Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT cygwin DOT com Mail-Followup-To: cygwin AT cygwin DOT com Delivered-To: mailing list cygwin AT cygwin DOT com In-Reply-To: <20050408205745.GQ1471@cygbert.vinschen.de> References: <20050401121143 DOT GD1471 AT cygbert DOT vinschen DOT de> <424D5C64 DOT 5050706 AT smousseland DOT com> <20050401160749 DOT GH1471 AT cygbert DOT vinschen DOT de> <4250E410 DOT 1060708 AT smousseland DOT com> <20050404103824 DOT GS1471 AT cygbert DOT vinschen DOT de> <9305417ca9a2f2601ad20b83801d2f92 AT smousseland DOT com> <20050404142507 DOT GT1471 AT cygbert DOT vinschen DOT de> <5c59a962471bd65c3c3f3f44c37b8bf3 AT smousseland DOT com> <20050406111752 DOT GG1471 AT cygbert DOT vinschen DOT de> <20050408205745 DOT GQ1471 AT cygbert DOT vinschen DOT de> Mime-Version: 1.0 (Apple Message framework v619.2) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <86032b7972f91acc07b53ed43999a665@smousseland.com> Content-Transfer-Encoding: 7bit Reply-To: cygwin AT cygwin DOT com From: kraken+spam AT smousseland DOT com Subject: Re: ipc, sockets and windows sp2 Date: Sat, 9 Apr 2005 10:55:17 +0200 To: cygwin AT cygwin DOT com X-IsSubscribed: yes > >> it works for me too for the testcase i provided last time. >> >> But there is still some issues when you run several semaphore-using >> program at the same time. > > Again the thread synchronization was the culprit. I reworked it once > again and I ran your testcase 5 times concurrently with different sleep > values, some in the same, some in different directories. It works for > me again. Please test and tell me that it solves all your problems > and nothing's left, ok? It solves all my problems and nothing's left ;) There is a strange stuff, but it won't perturb me : when i run my master/slaves programs, then run the testcase at the same time and kill it, the slave may have a socket read error if it was reading on the socket at the same time that testcase is launched. So i have to tune slave so i don't quit on socket read, but retry to read on the socket, and it's fine, as the socket read will fail only when the testcase is launched. Anyway, reproducing this would imply a server-and-client testcase, exchanging data on their socket, and using IPC, like the testcase I provided, and it depends on the children number. Just note it. Tuned like this, I made a test with a 1000 frames task, a local and remote slave, everything was ok, even when I ran two testcases at the same time I ran master and slave, so I'm verry happy, and I thank you so much for the great work you did. Kraken -- 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/