Re: Disappearing packages from wip : what is the policy

[Greg Troxel <gdt%lexort.com@localhost>]

Greg Troxel <gdt%lexort.com@localhost> writes:

> Robert Swindells <rjs%fdy2.co.uk@localhost> writes:
>
>> One example of a wip package that has been deleted is freecad,
>>
>> This seems to have been done as part of a bulk cleanup:
>>
>>     "remove packages with outdated distinfo files"
>>
>> The freecad package has sources on github so could have been updated
>> to pass the distinfo check, it also has a listed maintainer.
>
> git log shows
>
> commit 27326db8fb273e583c63d47d9ac8b2f2c6e8052d
> Author: Roland Illig <roland.illig%gmx.de@localhost>
> Date:   Sat Sep 14 20:09:47 2024 +0100
>
>     remove packages with outdated distinfo files
>     
>     The hash algorithm RMD160 has not been in use for several years now,
>     so every package that still uses it is suspicious of being unmaintained
>     and having no working MASTER_SITES, as otherwise it would have been
>     possible to update the distinfo file by running "bmake distinfo".
>     
>     grep -l ^RMD160 */distinfo | sed s,/distinfo,, > obsolete
>     git rm -r $(cat obsolete)
>     pkglint -F

This removed a massive number of packages:

  $ git log -1 --stat 27326db8fb273e583c63d47d9ac8b2f2c6e8052d|egrep Makefile|wc
      4969   19876  308080

I have already asked for the entire commit to be reverted, as it is a
change that clearly should have been discussed, and was not.   As I said
in that message, and have said before, (expanding slightly here) we have
norms that changes that are controversial should be discussed, and if a
change turns out to be controversial when not anticipated to be so
(certainly that can happen), it should be immediately reverted and we
can then discuss the change *with it not present in the tree*.