Re: [boost] Re: [regression] Discrepancy with display and results...

From: Aleksey Gurtovoy (agurtovoy_at_[hidden])
Date: 2004-08-03 04:39:18

Rene Rivera writes:
> Aleksey Gurtovoy wrote:
>
>> Oh, OK. Those are uploaded as a gzipped tar and then unpacked on the
>> SF. May be the job got terminated. In general, due to issues like
>> this, the Boost-wide results are way more trustworthy.
>
> In that case I would suggest to dispense with the XSLT reports in SF
> entirely. Instead just upload the generic report to SF, and upload the
> XSLT reports to your server for the Boost-wide reports.

That was the plan. Just didn't get there yet.

>
> The current situation is just confusing. And since the reports are not
> reliable it invalidates people looking at the latest results. 

I was hoping that by now everybody goes straight to the Boost-wide
reports. Any reason not to (besides a habit :)?

> You are basically wasting time on the second set of test you are
> doing if people can't use them.

It's *the same* set of tests, but you are right, there is no reason to
generate these reports now that the Boost-wide ones have stabilized
and are actually more reliable.

>
> And not uploading the large set has the benefit of saving disk space
> on SF ;-)

Sure.

-- 
Aleksey Gurtovoy
MetaCommunications Engineering