please improve getPublishedSource() created_since_date API description

Bug #685814 reported by Michael Vogt
10
This bug affects 2 people
Affects Status Importance Assigned to Milestone
Launchpad itself
Triaged
Low
Unassigned

Bug Description

I put a launchpadlib based changelog extraction tool into production now that uses getPublishedSources. It works great and is very speedy, however it appears it misses a source package every now and then. The way it works is that it wakes up via cron and asks for the latest changes since it was run last (curently that is every 15min).

It looks like the problem is releated to bug https://bugs.launchpad.net/soyuz/+bug/372270. It can take 1h (for the primaty archive according to the bug) between record created and published source available.

It would be nice if that could be documented in the API description. Something like
"""
Created Since Date in the database.

Please note that there can be a certain delay (up to 1h currently for the main archive) between a record being
created and publication. Keep that in mind when using this in combination with the "status" parameter.

Return entries whose date_created is greater than or equal to this date
"""

Tags: api lp-soyuz doc
Michael Vogt (mvo)
summary: - getPublishedSource() created_since_date confusing
+ please improve getPublishedSource() created_since_date API description
Changed in soyuz:
status: New → Triaged
importance: Undecided → Medium
tags: added: api doc
Revision history for this message
Robert Collins (lifeless) wrote : Re: [Bug 685814] Re: please improve getPublishedSource() created_since_date API description

Michael, are you using the 'most recent date you saw returned', or
'most recent date your script ran on' - the former should be safe, the
latter unsafe.

I agree we'll want to be clearer about this, but I'd like to note that
its a general principle when you have any sort of replication in play.

-Rob

Revision history for this message
Julian Edwards (julian-edwards) wrote :

On Monday 06 December 2010 18:06:25 you wrote:
> Michael, are you using the 'most recent date you saw returned', or
> 'most recent date your script ran on' - the former should be safe, the
> latter unsafe.
>
> I agree we'll want to be clearer about this, but I'd like to note that
> its a general principle when you have any sort of replication in play.

The code was written so that you could pass back in the actual timestamp of
the last record that you received and it will continue from where it left off.

That's the bit that needs clarifying in the documentation :)

Changed in launchpad:
importance: Medium → Low
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.