Front page | perl.perl5.porters |
Postings from July 2014
Re: Build fixes for cherry-picking to maint-5.20?
Thread Previous
|
Thread Next
From:
Brian Fraser
Date:
July 29, 2014 16:58
Subject:
Re: Build fixes for cherry-picking to maint-5.20?
Message ID:
CA+nL+naPRfv7ibRV4GZifGYUH_=8ofB8fqRDZo2Lf9jhxPAHeA@mail.gmail.com
On Tue, Jul 29, 2014 at 4:19 PM, Brian Fraser <fraserbn@gmail.com> wrote:
> On Tue, Jul 29, 2014 at 9:53 AM, Steve Hay <steve.m.hay@googlemail.com> wrote:
>> On 28 July 2014 12:27, Brian Fraser <fraserbn@gmail.com> wrote:
>>> On Mon, Jul 28, 2014 at 10:01 AM, Steve Hay <steve.m.hay@googlemail.com> wrote:
>>>> The following build/test fixes, mostly for Android, look good for
>>>> maint-5.20 to me. Any seconders and thirders?
>>>>
>>>
>>> Since I committed most of these, I'll abstain from actually voting,
>>> but a soft +1 to all of them, except
>>>
>>>> 329865d Makefile.SH: Fix cross builds with -Uhostgenerate
>>>
>>>
>>>> ef4ead6 Makefile.SH: Add missing files for make test on cross builds
>>>
>>> This. I'm not sure if it's needed for 5.20; I think it might be all
>>> due to changes in blead. The problem is that, if a file should be
>>> copied but isn't, then one test during all of 'make test' will fail.
>>> However, if we try to copy a file that doesn't exist in 5.20, then
>>> *all* of make test will fail.
>>>
>>
>> Would it be easy for you to test the current maint-5.20 branch to
>> determine if we need this change or not?
>>
>> (I think this is not something that I can test myself (on Windows...)
>> since it's a change to Makefile.SH.)
>
> Sure! I'll give this a shot when I get home in a couple of hours.
Tested, 5.20 does need the backport; customized.t fails without it.
It looks like it doesn't cherry-pick cleanly on 5.20, but the conflict
is pretty trivial.
Thread Previous
|
Thread Next