[PREP] Last Rounds for Parrot 0.9.1

[Reini Urban]

2009/2/17 Andy Dougherty <doughera at lafayette.edu>:
> On Mon, 16 Feb 2009, Mark Glines wrote:
>
>> kjstol wrote:
>> > If you have any:
>> > * critical bug fixes
>> > * updates
>> > * stuff to be deprecated and removed before 1.0
>> > * anything else you'd like to get in 0.9.1
>> >
>> > please commit/add this before tomorrow.
>>
>> Got a couple of TODO tests passing (pbc integer/number), but otherwise,
>> test and codetest look good on linux/x86-64 in r36806.
>
> I think the t/native_pbc/*.t tests should all be SKIPPED for the release.
> They dump core on 64-bit systems and "Unexpectedly pass" on some other
> systems.  They aren't telling us anything useful.

They "Unexpectedly pass" because you didn't delete the devel flag, and
didn't update the VERSION for testing the "release".

The new pbc's 1,2,4,5 were prepared for the 0.9.1 release, not for
0.9.0-devel testing.
I cannot prepare 2 and 6 pbcs as I have no big-endian boxes.

On devel they are TODO, on the release not.

But since strings are also broken now with 64-bit they could be skipped at all.
They only show that 64-32 interaction is broken, most likely due to
broken 64bit alignment. And we already know that.
But I would have liked to get more test reports on not-so known
systems, such as from big-endian boxes.

Skipped tests are less useful than passing todo tests.
-- 
Reini