develooper Front page | perl.perl5.porters | Postings from September 2014

Re: smoker constipation?

Thread Previous | Thread Next
From:
George Greer
Date:
September 24, 2014 03:51
Subject:
Re: smoker constipation?
Message ID:
alpine.LFD.2.11.1409232344080.7700@drei.m-l.org
On Tue, 23 Sep 2014, Dave Mitchell wrote:

> Having pushed a smoking branch, smoke-me/davem/cybers_methop_stuff,
> five days ago, and only having seen one result so far, I looked at:
>
> http://m-l.org/~perl/smoke/perl/linux/smoke-me_clang_sanitize=address/status
> http://m-l.org/~perl/smoke/perl/win32/smoke-me/status
>
> which, if read them correctly, seems to imply that the linux/ASan smoker
> is busy smoking branches from 2012, while the win32 smoker has a big list
> of candidate branches, but has concluded that none of them need smoking?

The win32 smoke-me & maint smokers were killed.  I think sometimes a 
test's watchdog triggers and it process kills too much, taking out the 
smoker as well.

My current smoke-me code puts a timestamp at the top of the status pages 
so you can see if the status file is stale, but the Win32 code is a little 
behind and lacks that.

I've updated the smoke-me code to refuse to smoke anything with a commit 
date over two weeks old, to avoid problems if the state is ever corrupted 
again.  I'm also tracking state directly instead of using whether the git 
repo local branch SHA is the same as the remote branch SHA.  (Which had 
other undesireable side-effects, so good to stop doing that regardless.)

I restarted the Linux smokers that were idle and will restart the rest 
tomorrow and update the Windows smoker.

-- 
George Greer

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About