[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Savannah-hackers-public] A public anonimized savane database for fu

From: Assaf Gordon
Subject: Re: [Savannah-hackers-public] A public anonimized savane database for future development
Date: Mon, 20 Mar 2017 00:38:39 +0000
User-agent: Mutt/1.5.23 (2014-03-12)


On Sun, Mar 19, 2017 at 05:15:28AM +0000, Tim Landscheidt wrote:
b) Having a script that deletes data sounds scary because
  Murphy will run it on the wrong database :-).

There is always such a risk, however I hope the chances are low.
This will have to be run very few times, and I run it as a user
that does not have write access to the real savane database (address@hidden).

  Also, to
  maintain privacy it's much easier to white-list data that
  is known to be suitable for publication than to delete
  everything else.

I humbly disagree with "easier". In the two and a half years
since I my last attempt, I did not find an easier (real world, practical) way to do it, and no one else came forward and suggested such way.

  I would (try to) rewrite that script to
  create a new database that replicates the schema and then
  copy selected tables and columns, applying filters where

The MySQL schema is available here:

If you can suggest a better script that takes one database (let's call it 'savane_agn') and creates a valid replicate with only 'whitelisted' data, I'll be happy to use it. You can use 'savane_demo' or 'savane_tim' as the target database in your scripts, I'll make sure it exists when I run it. The list of white-listed group IDs are in the SQL script I attached to the previous message in this thread.

Realistically, I'm not going to spend time on doing a white-list version.

I do take privacy very seriously, and I do not want to expose any user data. That is way I first make the files available only the savannah admins - and I'm happy to hear feedback.

As an intermediate step, I'm happy to share the files either with other gnu members (on fencepost), or privately with other volunters (e.g. you)
to have many people examinethe data.

- assaf

reply via email to

[Prev in Thread] Current Thread [Next in Thread]