Re-index on any/all -new- entries in 02packages.details #220

oalders opened this Issue Aug 8, 2012 · 0 comments


None yet
3 participants

oalders commented Aug 8, 2012

kentnl: [05:14:24] -> module -> == 404
[10:39am] dipsy: [05:14:25] [ signatures-0.07 - subroutine signatures with no source filter - ]
[11:03am] oalders: kentnl: signatures fixed with a re-index
[11:07am] mst: oalders: well, that sucks
[11:07am] mst: oalders: metacpan shouldn't have to reindex like that
[11:07am] oalders: mst: it shouldn't
[11:08am] mst: oalders: that was a dist that was uploaded with the wrong perms, then the perms were fixed and it was re-indexed at the PAUSE level
[11:08am] oalders: issue is things in the indexer get fixed, but we don't re-run the whole thing again after that
[11:08am] mst: yes, but in this case the problem was within the last 48h
[11:08am] mst: have you fixed it in the indexer since then?
[11:08am] oalders: mst: no. i just re-indexed it manually in this case
[11:09am] mst: so the bug is still there
[11:09am] oalders: the bug with this particular dist?
[11:09am] mst: no, the bug that meant metacpan didn't reindex it automatically
[11:11am] oalders: how should metacpan know to re-index that one? we're just watching the changes reported by rrr.
[11:11am] oalders: so, that doesn't pick up perm changes, does it?
[11:14am] mst: when you sync the PAUSE index
[11:14am] mst: you need to check for new entries
[11:24am] oalders: mst: that makes sense
[11:24am] oalders: will open an issue for that
[11:24am] mst: I thought you already had something like that
[11:25am] mst: since when a dist hasn't made it to the index yet
[11:25am] mst: you mark it as indexed about ... an hour later?
[11:25am] oalders: i'll have to check the logic, but i think there's a time limit on how far we go back
[11:29am] mst: it would seem to me that the answer would be to test any/all -new- entries in 02packages.details
[11:29am] mst: i.e. do a diff, reindex all lines that are new
[11:29am] mst: seems simple enough, and shouldn't have any more things to index than normal
[11:30am] mst: but would nail this edge case nicely

See also: #390

@yanick yanick pushed a commit to yanick/cpan-api that referenced this issue Mar 30, 2013

@monken monken add MIME header, refs #220 8f3d2d9

rose was assigned by oalders Dec 9, 2014

rose was unassigned by oalders Feb 12, 2015

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment