Re: git: 2ba4ed91621e - main - sysutils/createrepo_c: new port

From: Daniel Engberg <daniel.engberg.lists_at_pyret.net>
Date: Fri, 04 Oct 2024 03:34:41 UTC
On 2024-10-02T09:21:13.000+02:00, Pietro Cerutti <gahr@FreeBSD.org>
wrote:

> On Oct 01 2024, 17:04 +0000, Daniel Engberg <daniel.engberg.lists@pyret.net> wrote:
>>  Please review Porters Handbook, there are several "guidelines" to
>>  make
>>  
>>  the tree more consistent and reducing overall workload.
> 
> I have pushed a few cleanups in
> 771ce8599db53660f3aa7c1e31dba1c87852586c.

Thanks!

>>>   PORTVERSION --> DISTVERSION
>>  
>>  https://docs.freebsd.org/en/books/porters-handbook/book/#makefile-master_sites-github
>>  
>>  ...and stray(?) comment?
> 
> Thanks, corrected.
> 
>>  Correct maintainer? There's no reference and from what I can tell
>>  no
>>  
>>  submitted PR in Bugzilla.
> 
> The maintainer is correct. It's a person I know and the port was 
> 
> submitted to me privately. I didn't think anyone would want it 
> 
> publicized it in the commit message.

Just add something like "submitted via private email" or something
similar so it's clear of origin

>>  libxml2 dependency should use the GNOME helper
>>  
>>  https://docs.freebsd.org/en/books/porters-handbook/book/#using-gnome
>>  
>>  https://docs.freebsd.org/en/books/porters-handbook/book/#gnome-components
>>  
>>  sqlite3 dependency should also use a helper
>>  
>>  https://docs.freebsd.org/en/books/porters-handbook/book/#using-databases-uses
>>  
>>  CMAKE_ARGS can be replaced with CMAKE_OFF= ENABLE_PYTHON
>>  
>>  https://docs.freebsd.org/en/books/porters-handbook/book/#using-cmake
> 
> Thanks, corrected.
> 
>>  do-test section can be replaced with USES= cmake:testing
>>  
>>  https://cgit.freebsd.org/ports/tree/Mk/Uses/cmake.mk#n20
> 
> I wasn't able to do it with cmake:testing. Were you?

Upstream uses from what I can tell an rather old approach for test
target,

https://github.com/rpm-software-management/createrepo_c/blob/master/CMakeLists.txt#L160

The "EXCLUDE FROM ALL" part is what causes the issue

>>  Poudriere also lists quite a few warnings about dependencies and
>>  the
>>  
>>  REINPLACE_CMD commands. 
>>  
>>>   https://pdr2.bofh.network/data/141-diizzy/2024-10-01_11h36m29s/logs/createrepo_c-1.1.4.log
> 
> Nice. I wasn't aware of poudriere's Q/A phase. I used poudriere bulk
> to 
> 
> test the port but I don't see that in my logs. How do I enable it?
> 
> https://gahr.ch/poudriere/logs/bulk/14amd64-dev/2024-10-01_12h26m54s/logs/createrepo_c-1.1.4.log
> 
> Also, I have run `portlint -AC` multiple times as I was reviewing
> the 
> 
> port, and it was happy with what I had.
> 
> Yeah, I should skim through the handbook once in a while. I've been 
> 
> doing ports for a long time and I am probably missing out on the
> latest 
> 
> enhancements and developments.
> 
>>  It also seems to look for Doxygen by default
>>  
>>  which you probably want to disable to avoid potential plist
>>  issues.
> 
> Well, they don't provide a way to turn off docs, but I don't see an 
> 
> issue with it building docs if it finds doxygen. We won't be
> installing 
> 
> them anyway.

https://cmake.org/cmake/help/latest/variable/CMAKE_DISABLE_FIND_PACKAGE_PackageName.html
works :-)

> -- 
> 
> Pietro Cerutti
> 
> I have pledged to give 10% of income to effective charities
> 
> and invite you to join me - https://givingwhatwecan.org

I've hacked together a patch doing the final cleanups which fixes the
Q/A issues.

I've also left a comment in one of the submitted PRs as it breaks
build if optional dependencies are disabled.

https://pdr2.bofh.network/data/141-diizzy/2024-10-03_21h47m26s/logs/createrepo_c-1.1.4.log
(testport)

https://pdr2.bofh.network/data/133-diizzy/2024-10-03_22h05m43s/logs/createrepo_c-1.1.4.log
(bulk)

Not sure why your Podriere doesn't run QA on bulk builds :-/

Patch:

https://projects.pyret.net/files/freebsd-patches/createrepo_c-improve.patch

Best regards,

Daniel