I compromised: see https://svn.open-mpi.org/trac/ompi/changeset/25432.
If someone has strong feelings, go ahead and change it. This is a rare enough
issue that I don't think we need to spend much more time on it.
On Nov 2, 2011, at 2:22 PM, Barrett, Brian W wrote:
> To be honest, I don't care
It was too complex, Tim. Nobody wound up using it, and so some of us have found
a simpler alternative that seems to work, but hasn't been fully implemented
across the code yet (we're just doing it as we go).
And nothing solves the problem of errors from every proc when you direct launch.
On No
Brian,
I thought the OPAL_SOS stuff was supposed to be the way to fix this?
https://svn.open-mpi.org/trac/ompi/wiki/ErrorMessages
Has that effort faded or not worked?
On Wed, Nov 2, 2011 at 2:22 PM, Barrett, Brian W wrote:
> To be honest, I don't care so much today, I'm just fighting so that the
To be honest, I don't care so much today, I'm just fighting so that the
output doesn't get worse. At some point, we do need to figure out a
better way of dealing with error messages, but not today :).
Brian
On 11/2/11 11:53 AM, "Ralph Castain" wrote:
>Hmmmsince it was my bug that surfaced
Hmmmsince it was my bug that surfaced the problem, maybe the best answer is
to just return an error code. I'll slowly work thru the param registrations in
ORTE and make them all check the return code. I'm willing to look at OPAL as I
go, but someone else will have to deal with the OMPI layer
I really don't like our show_help at every level behavior (look at what
happens when MPI_INIT fails, you get a page per process of the same error
message from each level of the call stack). If you want to show_help and
abort on debug, that makes sense. It doesn't make any sense on a
production bu
Brian: you were the one that had an allergic reaction to #1 on the call.
Thoughts?
On Nov 2, 2011, at 1:23 PM, George Bosilca wrote:
> As it has been said, this is not something supposed to make it in a release.
> On the unfortunate case where it does, always having a show_help will ensure
>
As it has been said, this is not something supposed to make it in a release. On
the unfortunate case where it does, always having a show_help will ensure a
quick complaint on one of our mailing lists and increase the probability of a
[very] quick fix.
george.
On Nov 2, 2011, at 06:26 , TERRY
On 11/1/2011 7:48 PM, Jeff Squyres wrote:
So this was slightly different than the opinion that was discussed on the call
today, which was 2. The rationale for #2 was to punish developers, but if such
a bug did make it through to production, users wouldn't be annoyed with
show_help messages
On Nov 1, 2011, at 5:48 PM, Jeff Squyres wrote:
> So this was slightly different than the opinion that was discussed on the
> call today, which was 2. The rationale for #2 was to punish developers, but
> if such a bug did make it through to production, users wouldn't be annoyed
> with show_he
So this was slightly different than the opinion that was discussed on the call
today, which was 2. The rationale for #2 was to punish developers, but if such
a bug did make it through to production, users wouldn't be annoyed with
show_help messages all the time.
Does anyone have strong opinion
1
george.
On Nov 1, 2011, at 17:23 , Jeff Squyres wrote:
> Can you clarify -- I can parse your text multiple ways. Which are you voting
> for?
>
> 1. show_help + return error code in all cases.
> 2. if OPAL_ENABLE_DEBUG, show_help + exit(1), else silently return error code.
> 3. show_help.
Can you clarify -- I can parse your text multiple ways. Which are you voting
for?
1. show_help + return error code in all cases.
2. if OPAL_ENABLE_DEBUG, show_help + exit(1), else silently return error code.
3. show_help. if OPAL_ENABLE_DEBUG, exit(1), else return error code.
On Nov 1, 2011,
This is a much saner solution. We [mostly] stayed away from calling exit deep
into our libraries, there is no reason to add it now. I'll vote in favor of
show_help + return code.
george.
On Nov 1, 2011, at 15:14 , Jeff Squyres wrote:
> We talked about this on the call today.
>
> A good sugg
We talked about this on the call today.
A good suggestion was made: call show_help/opal_finalize/exit only when
OPAL_ENABLE_DEBUG is true. Otherwise, return an error code.
If no one objects to this, I'll commit this tomorrow.
On Oct 31, 2011, at 4:16 PM, Jeff Squyres wrote:
> WHAT: what to
15 matches
Mail list logo
