Mail Archives: cygwin/2006/09/07/16:51:33
On Thu, Sep 07, 2006 at 01:04:49PM -0700, Chris McIntosh wrote:
>Hello, I work at a company where we have a complicated Linux build. We
>use cygwin to mimic this build on windows for some of our windows
>centric people.
>
>At different times when cygwin forks it will throw a Resource
>temporarily Unavailble error. I am tempted to call this "Random"
>because it doesn't manifest itself until a random time, but once
>manifested it will continue to fail at the exact same spot. This spot
>is different for all users but once it does happen it continues to
>fail.
>
>This is typically in our make flow when it is forking or spawning
>another process to execute a command.
>
>Here is a snip
>
>../../../../..//class/mfs/engine/disp/gf100_engine.mfs
>/usr/bin/env: ../../../../..//bin/Win32/cfp: Resource temporarily
>unavailable
>makepp: Failed to build target gf100.xml
>
>One thing to note, the command that it fails on is different for all
>users. It actually fails when forking. Another example is
>
>/usr/bin/bash: fork: Resource temporarily unavailable
>make[4]: *** [ObjWin32/.PTreeVisitor.d] Error 128
>
>After looking at the cygwin source I found that EAGAIN which produces
>this message is mapped to many windows errors. So after more
>investigation I found it to be Win32 error 234 which is
>ERROR_MORE_DATA. I then traced this down to when cygwin calls
>CreateProcessAsUser in spawn.cc
>
>So my question, has anyone seen this before. I don't think it is a
>cygwin bug since it the win api returning this error, but maybe it is
>something to do with the call.
>
>Any pointers or tips would be helpful.
A google search for CreateProcess and ERROR_MORE_DATA shows a patch for
make which avoids putting empty environment variables in the environment
block sent to CreateProcess.
So, I've added a check for that to cygwin and generated a new snapshot at:
http://cygwin.com/snapshots/
Be advised that the snapshot version of Cygwin has been built with
debugging turned on, which means that it will be slower and noisier than
a normal cygwin release. In particular, there is a false positive but
inconsequential report about handle closing which can be generated
fairly easily.
cgf
--
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/
- Raw text -