Re: 141releng-armv7-quarterly stuck in run-depends status for octave-forge-20241116: 34:28:33 for package build so far

Mark Millard <[email protected]>
Newsgroups gmane.os.freebsd.devel.ports,gmane.os.freebsd.devel.arm
Message-ID <[email protected]>
Hello Robert and Philip,

Philip: This note indicates that the build does eventually finish.
So it is not a "need too kill the buildler" context.

On Feb 11, 2025, at 03:20, Robert Clausecker <[email protected]> wrote:

> Am Mon, Feb 10, 2025 at 08:08:43PM -0800 schrieb Mark Millard:
>> https://pkg-status.freebsd.org/ampere1/build.html?mastername=141releng-armv7-quarterly&build=525876c27745
>> 
>> shows that the build of octave-forge-20241116 has been stuck for 34:28:33,
>> only having reached the status run-depends. (Everything else has finished
>> building.)
>> 
>> The overall from scratch "bulk -a" time is at the moment around: 115:09:42 .
>> 
>> ===
>> Mark Millard
>> marklmi at yahoo.com
> 
> I observed the same thing.  Octave-forge is installing its various modules
> one by one, but with each module it judges its package database to be corrupt,
> so it reinstalls every module again, leading to O(n²) runtime.  I do not know
> what the reason for this is.

It did finish --after 39:49:02 .

As long as the O(n²) build-time problem exists: Is the
package's build worth the delays to most other package
builds that are done on on any specific builder machine?

ampere1 cycles through building and distributing:
141arm64-quarterly
141releng-armv7-quarterly
1341arm64-quarterly
134releng-armv7-quarterly

ampere3 is similar (default here is a.k.a. latest):
141arm64-default
141releng-armv7-default
1341arm64-default
134releng-armv7-default

If the armv7's all have that time problem, the
problem looks to adds days to the time to complete
a cycle of 4 types of builds.

> The reason why you see this only now is that I have recently pushed a patch
> to fix math/octave on armv7; previously it didn't build.=

Should the package be considered broken for armv7 until
it is fixed to build in a normal time frame instead of
an O(n²) tiem frame, just because it takes too long as
stands?

===
Mark Millard
marklmi at yahoo.com
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.