| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
This does many meta-things, but changes no functionality:
- get rid of stack, and use just cabal with a stackage snapshot instead
(why did I ever think stack was a good idea?)
- update the stackage snapshot to something halfway recent
- thus making builds work on nixpkgs-23.05 (current stable)
- separating out packages into their own cabal files
- use the GHC2021 set of extensions as default
- very slight code changes to make things build again
- update readme accordingly
- stylish-haskell run
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
this removes:
- the bbb properties
- all explicit mentions of rc3
- the weird script domain hacks (done via a substitution now)
- some (few) of the weirder code choices
it also adds some more type level witchery to deal with configs, which
for some reason seems to be the hardest problem of this entire program …
also the server now does inter-assembly dependency checking!
|
| | |
|
| |
|
|
| |
also don't keep adjusted maps around if not necessary
|
| |
|
|
| |
as annoying as yaml is, cabal's package format is somehow worse, apparently
|
| | |
|
| |
|
|
|
|
|
| |
don't add maps to the result if their lint result looks the same as that
of another list, just say it happened several times instead (this leads
to a rather confusing Eq instance for MapResult which implements a very
simple surface-level equality — perhaps change that later)
|
| | |
|
| | |
|
| |
|
|
|
| |
(audio links now have their own allowlist, which won't put things
through the dereferrer)
|
| |
|
|
|
| |
(suggestions are shown only if they have a Damerau-Levenshtein distance
<= 4, which seems to yield reasonably good results)
|
| |
|
|
| |
as per yesterday's discussion about extended scripting variables
|
| | |
|
| |
|
|
|
| |
these now have their own top-level config attribute which is essentially a
shorthand for setting one that's deeper nested.
|
| | |
|
| |
|
