Re: git: f98beca9766c - main - devel/glib20: add DOCS option [837447eabc74 still fails]

On 2022-Apr-17, at 20:44, Mark Millard <marklmi@yahoo.com> wrote:

> [At least this time I tested the right vintage:
> git: 837447eabc74 - main - devel/glib20: fix up DOCS generation Adriaan de Groot
> But it is still broken, but only for one "ld: error:" in my context:
> ld: error: unable to find library -lgobject-2.0 .]
> 
> On 2022-Apr-17, at 15:57, Mark Millard <marklmi@yahoo.com> wrote:
> 
>> [Ignore the report. I'm retrying after my screwup.]
>> 
>> On 2022-Apr-17, at 10:15, Mark Millard <marklmi@yahoo.com> wrote:
>> 
>>> From: Adriaan de Groot <adridg_at_freebsd.org> 
>>> Date: Sun, 17 Apr 2022 15:15:15 +0200 :
>>> 
>>>> On Sunday, 17 April 2022 11:12:11 CEST Jochen Neumeister wrote:
>>>>> I have it currently also at 13.0 i386:
>>>> 
>>>> My meson-fu is weak -- and since I need a faster or slower machine to actually 
>>>> reproduce the problem, I have selected a proxy problem: `ninja -j48 gio-doc`. 
>>>> That's an actual target defined by meson, and I believe it should "just 
>>>> build". It didn't, of course: missing dependencies on things-that-are-
>>>> generated-during-the-build.
>>>> 
>>>> Telling ninja that gio-doc needs to build the examples first is trivial, so I 
>>>> patch that in. It's definitely ugly to be doing {REINPLACE_CMD} on the 
>>>> generated build-system, but it's the best I could come up with.
>>> 
>>> What I got was (for [ 33% 463/1333]):
>> 
>> I never managed to do the merge --ff-only : just the fetch.
>> I'm rerunning bulk after actually having merged the update.
>> It will be a while if it works as all 3 jails on that machine
>> are to be updated in sequence and the armv7 one is the last
>> of the 3.
>> 
>>> ERROR: 
>>> . . .
> 
> I had tested building the wrong version.
> 
>>> Again a 16 core Cortext-A72 system (HoneyComb) using an armv7
>>> poudriere-devel jail.
>>> 
>> 
>> 
> 
> It still fails to build in my context. But this time
> there is only one "ld: error:". This was for:
> 
> . . .
> 

I finally had a bulk build targeting armv7 for which devel/glib20
built to completion. I've no clue if it was an accidental race
result vs. a systematic result. It was a build for updating
to be based on:

# ~/fbsd-based-on-what-commit.sh -C /usr/ports/
branch: main
merge-base: 27e36df36a9d47e844b1fa83d4136b63fee7c905
merge-base: CommitDate: 2022-04-28 02:19:08 +0000
27e36df36a9d (HEAD -> main, freebsd/main, freebsd/HEAD) databases/redis: Update to 7.0.0
n582007 (--first-parent --count for merge-base)


===
Mark Millard
marklmi at yahoo.com