[Teammetrics-discuss] The Database Phase

[Sukhbir Singh]

Hello!

repository.update()

Changes:

+ I have done away with the username requirement and will go withe
Postgres method you suggested. For now, before running the script,
please append _user='<username>'_ to lines 168 and 391. I can't wait
to remove this weak hack, but for now we will let it go this way.

+ There was a flaw in the script that could have prevented lists from
being parsed. I was saving the checksums *downloading* the mbox
archives and not after *parsing* them. This has been fixed and luckily
I noticed it soon enough. So now the hash is written _after_ the mbox
has been parsed.

+ Table name has been changed to 'listarchives' instead of
'listarchive' (reflected in archives.sql).

+ Pushing the information into the database works completely, except
encoding errors and spam. We will tackle that soon, but tomorrow I am
going to get rid of permissions problem first (it's annoying).

And now, some fun facts! :

I).

liststat=# SELECT name, COUNT(name) FROM listarchives GROUP BY name;
     name      | count
---------------+-------
 Andreas Tille |     5
 Sukhbir Singh |     8
(2 rows)

II).

liststat=# SELECT name, SUM(msg_raw_len) FROM listarchives GROUP BY name;
     name      | sum
---------------+------
 Andreas Tille | 8224
 Sukhbir Singh | 7297
(2 rows)

So I think that saving the message body was a good idea; we will see
how it stacks up :-)