Open MPI logo

Open MPI Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Development mailing list

Subject: [OMPI devel] Fwd: OMPI / Coverity: not running?
From: Jeff Squyres (jsquyres_at_[hidden])
Date: 2008-06-16 17:49:47


FYI -- in case you hadn't already noticed, the Coverity results are a
bit out of date (last run was 5/28/08). It looks like they're going
through some growing pains -- see my mail to David below, and his
reply below that.

Begin forwarded message:

> From: David Maxwell <dmaxwell_at_[hidden]>
> Date: June 10, 2008 10:17:02 AM EDT
> To: Jeff Squyres <jsquyres_at_[hidden]>
> Subject: Re: OMPI / Coverity: not running?
>
> On Fri, Jun 06, 2008 at 08:16:31AM -0400, Jeff Squyres wrote:
>> David --
>>
>> It looks like the last run of OMPI on scan2 was 5/28 (run ID 145).
>> It also
>> looks like the same SVN r number tarball was used for all the runs
>> since
>> 5/18 (r18449). Our SVN is currently up to r18604. Is something
>> wrong?
>>
>> Some other random observations:
>>
>> - The number of files decreased dramatically between run IDs 141 and
>> 142 (5/18 and 5/21) -- you've mentioned this problem before.
>> - The LOC counts between runs varies wildly, even when the number of
>> files is more-or-less constant. Max LOC I see is somewhere around
>> 430k; min LOC is somewhere around 100k.
>> - The number of files and LOC vary wildly, even when the same SVN r
>> number was used (e.g., the 5/18 - 5/28 runs).
>>
>> Can you look into this? As stats, the # files/LOC don't really
>> matter to
>> us -- what matters is that it actually scans all the files that it
>> can.
>>
>> Additionally, can you tell us what extra libraries you installed to
>> configure against OMPI? Remember that OMPI conditionally compiles
>> a bunch
>> of its plugins, depending on what back-end library/software support
>> exists
>> on the machine where ./configure is run. Additionally, we might be
>> able to
>> tune up your OMPI ./configure command to get greater coverage of the
>> overall code base.
>>
>> Thanks!
>>
> Hello. This is a semi-automated reply (form-letter). I am not able
> to reply to your message right now. I am receiving email requests
> much more quickly than I can process them.
>
> My current projects include:
>
> Configuration of new servers for the Scan website and build hosts -
> The existing database server is almost out of disk space. I can not
> add additional projects, or new runs of existing projects.
> Migration to
> the new database server is my top priority. Migration to the new
> build
> servers is a secondary priority. Please be patient.
>
> Configuration of a ticketing system to track outstanding requests -
> While I have tried to live off an inbox filing-system, this is no
> longer
> possible. I need to provide a way for people to query the status of
> their request without my involvement. You will receive a ticket
> number
> for your request as soon as testing of the new system is complete.
>
> Completion of the automated build-request system -
> For those of you on the self-build trial, the new build and database
> servers are the blocking point on finishing the automation of this
> system. After that, you will be able to submit builds without
> requiring
> hands-on time from myself or Erinn.
>
> Once the immediate project backlog is handled, we will process
> requests
> as quickly as possible. While I'm likely to send you this queue
> message
> in order from the most recent end of my mailbox, tickets will preserve
> the Date of your earliest email.

-- 
Jeff Squyres
Cisco Systems