Re: [aur-general] Delete Git tree for new package with broken state

2019-03-28 Thread lambdadroid via aur-general
On Thu, Mar 28, 2019 at 04:05:06PM +0100, Robin Broda via aur-general wrote:
> On 3/27/19 1:43 PM, lambdadroid via aur-general wrote:
> > Hi,
> > 
> > I've been trying to upload packages from archlinux-me176c [1] to the
> > AUR. My scripts create a subtree split of one package, then do quite
> > a bit of other magic and upload it to the AUR.
> > 
> > I've been having a few problems with "linux-me176c"
> > because I used to have large patch files in the repository (> 250 KiB).
> > So I modified the scripts to cut off the history after I switched to
> > fetching from a Git repository instead.
> > 
> > This works fine, but while testing I seem to have uploaded a broken
> > subtree or some temporary commit, so the new generated subtrees no
> > longer match what is uploaded on the AUR - making it impossible for me
> > to update the package (without force-push).
> > 
> > Deleting the package on the AUR was accepted automatically because of:
> >Deletion of a fresh package requested by its current maintainer.
> > but I should have read on the wiki first that this does not delete the
> > Git tree. :)
> > 
> > Is there any chance to remove the Git repository of this new package,
> > or should I somehow try to recover that broken state?
> > 
> > Thanks,
> > lambdadroid
> > 
> > [1]: https://github.com/me176c-dev/archlinux-me176c
> > 
> 
> there's no broken state. pull from the AUR remote and rebase your commits on 
> top of that
> 

I would like to do this but I'm not sure how this would work in my case:
 
I push to the AUR using an automated subtree setup (very similar to 
aurpublish [1]). For each subdirectory that represents an AUR package, 
it splits the commit history using "git subtree", and pushes it to the 
AUR.

I've taken a closer look at the Git repo I pushed and it looks like I 
only accidentally pushed a temporary commit to "linux-me176c" that no 
longer exists in my main repository. So the new subtrees I generate no 
longer match what exists on the AUR.

I would prefer to remove that commit since there is a proper commit 
message on my main repository, but I'm also not sure how to solve this 
situation with a rebase:

 - I have already pushed other packages using the new state, so if I 
   rebase the main repo, it'll stop working for those other packages
   (because the other subtrees will no longer match)

 - How do I rebase a subtree back into the main repository?

Is there any chance that you (or someone else) could simply remove the 
last commit with a force-push? There is absolutely no history that is 
getting lost here, since this is a completely new package.

Thanks,
lambdadroid

[1]: https://github.com/eli-schwartz/aurpublish

> -- 
> Rob (coderobe)
> 
> O< ascii ribbon campaign - stop html mail - www.asciiribbon.org


Re: [aur-general] Delete Git tree for new package with broken state

2019-03-28 Thread Robin Broda via aur-general
On 3/27/19 1:43 PM, lambdadroid via aur-general wrote:
> Hi,
> 
> I've been trying to upload packages from archlinux-me176c [1] to the
> AUR. My scripts create a subtree split of one package, then do quite
> a bit of other magic and upload it to the AUR.
> 
> I've been having a few problems with "linux-me176c"
> because I used to have large patch files in the repository (> 250 KiB).
> So I modified the scripts to cut off the history after I switched to
> fetching from a Git repository instead.
> 
> This works fine, but while testing I seem to have uploaded a broken
> subtree or some temporary commit, so the new generated subtrees no
> longer match what is uploaded on the AUR - making it impossible for me
> to update the package (without force-push).
> 
> Deleting the package on the AUR was accepted automatically because of:
>Deletion of a fresh package requested by its current maintainer.
> but I should have read on the wiki first that this does not delete the
> Git tree. :)
> 
> Is there any chance to remove the Git repository of this new package,
> or should I somehow try to recover that broken state?
> 
> Thanks,
> lambdadroid
> 
> [1]: https://github.com/me176c-dev/archlinux-me176c
> 

there's no broken state. pull from the AUR remote and rebase your commits on 
top of that

-- 
Rob (coderobe)

O< ascii ribbon campaign - stop html mail - www.asciiribbon.org


Re: [aur-general] plethora of spams

2019-03-28 Thread Daniel M. Capella via aur-general
On March 20, 2019 3:49:00 AM EDT, marnout  wrote:

~snip~

>P.S.
>I have a lot of free time and i can participate by doing certain tasks 
>for the community. For instance I can clean up in aur under the
>guidance 
>of a TU.

https://wiki.archlinux.org/index.php/DeveloperWiki:AUR_Cleanup_Day

Of course requests[1] made outside of that day are also considered. :)

[1]: https://wiki.archlinux.org/index.php/Arch_User_Repository#Other_requests

--
Best,
polyzen


Re: [aur-general] Spring cleaning (moving orphaned packages from [community] to AUR)

2019-03-28 Thread Antonio Rojas via aur-general
El jueves, 28 de marzo de 2019 10:35:46 (CET), Alexander F Rødseth via 
aur-general escribió:

You're welcome! I moved the packages to AUR first, then ran the db-remove
script later, that also removed the packages from SVN. Using `svn log
grumpy`:

r444125 | arodseth | 2019-03-27 10:28:54 +0100 (on., 27 mars 2019) | 1
line
db-remove: grumpy removed by arodseth



db-remove only removes the corresponding /repos entry from svn. The svn 
dirs themselves are still there:


https://git.archlinux.org/svntogit/community.git/tree?h=packages/neverball


Re: [aur-general] Spring cleaning (moving orphaned packages from [community] to AUR)

2019-03-28 Thread Alexander F Rødseth via aur-general
Antonio Rojas wrote:
> Thanks for the much needed cleanup. It seems you forgot to remove them
from svn though.

You're welcome! I moved the packages to AUR first, then ran the db-remove
script later, that also removed the packages from SVN. Using `svn log
grumpy`:

r444125 | arodseth | 2019-03-27 10:28:54 +0100 (on., 27 mars 2019) | 1
line
db-remove: grumpy removed by arodseth

-- 
Sincerely,
Alexander F. Rødseth / xyproto