Aiming for "release"

[Rob 'Commander' Pike]

Things have been a bit churny lately (mea culpa) and are going to get
more so (russa culpa) in the next few days. Many of these changes
have been planned for a while, but they will introduce problems.

I'd like to suggest we aim for a stable state soon, so we can tag a
"release", and propose we aim for the weekly two weeks from now to be
stable enough for it to be the candidate.

Thus, other than the big changes already underway, I'd like to hold
off significant changes for a while and concentrate on bug fixes and
stability for the next couple of weeks.

So two weeks from now, we'd like to have a solid code base with bug
and stability fixes and the following new things installed:

- new reflect API
- new reflect implementation
- new net setup (using local libraries for name resolution, etc.)

If there's something you think should be part of this set, please say
so now so we can plan properly. Other major changes will most likely
be postponed unless a compelling case is presented.

-rob

[Robert Griesemer]

Just FYI: I have a larger (1KLOC) CL forthcoming that is implementing
imports and a new type representation for AST nodes. It is very local
at this point and doesn't affect anything else so I like it to be part
of the set.

Otherwise I agree that a round of bug fixing and stabilization is a
very good idea.

- Robert

[Lorenzo Stoakes]

On 5 April 2011 17:21, Rob 'Commander' Pike <r...@google.com> wrote:

>
> If there's something you think should be part of this set, please say
> so now so we can plan properly.  Other major changes will most likely
> be postponed unless a compelling case is presented.
>

I'd like to see the issue 667 bug fix go in :-)

--
Lorenzo Stoakes
http://www.codegrunt.co.uk

[Mikio Hara]

On Tue, Apr 5, 2011 at 4:21 PM, Rob 'Commander' Pike <r...@google.com> wrote:

> So two weeks from now, we'd like to have a solid code base with bug
> and stability fixes and the following new things installed:

(snip)

> - new net setup (using local libraries for name resolution, etc.)

I have a small CL <http://codereview.appspot.com/4360047/>
regarding a bit new net staff you suggested but I would like to
see and wait for forthcoming CL from core team.

-- Mikio

[Rob 'Commander' Pike]

I'm still aiming for a stable release in a few days, but seeing little
evidence that others agree. Please, other than the things we agreed
on, please favor fixes and stability improvements over new features.

I'm partly to blame for not pushing back on some of the churn. I will
be more vigilant in the days ahead.

-rob

[peterGo]

rob,

On Apr 15, 11:17 am, "Rob 'Commander' Pike" <r...@google.com> wrote:
> I'm partly to blame for not pushing back on some of the churn. I will
> be more vigilant in the days ahead.

This is going to break the Windows port for a "nice-to-have" feature.

Issue 4430054: code review 4430054: gofmt: add -diff
http://codereview.appspot.com/4430054/

It's a same sort of problem as gotest. It's running a command.

Can we wait until it's fully tested?

Peter

[Russ Cox]

It still builds right?  Otherwise gofix wouldn't.  It won't work if someone uses the flag but that's fine.

[peterGo]

Russ,

It builds but crashes when executed with the flag.

I understand that that may be OK for the tip or perhaps the weekly but
it's more unstable than I would expect from a stable release. If it is
going to fail shouldn't it fail gracefully?

Peter

[David Crawshaw]

Is this a bug in exec.LookPath? It expects PATH to be colon-separated,
is that true on windows? (It's been a while.)

[David Crawshaw]

Nm, reading the wrong file.

[peterGo]

David,

We are working on a fix for this problem in gofix. When that's
finished, we should be able to clone the solution to fix gofmt.

Peter

On Apr 19, 2:39 pm, David Crawshaw <craws...@google.com> wrote:
> Is this a bug in exec.LookPath? It expects PATH to be colon-separated,
> is that true on windows? (It's been a while.)
>