FS#17325 - {archweb} Move maintainer information off packages table

Attached to Project: Arch Linux
Opened by Dan McGee (toofishes) - Monday, 30 November 2009, 06:12 GMT
Last edited by Dan McGee (toofishes) - Thursday, 29 April 2010, 15:52 GMT
Task Type Feature Request
Category Web Sites
Status Closed
Assigned To Dan McGee (toofishes)
Architecture All
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

On Wed, Nov 25, 2009 at 9:20 AM, Allan McRae <allan@archlinux.org> wrote:
> Andrea Scarpino wrote:
>>
>> Oh no!!
>> This happened again! Or someone orphaned 500 packages cause we have
>> ~2600 orphans now; more then 1700 in [extra].
>> Can you check your packages?
>
> It seems to have occurred only on i686 for packages starting with "p"
> onwards.

So it looks like we aren't sure what happened here, which sucks.

This can be addressed if we make a more substantial code change in the
application. Right now, the maintainer info is stored on the package
itself, so if reporead decides that package should get zapped (which I
have no idea at all why that happens), then it also kills the
maintainer info.

Instead, we could store pkgname <-> maintainer mappings in a separate
table. This has two distinct advantages. One, the maintainer field is
not dependent anymore on architecture or repository- by adding one
"dan" <-> "pacman" mapping entry, I become the maintainer for pacman
in all architectures and all repositories. The other advantage of
course is deleting a package does not delete maintainer info. Yes, we
might have some stale maintainer info in the table, but who cares- we
don't have more than 10,000 pieces of data in the puzzle here.

One question- should it be based off of pkgname or pkgbase?
This task depends upon

Closed by  Dan McGee (toofishes)
Thursday, 29 April 2010, 15:52 GMT
Reason for closing:  Fixed
Comment by Allan McRae (Allan) - Monday, 30 November 2009, 06:29 GMT
Definitely pkgbase. That way all split packages are treated as one package.
Comment by Thomas Bächler (brain0) - Thursday, 29 April 2010, 15:50 GMT
Ehm, Dan, didn't you implement exactly this recently? Shouldn't we close this?
Comment by Dan McGee (toofishes) - Thursday, 29 April 2010, 15:52 GMT
Yeah, and I already closed  FS#11312  which was similar stuff.

Loading...