The debian Perl group's Quality Assuarance software

This page documents perl scripts developed to monitor the condition of each package the debian-perl group maintains. (We should get a real name for them :) )

These perl scripts have recently been rewritten to clean up the code, ease improvements and make them usable by other groups. This page is meant to be a place to discuss development and provide documentation.

Current components:

Things to-do:

Classification issues

Classifying the different package status has proven to be confusing and hard to get right. So, I think we should first agree "on paper".

SVN released version

SVN unreleased version

upstream version

archive most recent version

archive other version

category

notes

missing changelog

Error

missing watchfile

Error

none

1.1

1.0

*

*

Needs upgrade / invalid?

could be work in anticipation

1.0

1.1

1.0

*

*

WiP / not shown / needs upgrade / invalid?

none

1.0

1.0

*

*

Needs upgrade / WiP?

none

1.0

1.1

*

*

Needs upgrade?

1.0

*

0.9

*

*

Invalid version

1.0

*

*

1.1

*

Invalid version

1.0

none

1.0

0.9

*

Needs upload

1.0

none

1.0

1.0

*

OK

1.0

none

1.1

*

*

Needs upgrade

1.0

none

1.1

0.9

*

Needs upload + upgrade?

1.0

1.1

1.1

0.9

*

Needs upload + upgrade?

1.0

1.1

1.1

1.0

*

Needs upgrade / WiP?

1.0

none

1.0

1.1

*

Invalid version

1.0

1.1

1.1

*

*

Needs upgrade / WiP?

and many more cases....

A different approach

If we have the following variables

we could try to define categories and conditions as follows:

Wishlist bugs

  1. <man-di> Tincho: 1) would it easy to show unstable and experimental for packages with versions in experimental?

<Tincho> man-di: uhm, seems sane. It is easy, but AFAICR there was no hardcoded suite, except for unstable. I should look into a way of doing it in a elegant way, but yes, count on that

  1. (for man-di too) having a max-ttl on watch data, using cached data with only a info msg when download fails, and upgrading to warn (and visible error in qareport) when max-ttl is reached?

Users of the scripts

See PET project page