying problem which is the actual thing that >> needs fixing? > /home/rpeckhoff/cygwin-1.5.25-11/winsup/cygwin/syscalls.cc: > In function > 'FILE* popen(const char*, const char*)': > /home/rpeckhoff/cygwin-1.5.25-11/winsup/cygwin/syscalls.cc:323 8: error: > jump to label 'err' > /home/rpeckhoff/cygwin-1.5.25-11/winsup/cygwin/syscalls.cc:3230: error: > from here /home/rpeckhoff/cygwin-1.5.25-11/winsup/cygwin/syscalls.cc:3233: > error: crosses initialization of 'fhandler_pipe* fh' > /home/rpeckhoff/cygwin-1.5.25-11/winsup/cygwin/syscalls.cc:323 8: error: > jump to label 'err' > /home/rpeckhoff/cygwin-1.5.25-11/winsup/cygwin/syscalls.cc:3210: error: > from here /home/rpeckhoff/cygwin-1.5.25-11/winsup/cygwin/syscalls.cc:3233: > error: crosses initialization of 'fhandler_pipe* fh' Ah, yes, that is potentially quite a nasty thing to happen. I wonder if there's a code path where it could end up using the uninitialised 'fh'? I'll take a look. BTW, you're stepping into unknown territory trying to build the DLL with such a recent GCC. The distro compiler is still on 3.4.4 for a number of reasons (see archives passim ad nauseam), so you're probably the first to discover these issues; thanks for the report! cheers, DaveK -- Can't think of a witty .sigline today.... -- 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/