[boost] Re: Regression results for Intel 8 online

From: Michael Glassford (glassfordm_at_[hidden])
Date: 2004-07-20 15:38:40

Martin Wille wrote:
> Michael Glassford wrote:

[snip report of failures in various libraries]

>> The read_write_mutex is new, so it's not a true regression from the 
>> previous 100% pass rate. I'm currently working on the scheduling 
>> algorithm to make it pass its unit tests and hopefully deal with 
>> reported deadlocking issues.
> 
> 
> Please, fix that quickly. 

Believe me, I've been trying.

 > Someone fixed the cause of the compile errors
> for the more conforming compilers. 

I fixed many compile errors, and had help with a couple of others.

 > So, now, I can expect the tests to
> deadlock for every compiler. This is a showstopper for testing, IMHO.
> I'm considering to remove the test from the Jamfile until the problem
> has been fixed.

I've just checked in the changes that I've been working on. I can't 
guarantee that they fix the deadlocks until they're tested on the 
platforms that were experiencing problems, but I hope they do. I'll keep 
an eye on the regression testing reports, but please let me know one way 
or the other as soon as you find out anything.

Thanks,

Mike