It turns out the pg_format thing is just too slow for my use-cases
most of the time, even when pooling the mf. Most queries stay 90%+ in
the perl script, even though they are very fast to execute on their
own, screwing up the traces a lot.
So instead I replace the `postgres-simple` quasi-quoter that strips
whitespace (and tends to screw up queries anyway) with a simple one
that just removes the outer indentation up to the first line.
Why did I spend so much time on pg_format haha
Change-Id: I911cd869deec68aa5cf430ff4d111b0662ec6d28
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12138
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
Parked projects: I might pick these up again, but don’t want to
maintain them right now. Might delete later, or reinstate.
Change-Id: Ieb4f51dfaeba33c84820ca804b57c016e05bf566
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12137
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
Ideally there’d be a better generic abstraction of doing basic http
calls (with tracing) in the future, but for now just reexport.
Change-Id: Id7548739ea62e9172f2773f8db79fe726096b7f1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12136
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
There’s a bunch of duplication in how http client things are done,
let’s move that all to a single module.
Change-Id: Ic08c9bce49d562e4fa640a5bdfc15973a28a7bcb
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12135
Reviewed-by: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
Start of an effort to make the app work without javascript
enabled (graceful degradation yay).
We use a trick where buttons are nested into a form element, passing
their value as input; this should be better than depending on
`hx-vals`. If htmx is disabled, just redirect and reload the full page
instead of sending back the snippet. Probably depends on the use-case
of each snippet though.
Change-Id: I6c73e624c4bd29b1cbd5492b2f84f48102edc68b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12056
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
The redacted URL needs the torrent group id to go to the album page,
not the torrent id.
Also open the external link in a separate tab.
Change-Id: Ie6d73cb02137855bb388859d2800968cd56ece49
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11745
Reviewed-by: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
Autosubmit: Profpatsch <mail@profpatsch.de>
We’re using the redacted id, not the database id.
This is more stable.
Change-Id: I4d24e5bc6cc469d811cb1402a83157359576ef77
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11683
Autosubmit: Profpatsch <mail@profpatsch.de>
Reviewed-by: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
This took a while to figure out, but essentially sorting a thing with
one million queries takes a long time, as compared to doing it on a
small subset of things and then joining against the final ordering.
The generated column helps, too.
Change-Id: I1bf283e2be060748eebda92576e3d062c51a6777
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11681
Autosubmit: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
Generated columns!!
This reduces the query time for distinct on somewhat!
Change-Id: Ic004d327b719a9f44a224ec87709992ce3d8f160
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11676
Autosubmit: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
Simple artist album page that only shows albums the artist was
involved with.
Change-Id: Icff34afc6d1b39b6fb17765c1b3ea500dd4b4d95
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11675
Tested-by: BuildkiteCI
Autosubmit: Profpatsch <mail@profpatsch.de>
Reviewed-by: Profpatsch <mail@profpatsch.de>
The seeding weight would slow down the query quite a bit, so let’s
move it into a procedure and add an index onto the torrents table that
caches the result.
Baba’s first pl/SQL function!
Change-Id: I3bc6919b115c02b9c9aa74702fac0a8bbc66d2c1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11674
Autosubmit: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
The logic around transmission session handling was f*cked, this fixes
that.
We use an IORef instead of an MVar, since we want to unconditionally
write the new value. Even if multiple requests race, I *hope* that
transmission returns the same session id, otherwise we might get a
request loop. But it should be fine. (The semantics is not nicely
documented in the RPC docs.)
Additionally, log the session ids in the requests.
Change-Id: Id7d33f8cb74cb349e502331cad5eb5abe8a624cd
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11673
Autosubmit: Profpatsch <mail@profpatsch.de>
Reviewed-by: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
apparently the `torrent_group` json only contains one artist, while
the torrent itself contains a list of them.
This is important, because we need the artist id and the
`torrent_group` does not provide it, only the torrent one.
Change-Id: I3db45f454f14e89ea8c8dafba2065ecd55f5bcda
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11670
Autosubmit: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
We want the user thingy to see which error happened; it also gets
logged in the traces as before.
There’s another function which we should replace as well at one point.
Change-Id: I3d49edccd0e2088a45ac0138af9536b40dfa6848
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11660
Tested-by: BuildkiteCI
Autosubmit: Profpatsch <mail@profpatsch.de>
Reviewed-by: Profpatsch <mail@profpatsch.de>
The pool library would always take out the most recently used perl
resource again, and since that is the one that we just spawned, we’d
be back at square one. Instead, we try to find an older one (or up to
200ms old) and use that instead, because that should be the one with
the fastest response time.
Okay, that was enough bullshit lol.
Change-Id: I6b999e682d02ab03206a9d1b707edf16daa04a0d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11657
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
Autosubmit: Profpatsch <mail@profpatsch.de>
Instead of opening a separate span, we just add events for start and
end of formatting.
Change-Id: I26f6792dfdcd23c01cff415fa0f436d6a22d93fe
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11655
Autosubmit: Profpatsch <mail@profpatsch.de>
Reviewed-by: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
It does chip of the init overhead of like 50–100ms, even though the
formatting still takes quite some time (up to 200ms for more complex
expressions).
Maybe we need some simplistic formatter in the future that just splits
on parens? It’s not an easy problem …
Change-Id: I2ce951e6b3c2dc56294b1bdab913480727b50f0b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11654
Autosubmit: Profpatsch <mail@profpatsch.de>
Reviewed-by: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
When showing the best torrents table, we want to be able to filter for
“only downloaded”.
Change-Id: Ibfe9212f3d422d305f858fbef023ee985c1183d6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11650
Autosubmit: Profpatsch <mail@profpatsch.de>
Reviewed-by: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
The pretty renderer would add extra newlines, making the html
different.
Change-Id: I28496cbec61d4c9c63f657a499a1990f891949ee
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11639
Reviewed-by: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
Autosubmit: Profpatsch <mail@profpatsch.de>
Might have made a mistake when refactoring the cabal file.
Change-Id: Ic45ac9bf0272784ef77a2dfcc59ce2dbd421113c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11637
Autosubmit: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
Update all 3p/sources as we do normally except
- agenix which is still pinned to 0.15.0
- nixpkgs (unstable) which we bump to the HEAD of the staging-next
branch. This branch includes the downgrade of xz from 5.6.1 to
5.4.6 (https://github.com/nixos/nixpkgs/commit/d6dc19adbd). It
also includes the second haskell-updates rotation with GHC 9.6.4
which contains a few build fixes that seem to be required to get
our Haskell targets to work.
Note that this only reverts xz to a version that doesn't contain the now
known backdoor (CVE-2024-3094) which may or may not actually affect
NixOS. Additionally reverting to a version before the malicious
contributor's involvement may be difficult, but prudent:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1068024
Changes required by the updates:
- //3p/overlays/haskell:
- Update ihp-hsx to latest master to fix build with Stackage LTS 22.
- Update tmp-postgres to latest master to work around failure with
ansi-wl-pprint >= 1.
- Patch punycode for mtl >= 2.3.
- //users/Profpatsch:
- Clean up some warnings, mostly about unused dependencies
- my-prelude: Fix build with ghc-boot-9.6.4
- cas-serve: Use crypton over unmaintained cryptonite
- ical-smolify: skip in ci, iCalendar would require heavy patching to
work with Stackage LTS 22.
- //users/{wpcarro,aspen,flokli}:
Disable home-manager / nixos configuration builds that seem to have
transient failures that should disappear as we move away from
staging-next and closer to an actual channel release.
Change-Id: I5cca48e101041c3aedc1d9932dbca2cac885fcc1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11289
Tested-by: BuildkiteCI
Autosubmit: sterni <sternenseemann@systemli.org>
Reviewed-by: sterni <sternenseemann@systemli.org>
Reviewed-by: tazjin <tazjin@tvl.su>
First stab at factoring out handlers into a generalized handler
function.
This is still kind of confusing, but can be simplified later.
Change-Id: I42da047de83f6d489337d57059f85f793313443a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/11245
Autosubmit: Profpatsch <mail@profpatsch.de>
Reviewed-by: Profpatsch <mail@profpatsch.de>
Tested-by: BuildkiteCI
