Bug #7161: Support more than 24 HTTP mirrors
Enable the mirror pool dispatcher on all website pages that need it
This is about having our JS mirror pool dispatcher loaded and active on web pages that need it.
#9 Updated by intrigeri over 3 years ago
In https://mailman.boum.org/pipermail/tails-dev/2016-February/010262.html (#10284) we've decided to not do this, and instead use the fallback DNS round-robin pool for anyone not using DAVE.
#16 Updated by u over 3 years ago
- Status changed from Rejected to Confirmed
- Assignee set to intrigeri
- % Done changed from 100 to 60
I am reopening this. Some links on the website should use the mirror pool dispatcher: news/* (Call for testing) and the newly to be added back link to to the ISO and signature.
#25 Updated by intrigeri over 3 years ago
- Status changed from Confirmed to In Progress
- Assignee changed from u to intrigeri
Tested the branch locally as best as I could, and it seems to work :)
We now have two options:
- merge this into master, and have the JS use a
mirrors.jsonthat only has dl.amnesia.boum.org
- wait until #8635 is done and
mirrors.jsonpoints to the new mirrors pool
Let's think about it more.
#32 Updated by intrigeri over 3 years ago
Note that there's ongoing work on the IA, that may add more links that need to get the right CSS class, so before we merge this we'll need to do a last complete search for such links (and then, we'll need to communicate with the IA developers and doc writers, to let them know how links to dl.a.b.o should look like nowadays).
#37 Updated by u over 3 years ago
Depending on https://labs.riseup.net/code/issues/11405 we might want to include this script only on non IE browsers. Using conditional comments :
<!--[if !IE]><!--><script>our script</script><!--<![endif]-->
.. but this will not work in recent IE browsers as support for this kind of comment has been deprecated..https://msdn.microsoft.com/library/hh801214(v=vs.85).aspx So I'm not sure it's actually a good solution.
#44 Updated by intrigeri over 3 years ago
- % Done changed from 60 to 70
I've merged master again, and checked that any new dl.a.b.o or
stable_i386_iso_url occurrence had the right CSS class. Then I've tested locally, it works fine, modulo a bug in the JS (
reject() does not exist, so our various sanity checks are no-ops in practice) that we've just discovered. Once that bug is fixed, I'll test again and will send this to sajolida's plate for a last review, and I'll let him merge! :)