io_rewiring branch needs testing

[Will Coleda]

On Sun, Jun 7, 2009 at 9:16 PM, Andrew Whitworth<wknight8111 at gmail.com> wrote:
> On Sun, Jun 7, 2009 at 8:25 PM, Will Coleda<will at coleda.com> wrote:
>>> I merged trunk into branch so we could try to get benefit from some of
>>> the improvements that have been made to trunk since the branch
>>> started. There are some packfile-related failures because we bumped
>>> PBC_COMPAT and we're fixing that now. Could you do me a favor and run
>>> the tests again with trunk and branch at r39444, which is when I
>>> merged? that should show a pretty even footing.
>>
>> trunk @39444 == trunk @39443
>
> Sorry, I should have been more clear. I merged trunk INTO the branch,
> not the other way around.

Yes, I know; just commenting on your request to test trunk @39444 when
I had already tested it at an equivalent revision.

> iorewiring @39444 != io_rewiring at 39443 != trunk @39444
>
>> partcl isn't using any custom IO (yet). Everything is, I think, just
>> straight opcodes.
>
> Striaght opcode on vanilla IO PMCs?

Yes. I'm not explicitly creating ANY pmcs for IO, let alone custom ones.

>> Branch now completes. 'prove' results:
>>
>> Files=70, Tests=1257, 215 wallclock secs ( 0.30 usr  0.37 sys + 195.62
>> cusr  9.39 csys = 205.68 CPU)
>> Result: FAIL
>>
>> So, that's 215s now in branch vs. 221s a few hours ago in trunk. (only
>> ran 'prove' once each time)
>
> So that's a small improvement, which I guess is better then nothing.
> At least it's not a regression, which is always something.
>
> --Andrew Whitworth

Without a more thorough benchmark test, that could just be line noise.
(That's why I pointed out I just ran prove once each time.)

Regards.

-- 
Will "Coke" Coleda