Closed metp closed 4 years ago
Even though my implementation was using Lukas' suggested approach I would rather stick to Ralf's now as Patchwork recently allows to search for msgids via api. Lukas, your first use case is implementation independent from PaStA, the second one is also now satisfied when using pubins. It only requires updating Patchwork to 2.2.
Edit: Example: http://patchwork.ozlabs.org/api/patches/?msgid=200809050416.27831.adetsch%40br.ibm.com
Hi @rsarky, "Official way": https://patchwork.readthedocs.io/en/latest/development/installation/#import-mailing-list-archives. However this only works for mailing lists archives that use pipermail (e.g those on https://lists.ozlabs.org/listinfo/). https://www.kernel.org/lore.html uses public inboxes (pubins). You can write a bash script in case you want to import those (don't have mine anymore ). Either clone your desired pubin and let the script iterate trough the git repository. Every commit in pubin relates to one mail. So you could possibly create your own mbox file and import it with https://patchwork.readthedocs.io/en/latest/deployment/management/#parsearchive. Or you could try to use pubin's export to mbox option. Unfortunately only single threads are downloadable.
Thanks a lot for this @Honeybyte So currently what I have done is on say the alsa-devel archive I wanted to download the archive in mbox format. Instead of writing a script for converting the pubin repo into one, I kind of found a shorter way for now. I searched for all mails using the query https://lore.kernel.org/alsa-devel/?q=s%3A* in the archive. This gives an option to download the results in mbox form. Although I do think a script converting pubin repos to mboxes will be useful. Maybe it can be a binary in PaStA? As we have data structures for PubIn in Mbox.py in pasta.
Edit: imported patches into patchwork:
Even though my implementation was using Lukas' suggested approach I would rather stick to Ralf's now as Patchwork recently allows to search for msgids via api. Lukas, your first use case is implementation independent from PaStA, the second one is also now satisfied when using pubins. It only requires updating Patchwork to 2.2.
Okay, I got that. So if we want to support private mailboxes, where no public inbox repository exists. we would need to still need to create a public inbox repository somehow, either directly from the email account or by exporting from patchwork, and then we create a public inbox repository and provide that to pasta. Right?
@Honeybyte Did you also make some changes to some network configuration for docker on patchwork? Getting an error while connecting to port 8000 from PaStA's container. (connection refused
) Connecting from to port 8000 from host works though
Edit: Ignore. Fixed this. The fix was using the patchwork_default
container's IP address
Even though my implementation was using Lukas' suggested approach I would rather stick to Ralf's now as Patchwork recently allows to search for msgids via api. Lukas, your first use case is implementation independent from PaStA, the second one is also now satisfied when using pubins. It only requires updating Patchwork to 2.2.
Okay, I got that. So if we want to support private mailboxes, where no public inbox repository exists. we would need to still need to create a public inbox repository somehow, either directly from the email account or by exporting from patchwork, and then we create a public inbox repository and provide that to pasta. Right?
Yes. Hmm unfortunately not a nice routine...
I reckon this can be closed now, @rralf ?
Yes, thanks.
WIP: Add setup instructions
closes #9