Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Unsubscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT sourceware DOT cygnus DOT com Delivered-To: mailing list cygwin AT sourceware DOT cygnus DOT com Message-ID: <37B8A12E.48B6@worldnet.att.net> Date: Mon, 16 Aug 1999 19:39:26 -0400 From: Paul McFerrin Reply-To: paulmcferrin AT worldnet DOT att DOT net X-Mailer: Mozilla 3.04Gold (Win95; I) MIME-Version: 1.0 To: cygwin AT sourceware DOT cygnus DOT com Subject: Re: Serious I/O bugs in B20 release References: <37B7AA91 DOT 1549 AT worldnet DOT att DOT net> <19990816123117 DOT B870 AT cygnus DOT com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Thank you for your reply. I have subscribed to the mailing list. FYI - cat and my cpio DOES it's open with O_BINARY so the reads are in binary mode. Thus control-Z should not be my problem. -paul mcferrin Chris Faylor wrote: > > As you're probably aware, cygwin is essentially a free software project. > Like most free software projects, it has a mailing list where you can > report problems like this and get help from the community of people > using the product. > > The mailing list is cygwin AT sourceware DOT cygnus DOT com. You can subscribe > to it by sending email to cygwin-subscribe AT sourceware DOT cygnus DOT com. > > Your problems are probably due to the fact that your file contains > a CTRL-Z character, which is interpreted as an EOF. You should investigate > the mount command which has a '-b' option which will force all reads > and writes to be in 'binary' mode. > > You should also check out the various other resources available at > our web page: http://sourceware.cygnus.com/cygwin/ > > I've redirected followups to the mailing list. I would appreciate it > if you would send any further comments or bug reports there. > > -chris > > On Mon, Aug 16, 1999 at 02:07:13AM -0400, Paul McFerrin wrote: > >Folks: > > > >What is the mechanism for reporting bugs in CYGWIN? I have a read(2) > >bug to report when reading files with "holes" in allocated blocks. For > >examplem I can do the following: > > > > $ cat file_A >file_B > > > >where file_A contains over 1MB of data amd file_B will end up having > >less than 800 bytes! This is with the distributed "cat". > > > >I've ported Free-BSD cpio command and I can't read the cpio archive > >created without getting a premature EOF because of the holes in the > >file. The cpio and "cat" command are doing simple read(fd, buff, len) > >and are returning a "0" when it encounters a hole. > > > >The "cp" command copies these files just fine because of the code > >present to handle holes. > > > >I'm including a sample file (zz.gz) that will demostrate the problem. > >Just do: > > > > $ gunzip zz > > $ cat zz >yy > > $ ls -l zz yy > > > >I've also discovered that when using "cat" the output file will be > >BIGGER! Just do: > > > > $ cd /tmp > > $ cat /etc/termcap >term > > $ ls -l /etc/termcap term > > $ cmp -l /etc/termcap term > > > >The files will not be identical. > > > >BTW: I've seen this type of read problems with Delorie's djgpp > >package. Did he write some of this code??? :+) > > > >If you are not the correct one to report "bugs" to, please ley me know. -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com