Re: NetBSD current's native xorg support

[Ryo ONODERA <ryo_on%yk.rim.or.jp@localhost>]

Hi,

From: Greg Troxel <gdt%ir.bbn.com@localhost>, Date: Wed, 26 Mar 2014 19:41:24 
-0400

> 
> Ryo ONODERA <ryo_on%yk.rim.or.jp@localhost> writes:
> 
>> From: Ryo ONODERA <ryo_on%yk.rim.or.jp@localhost>, Date: Sun, 23 Mar 2014 
>> 03:13:51 +0900 (JST)
>>
>>> From: Thomas Klausner <wiz%NetBSD.org@localhost>, Date: Sat, 22 Mar 2014 
>>> 16:06:30 +0100
>>> 
>>>> On Sat, Mar 22, 2014 at 11:36:06PM +0900, Ryo ONODERA wrote:
>>>>> NetBSD current has newer native xorg libraries including freetype 2.5.3 
>>>>> now.
>>>>> I would like to update pkgtools/x11-links for NetBSD curernt's native 
>>>>> xorg.
>>>>> And builtin support for freetype 2.5.3.
>>>>> 
>>>>> Patches are attached.
>>>>> 
>>>>> Can I commit thse patches?
>>>> 
>>>> Please wait for after the freeze.
>>>> 
>>>> It only helps one platform (NetBSD-current), but ALL bulk builds will
>>>> basically have to restart building all X packages.
>>> 
>>> Is see.
>>> I will commit them after freeze.
>>> Thank you.
>>
>> Branching pkgsrc-2014Q1 is postponed, and some PR should be resolved by
>> these patches.
>> I feel that these patches should be committed before branching...
> 
> I'm trying to understand the consequences here.
> 
> Are you saying that there are significant numbers of people who build
> From the pkgsrc stable branch on NetBSD-current systems?  Or are there
> binary packages available which are from stable branches but built
> on/for some verison of NetBSD current?
> 
> Or is this just about a fix (which seems to be not entirely clearly 100%
> arrived at so far) being committed a week earlier for people who will
> stay on pkgsrc-current with NetBSD-current?
> 
> The real goal of the stable branch is to provide a high-quality set of
> packages for people who either
> 
>   use binary packages from bulk builds (mostly), or
> 
>   build from source, but when they try everything builds
> 
> We've been struggling with not test-compiling packages because of other
> failures, and as those get fixed we get more test coverage.  An
> extension of the freeze (which I haven't seen announced but which seems
> like a good idea) is primarily a reflection that we didn't get the
> expected testing from the first week.  It's in the second week or so of
> a freeze that it's particularly important to avoid trouble with bulk
> builds, and with a three week freeze we'd essentially have a week that
> didn't feel stable and what passes for 1st and then 2nd weeks.
> 
> I realize that fairly few people are doing bulk builds.  But they are
> very important, and we should be trying to make doing bulk builds during
> freezes as good an experience as we can to encourage more people to do
> it.
> 
> So I would like to hear your thoughts about whether waiting would hurt
> users of the forthcoming stable branch.

Some actively working users are discouraged by this build failure.
And stable branch does not support NetBSD current for 3 months.
My concern is these.

Of course bulk build is important, and I cannot discourage bulk builders.

--
Ryo ONODERA // ryo_on%yk.rim.or.jp@localhost
PGP fingerprint = 82A2 DC91 76E0 A10A 8ABB  FD1B F404 27FA C7D1 15F3