Mutex Timout with Linux, v3.16

Post by **jclausius** » Mon Feb 06, 2006 4:14 pm

Are *multiple* Vault CLCs running at the same time when one of them is sent the SIGINT?

If so, that would describe the conditions of the bug Dick mentioned.

Post by **jclausius** » Tue Feb 07, 2006 10:57 am

I've heard back from Ximian's Dick Porter:

This was in reference to the 60 second timeout

Dick Porter wrote:This is the bug I noticed when I was composing the original response.

I next asked about a way to release a named mutex.

Dick Porter wrote:There's a tool in the source tree that's not installed in the binary packages. Look in mono/mono/handles at hps; when run with an argument it clears out all old shared handles.

The locked named mutex will be one of them if more than 60 seconds have elapsed since a process "pinged" it.

As for the bug itself:

Dick Porter wrote:Fixing the bug is next on my list after the current task.

Post by **jclausius** » Tue Mar 14, 2006 11:10 am

A quick update from Ximian's Dick Porter:

Dick Porter wrote:I've just committed a fix for this on both head and 1.1.13 branches. It will be included next time a 1.1.13 release is made.

- Dick