Scala 2.8.1 RC3 - details

On 10/14/2010 03:17 PM, Antonio Cunei wrote:
> Hello everyone,
>
> We were basically ready to push out RC2 as the final release, but decided
> to issue one more RC in order to address two additional regressions.
>
> We decided not to consider any further patches at this time since we are
> in a pretty advanced stage in the RC release cycle, and every additional
> change has the potential to impact on stability and/or binary
> compatibility.
>
> The only additional fixes in RC3 are:
>
> #3894 - Regression against 2.8.0 in Scaladoc
> #3915 - Compiler crash (also regression against 2.8.0)
>
> Unless we receive reports of some serious additional regression, RC3 will
> become the final one week from now. It's time to test! :)

There was an additional bug found in RC2 for #3605 that I committed a
fix for. I'm not sure if this warrants leaving the RC as-is, pulling in
the new fix, or removing #3605 from 2.8.1 entirely :( I wouldn't call
#3605 serious, but I'm not sure what the process would be here.

Thanks,

Derek

On Fri, October 15, 2010 12:37 am, Derek Chen-Becker wrote:
> There was an additional bug found in RC2 for #3605 that I committed a
> fix for. I'm not sure if this warrants leaving the RC as-is, pulling in
> the new fix, or removing #3605 from 2.8.1 entirely :( I wouldn't call
> #3605 serious, but I'm not sure what the process would be here.
>

Hello Derek,

Ticket #3605 is currently still marked as open, with the last message
suggesting a move to 2.8.2. It actually appeared to be still in progress.

In any case, as mentioned in the previous message, we decided to include
really only the major regressions this time. In general, at the beginning
of the release cycle we are very ready to accept all suggestions for the
patches that seem useful to include; however, as we progress through RCs,
we really put the main priority on stability, meaning that only tickets
that are significant problems or regressions can be included: every new
patch added to the code has the potential do destabilise something else.
Also, we cannot keep including new stuff, or we will never get to a final
release :)

This time, as we are now really close to the final, we only fixed those
two regressions. The remaining fixes will have to be delayed just a little
bit longer, to the next 2.8.2, or a 2.9.

Toni

On Fri, Oct 15, 2010 at 12:03 AM, Antonio Cunei wrote:
> This time, as we are now really close to the final, we only fixed those
> two regressions. The remaining fixes will have to be delayed just a little
> bit longer, to the next 2.8.2, or a 2.9.

I nominate ticket #3923[1] for 2.8.2. The title is misleading, the
actual problem is that (Nil == any empty mutable seq) would return
false. Paul has fixed it in trunk:

https://lampsvn.epfl.ch/trac/scala/changeset/23260

Best,
Ismael

[1] https://lampsvn.epfl.ch/trac/scala/ticket/3923

On 10/14/2010 05:03 PM, Antonio Cunei wrote:
> On Fri, October 15, 2010 12:37 am, Derek Chen-Becker wrote:
>> There was an additional bug found in RC2 for #3605 that I committed a
>> fix for. I'm not sure if this warrants leaving the RC as-is, pulling in
>> the new fix, or removing #3605 from 2.8.1 entirely :( I wouldn't call
>> #3605 serious, but I'm not sure what the process would be here.
>>
>
> Hello Derek,
>
> Ticket #3605 is currently still marked as open, with the last message
> suggesting a move to 2.8.2. It actually appeared to be still in progress.

I reopened the ticket because I committed the new fix and I wasn't sure
what to do about inclusion. More details below.

>
> In any case, as mentioned in the previous message, we decided to include
> really only the major regressions this time.

I completely agree. I thought that #3605 had already been included in
one of the 2.8.1 RCs (it's listed in your announcement as one of the
fixed bugs) so I was confused about what to do, but if it was never
included in the first place then there's no problem at all leaving it
until 2.8.2.

Thanks,

Derek