Commit Graph
2464 Commits
Author SHA1 Message Date
zkldi d617060e6d feat: hacky, but legendary git log parser 2022-09-19 23:30:18 +01:00
zkldi dd22ecf8b3 refactor: move listing collections into its own fn 2022-09-19 23:29:33 +01:00
zkldi 7f7c48311e fix: have proper types on Environment.nodeEnv
it was bugging me because i couldn't remember whether it was
dev or development.
2022-09-19 23:29:18 +01:00
zkldi c39822570b feat: add local development middleware 2022-09-19 21:40:49 +01:00
zkldi 1f9d9e8a7f Merge pull request #675 from Bottersnike/feat/childprocess-exec-cwd 2022-09-19 01:06:28 +01:00
zkldi 806fc5789a fix: new Error in JS takes a string, not an object
future employers, please ignore this small mishap on part of the
contributor. Consider hiring them, they're very good at their job
and make excellent coffees.
2022-09-19 00:49:16 +01:00
zkldi b8f8305476 feat: allow running offline
i have no internet at the moment :)
2022-09-18 12:06:03 +01:00
zkldi c0ec92463d fix: just always use absolute urls in tables
raja i hate you
and you're bringing me down
2022-09-15 10:55:38 +01:00
zkldi 424c8de103 fix: invalid body.json pointer in sgl tables 2022-09-15 10:10:32 +01:00
zkldi 781ce9a289 fix: change invalid stub page message 2022-09-14 21:31:35 +01:00
zkldi 427c155b15 fix: only filter iidx versions in one place
this fixes a bug where casthour scores would be rejected in one place
but not the other.
2022-09-14 18:51:44 +01:00
zkldi 0f241e720e fix: remove more instances of instanceof
it's completely broken and no longer honoured. I *still* to this day do
not know why this is a bug on our systems. Maybe I will investigate it
further at some point, but for now, instanceof is banned in the codebase
as it doesn't seem to fucking work.

normally i refrain from swearing in a codebase i know future employers
will check, but this has to be one of *the* most frustrating bugs we've
had in years.
2022-09-13 23:28:30 +01:00
zkldi 95acde4505 fix: improve local binding explanation 2022-09-12 03:38:27 +01:00
zkldi 9358ddb63a style: fix lines-around-comment
this is a cadence bug and it will be fixed at some point
2022-09-12 02:49:15 +01:00
Bottersnike 61981c57b9 Modify database-seeds/repo.ts to not use cd 2022-09-12 02:46:15 +01:00
zkldi 8a2058fdb3 feat: allow casthour scores from kailua iidx 2022-09-09 21:50:22 +01:00
zkldi e6b6db16f7 fix: remove unused endpoint 2022-09-09 15:08:56 +01:00
zkldi b02ec4cd8f fix: support AssistEasy clears in raja IR 2022-09-07 11:00:23 +01:00
Christopher Cooper 73d483d24d add exscore to sdvx eamuse csv imports 2022-09-05 16:22:02 -04:00
zkldi 6e4986a91f Revert "fix: make symbols unique"
This reverts commit 73d2190f0e.

This actually results in table names in-game being displayed as
"sgl-EC EC ...", which sucks. Mistake on my part.
2022-09-05 00:07:30 +01:00
zkldi 3ce11d8844 fix: don't depend on reqTachiData for game
the only possibility for game here is "bms". we dont actually
have access to ReqTachiData here because of the bizarre
mounting stuff.
2022-09-04 20:38:21 +01:00
zkldi 73d2190f0e fix: make symbols unique
not sure what they're actually used for tbh.
2022-09-04 20:21:49 +01:00
zkldi 67485e9837 refactor: improve performance of song lookups
use lookup tables to optimise looking up songs; the former solution
was O(n^2), this one is O(1) with a small hit to memory.
2022-09-04 20:03:31 +01:00
zkldi c560df37d8 fix: formalise logged errmsgs 2022-09-04 20:00:24 +01:00
zkldi 7b174fb3ad docs: improve documentation for CreateAndMountTable 2022-09-04 19:51:56 +01:00
zkldi c1507eb94c fix: explicitly document gpt-specifity of content endpoints
we also don't want to take advantage of an express gimmick, which
is that it likes to prefer the most "specific" route to a path.

I'd rather be far more explicit and mount these routes first.
2022-09-04 19:48:38 +01:00
noelle 9c687c230e remove unused imports from test 2022-09-05 03:29:00 +09:00
noelle 9d8bf7f421 initial sieglinde ec/hc table support 2022-09-05 03:28:04 +09:00
zkldi 4939149abd feat: include our assets in localdev 2022-09-04 15:57:10 +01:00
zkldi 9c4afe2d8f fix: don't verify emails if EMAIL_CONFIG isn't set 2022-09-03 13:54:22 +01:00
zkldi 33263e0103 fix: don't have invite codes by default 2022-09-03 13:49:45 +01:00
zkldi 84387ba78b fix: improper logging of classes in score import 2022-08-31 01:48:52 +01:00
zkldi dd821e0a69 refactor: use tagged unions instead of instanceof
Instanceof appears to be somewhat broken in typescript. Still looking
into it.
2022-08-31 01:48:52 +01:00
zkldi eeb7d21602 fix: pin server ts target at es2020
I'm *pretty* sure it should be inheriting this from the parent tsconfig,
but it doesn't seem like it is. instanceof doesn't work if we're
compiling for ES3, and I think we might be doing that?
2022-08-30 13:43:59 +01:00
zkldi 47c893f1eb fix: set individualDifference on dp tiers
(it's invalid to not have this set, not sure why it wasn't)
2022-08-27 16:39:19 +01:00
zkldi 078a34afc2 feat: allow forcing local seeds via env var
Every-so-often we want to run a script that works against local seeds
rather than a remote. It's inconvenient to have a variable passed to all
of the times seeds are clones, so this env var allows this to be
overridden globally. Likely only for script usage; not even documented.
2022-08-27 16:37:56 +01:00
zkldi 5eb3c51849 feat: add importType to import context 2022-08-25 23:57:44 +01:00
zkldi bc1918454e test: update how POST /ir/barbatos checks scores 2022-08-25 23:48:56 +01:00
zkldi bf5828c5f2 feat: barbatos sdvx6 support 2022-08-25 23:39:17 +01:00
zkldi 5db86c29a3 fix: inherit prototypes, not classes 2022-08-25 17:12:38 +01:00
zkldi afe6e0ba84 fix: enforce prototypes in converter-failures
Typescript sometimes decides to "compile out" prototype chains like this
we have to *enforce* that these classes have the right inheritance,
because we do instanceof checks to determine what kind of error was
thrown. We could switch to a tagged union approach, but this seems more stable.
2022-08-25 14:40:19 +01:00
zkldi 05cc2ac02a feat: add migration script for removing jubeatNightScores 2022-08-25 13:09:08 +01:00
zkldi d7900638ba fix: defer discord transport submissions if we get 429d
This was giving us brownout problems. When we tried to restart all of
our things, only some of them would connect to discord. The rest
would be 429d, and not boot.

This stopped us from firing up score workers, amusingly.
2022-08-24 09:25:25 +01:00
zkldi 674568a915 docs: sync-state is inefficient, but we don't care 2022-08-22 05:07:39 +01:00
zkldi c509ab1c28 feat(server): add "recalc-everything" to package.json 2022-08-22 04:28:12 +01:00
zkldi 6463e39b0b test: fix faulty test with invites 2022-08-22 03:34:47 +01:00
zkldi 308245a2ee Merge pull request #642 from TNG-dev/zkldi/issue-637 2022-08-22 03:25:50 +01:00
zkldi 16b9ecc32c fix: DON'T DELETE THE LOCAL WORKING SEEDS
i've made this option even more obvious now given that i just messed it up
2022-08-22 02:33:53 +01:00
zkldi 877feb08b3 fix: dont ever delete local seeds 2022-08-22 02:08:35 +01:00
zkldi b72ffeec57 feat: support syncing database-seeds from local 2022-08-22 02:06:30 +01:00