about summary refs log tree commit diff
AgeCommit message (Collapse)AuthorFilesLines
2021-03-21 r/2307 chore(tazjin/tverskoy): Enable SSH agent, Spotify, mullvad, ...Vincent Ambo1-0/+6
Just a few more missing things. Change-Id: I54b633296d91c6f8dce72d900d7d38b8af4ae419 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2625 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in>
2021-03-21 r/2306 feat(tazjin/tverskoy): Enable redshift & geoclue2Vincent Ambo1-3/+4
... also nuke that Bluetooth headset daemon, it doesn't work anyways. Change-Id: Iee884046a2bf2718767442ec62b2f731e14e39e8 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2624 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in>
2021-03-21 r/2305 feat(tazjin/tverskoy): Configure ephemeral home bind-mountsVincent Ambo1-1/+44
Uses the impermanence module to configure the bind-mounts for the few applications that are allowed to persist stuff between boots. For now this setup uses ZFS rollbacks to get /home back into a clean state, where the `tazjin-clean` snapshot is a partition with only the `/home/tazjin` folder and the correct partitions on it. This gives me enough scratch space to e.g. download stuff larger than my RAM without accumulating state. Change-Id: I7cdb2276f087ea62201690cb8b36ae074203f87c Reviewed-on: https://cl.tvl.fyi/c/depot/+/2623 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in>
2021-03-21 r/2304 chore(3p/gerrit_plugins): Build plugins separately in CIVincent Ambo1-2/+1
Change-Id: I76aeaa879e43b1fd08440f60a4f320bb7b48c6ab Reviewed-on: https://cl.tvl.fyi/c/depot/+/2630 Reviewed-by: sterni <sternenseemann@systemli.org> Tested-by: BuildkiteCI
2021-03-21 r/2303 chore(3p/gerrit): Update pinned dependency hashVincent Ambo2-3/+3
Change-Id: I1807ac1771d23b00e3a78cc36fb2f82a12900993 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2629 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2021-03-21 r/2302 fix(gs/emacs): Drop explicit load of slackGriffin Smith1-2/+0
This is getting loaded by use-package now Change-Id: I591629ed26ffac71a0df04d51c10b9290ebd76ff Reviewed-on: https://cl.tvl.fyi/c/depot/+/2428 Reviewed-by: glittershark <grfn@gws.fyi> Tested-by: BuildkiteCI
2021-03-21 r/2301 fix(tazjin/emacs): Ensure company is installedVincent Ambo1-0/+1
I don't know where this went. Change-Id: I82d27e273e38ffeffe683d6ff2236d383c7b1fcf Reviewed-on: https://cl.tvl.fyi/c/depot/+/2622 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in>
2021-03-21 r/2300 feat(tazjin/tverskoy): Configure impermanence moduleVincent Ambo1-0/+17
This module is responsible for mounting persistent storage into the ephemeral root disk of this machine. Currently only very basic configuration and data are retained, and hopefully that won't change. Change-Id: If800cbee60b7b3c5b8c457b9b332a0c05c33f20e Reviewed-on: https://cl.tvl.fyi/c/depot/+/2621 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in>
2021-03-21 r/2299 feat(tazjin/nixos): Initial check in of new host (tverskoy)Vincent Ambo3-1/+206
This is my new X13 AMD Thinkpad, on which many fun things will be done. Change-Id: I4de114a8c5ebb37d2f4844f407d2dc0e7cc9557e Reviewed-on: https://cl.tvl.fyi/c/depot/+/2620 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in>
2021-03-21 r/2298 fix(tazjin/emacs): Use ace-window from MELPA instead of ELPAVincent Ambo1-1/+1
This package definition is broken for ELPA in nixpkgs, but adisbladis is working on it. Change-Id: Ibe140a2daf7a3a294dae4c6340be33db05a47a44 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2619 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in>
2021-03-21 r/2297 feat(3p): Import nix-community/impermanence moduleVincent Ambo1-0/+12
Change-Id: I5b9386b1db74dafe921d7957bc726cfba403c155 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2618 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2021-03-20 r/2296 feat(gs/achilles): Implement very basic monomorphizationGriffin Smith12-19/+430
Implement very basic monomorphization, by recording type variable instantiations when typechecking Call nodes and then using those in a new hir Visitor trait to copy the body of any generic decls for each possible set of instantiation of the type variables. Change-Id: Iab54030973e5d66e2b8bcd074b4cb6c001a90123 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2617 Reviewed-by: glittershark <grfn@gws.fyi> Tested-by: BuildkiteCI
2021-03-20 r/2295 fix(gs/achilles): Get rid of universalization stepGriffin Smith2-30/+67
The step of "universalizing" function expressions was conflicting with top-level ascriptions for polymorphic function declarations: universalization generates universal type variables, and top-level polymorphic ascription *also* generates universal type variables, and the two were conflicting with each other when unifying. Let's just get rid of this now, and we can bring it back in a more principled manner once we do actual let-generalization (which there's still an ignored test case for) Change-Id: Idc08c8cb5ac92d1e6e1e63c9b8729176cab73f44 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2616 Tested-by: BuildkiteCI Reviewed-by: glittershark <grfn@gws.fyi>
2021-03-20 r/2294 feat(gs/achilles): Prefix top-level ascriptions with `ty`Griffin Smith2-2/+11
This makes parsing less ambiguous, which is nice (we can continue to not actually care about indentation!) and aligns nicely with `fn` for the declaration itself. Change-Id: Id48f064e2a1e01c5105297be355d0991b312b76d Reviewed-on: https://cl.tvl.fyi/c/depot/+/2615 Tested-by: BuildkiteCI Reviewed-by: glittershark <grfn@gws.fyi>
2021-03-20 r/2293 feat(gs/achilles): Implement extern decls, for glibc functionsGriffin Smith9-28/+147
Implement extern decls, which codegen to LLVM as forward-declared functions, and use these as a hook into calling glibc functions. We can print to the terminal now! The integration tests can test this now. Change-Id: I70af4546b417b888ad9fbb18798db240f77f4e71 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2614 Tested-by: BuildkiteCI Reviewed-by: glittershark <grfn@gws.fyi>
2021-03-20 r/2292 feat(gs/achilles): Codegen string literalsGriffin Smith1-1/+5
Codegen string literals to LLVM as (for now) global C string pointers Change-Id: I6dcd7fa25a7806a2f708a8e9275c9a01174fd0cf Reviewed-on: https://cl.tvl.fyi/c/depot/+/2613 Tested-by: BuildkiteCI Reviewed-by: glittershark <grfn@gws.fyi>
2021-03-20 r/2291 chore(gs/achilles): Integrate with the depot buildGriffin Smith5-10/+32
Get achilles building in Nix as part of the depot's build tree. This involved making it work with stable rust, since the depot only exposes stable rust to sub-packages, which turned out to be fairly straightforward. Also adds libffi as a new top-level expose, since it's required to build achilles Change-Id: I5f6dedb26c0b81ec258aedde1973e74903c07ece Reviewed-on: https://cl.tvl.fyi/c/depot/+/2612 Reviewed-by: sterni <sternenseemann@systemli.org> Reviewed-by: tazjin <mail@tazj.in> Tested-by: BuildkiteCI
2021-03-20 r/2290 chore(3p/exposed): Add glittershark to OWNERSGriffin Smith1-0/+1
Change-Id: I17474fbee75666d825b4c4db1af37fff1e214818 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2611 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in> Reviewed-by: sterni <sternenseemann@systemli.org>
2021-03-19 r/2289 fix(gs/xanthous): update xanthous.cabal, make CI catch such errorssterni3-13/+140
Since xanthous has a checked-in package.yaml and cabal file, the haskellPackages build infrastructure will use the package.yaml file for all builds. The resulting problem is that our CI won't actually catch build failures that would be observable with cabal or when building from the sdist. We fix that by filtering out the package.yaml file in pkg.nix additionally to the filters specified in .gitignore. For this we need gitignoreFilter from gitignore.nix which we expose as part of a functor set from third_party.gitignoreSource to maintain interface compatibility. Change-Id: I337185f484d2027341f38031dcd78898706904eb Reviewed-on: https://cl.tvl.fyi/c/depot/+/2609 Tested-by: BuildkiteCI Reviewed-by: glittershark <grfn@gws.fyi> Reviewed-by: tazjin <mail@tazj.in>
2021-03-19 r/2288 chore(gs/xanthous): fix for hgeometry 0.12.0.1sterni2-16/+2
The breaking removal of triangulationEdges was not that big of a deal after all: It was just renamed to edgesAsPoints apparently, so the fix is easy enough and we can save one override. hgeometry-combinatorial's doctests seem to trigger some kind of GHC dynamic linking bug (https://github.com/noinia/hgeometry/issues/132) so we disable the tests. Change-Id: Iba2a64cade4d1a55fa4b81846e1116f282d4590a Reviewed-on: https://cl.tvl.fyi/c/depot/+/2608 Tested-by: BuildkiteCI Reviewed-by: glittershark <grfn@gws.fyi>
2021-03-19 r/2287 fix(tazjin/emacs): Set notmuch custom variables through :customVincent Ambo2-5/+4
It seems like some of them end up file-local anyways, and the explicit call to `custom-set-variables` in mail-setup.el had seemingly no effect. Change-Id: Iad5011b5f8348b1ca5973813995c9644ac85ddf5 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2610 Reviewed-by: tazjin <mail@tazj.in> Tested-by: BuildkiteCI
2021-03-19 r/2286 fix(gs/xanthous): fix build failures caused by dependency updatessterni7-11/+50
The following changes in dependencies of xanthous broke the build and have been fixed in this CL. Thus we can reenable CI for xanthous. * random 1.2.0 removed the Read instance for StdGen, so we need use System.Random.Internal to un-newtype StdGen into an SMGen in the appropriate places as that type still has a Show and Read instance. Requires a new direct dependency on splitmix as well. * witherable 4.0 renamed Data.Witherable into Witherable and no longer exports Filter. * random 1.2.0 probably also broke the Function instance for GameState which contains a StdGen. I'm not exactly sure which change exactly triggered this, but the fix is easy enough: We implement a Function instance for SMGen using functionShow allowing us to write a Function instance for StdGen using functionMap. I've put these instances into Xanthous.Orphans. * hgeometry 0.12.0.0 removes the triangulationEdges function (which is also not mentioned in the changelog, so I'm not sure if there's a replacement yet). Fix by pinning to 0.11.0.0 for now. * hedgehog-classes: relax bounds on semirings Change-Id: I3617d8916d753b386c9fa80062be6bcbdfee0131 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2607 Tested-by: BuildkiteCI Reviewed-by: glittershark <grfn@gws.fyi>
2021-03-19 r/2285 chore(3p/haskell_overlay): remove obsolete overridessterni4-253/+2
I actually wanted to check up on regex-tdfa-text in owothia, but realized it was actually in a dependency. When porting the patch for chatter to nixpkgs, I wondered if we could get rid of other overrides or if we need to fix anything else in upstream. * aeson, attoparsec, cassava, psqueues, hedgehog: jailbreaks are no longer necessary * fgl, fgl-arbitrary: upstream has the versions pinned by now * hgeometry, hgeometry-combinatoral: upstream has moved past the pinned versions, but we don't need to keep them downgraded as xanthous's build is not broken by them. * random-source: the upstream compiler shouldn't crash anymore, additionally upstream has the version pinned here currently * semialign: upstream also has 1.1.0.1 by now * splitmix: splitmix has been fixed upstream and haskellPackages has moved past 0.1 * hspec-core: test suite passes or upstream has disabled it as well * QuickCheck: upstream advanced to the same version * vinyl: upstream moved past the pinned version, causes no build failures * comonad-extras: has been fixed upstream Change-Id: I34eff81ceaac005f2ad90dd9c1d3e623b8da91c0 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2606 Tested-by: BuildkiteCI Reviewed-by: glittershark <grfn@gws.fyi>
2021-03-19 r/2284 chore(3p): bump NixOS channels to 2021-03-18sterni5-11/+11
Making this a monthly service apparently. Necessary changes: * 3p: expose emacs27 instead of emacs26 which got removed users/tazjin/{camden, frog}: switch from emacs26 to emacs27 * 3p/lieer: google_api_python_client got renamed to google-api-python-client Change-Id: I1011665d10eebc99990addbef6a8a6b000b93896 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2605 Tested-by: BuildkiteCI Reviewed-by: glittershark <grfn@gws.fyi> Reviewed-by: tazjin <mail@tazj.in>
2021-03-15 r/2283 refactor(ops/dns): use drvTargets for meta.targets populationsterni1-7/+8
Since we have a dedicated util for this, we may as well use it to reduce code duplication. Change-Id: Ie52647be8c786d0b6a4dceb2fa6778b94625fafc Reviewed-on: https://cl.tvl.fyi/c/depot/+/2604 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in>
2021-03-15 r/2282 feat(users/Profpatsch): build attrset members on CIsterni11-43/+58
Setting meta.targets to include all derivations in the different package sets in Profpatsch's user folder makes them checked by CI until they do the readTree refactor as promised. To reduce code duplication we handle this in a simple function which is exposed from nix.utils which may be a good place for depot specific bits and bops we accumulate over time. To get around the issue of too nested sets we perform the following renames: * users.Profpatsch.tests gets moved into its own directory * users.Profpatsch.arglib.netencode now lives in its own file instead of the default.nix * users.Profpatsch.netstring.tests gets moved into its own directory Change-Id: Icd039c29d7760a711c1c53554504d6b0cd19e120 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2603 Tested-by: BuildkiteCI Reviewed-by: Profpatsch <mail@profpatsch.de>
2021-03-15 r/2281 feat(ops/dns): Configure tvl.su zoneVincent Ambo2-0/+30
Change-Id: I6016d92e9c231a257e06644dfcf44a4aaa12ac4d Reviewed-on: https://cl.tvl.fyi/c/depot/+/2601 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org> Reviewed-by: lukegb <lukegb@tvl.fyi>
2021-03-15 r/2280 feat(ops/dns): Import tvl.fyi DNS zone into depotVincent Ambo4-0/+59
Imports the current state of the tvl.fyi zone and configures simple CI checks on the file format. No deployment automation exists for this (yet?). Change-Id: Ia7d72e02b9f6d3adef994c5dc1898cc0df9dfcfb Reviewed-on: https://cl.tvl.fyi/c/depot/+/2600 Tested-by: BuildkiteCI Reviewed-by: glittershark <grfn@gws.fyi> Reviewed-by: sterni <sternenseemann@systemli.org>
2021-03-15 r/2279 merge(glittershark/achilles): Subtree import at 'b93268085a'Vincent Ambo33-0/+4508
Imported from https://github.com/glittershark/achilles/ git-subtree-dir: users/glittershark/achilles git-subtree-mainline: 4d193f239525954631ba9d789ca3aea9a2f4e14d git-subtree-split: b93268085aab14c80a400c299da5d04d2781098e Change-Id: I64a583b454bbe03e20358ad7808939a4cbc212ba
2021-03-14 Implement top-level ascription of declarationsGriffin Smith5-15/+73
2021-03-14 Universally quantified type variablesGriffin Smith17-112/+635
Implement universally quantified type variables, both explicitly given by the user and inferred by the type inference algorithm.
2021-03-14 Make string and bool parsing completeGriffin Smith2-5/+10
2021-03-14 Add string support to the frontendGriffin Smith10-11/+95
2021-03-13 Add the start of a hindley-milner typecheckerGriffin Smith20-78/+980
The beginning of a parse-don't-validate-based hindley-milner typechecker, which returns on success an IR where every AST node trivially knows its own type, and using those types to determine LLVM types in codegen.
2021-03-13 r/2278 feat(users/adisbladis): Add to usersAdam H1-0/+5
Change-Id: I2a3532605c602dd6ba44a6c723333db219a55907 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2599 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in>
2021-03-13 Allow exprs+bindings to optionally be ascriptedGriffin Smith6-38/+264
2021-03-13 Factor out expr parser into its own moduleGriffin Smith4-495/+510
2021-03-11 r/2277 feat(users/sterni/htmlman): hyperlink .Xr macro in outputsterni1-2/+32
We make use of the -O man=… option of mandoc(1) which allows to convert cross references via the .Xr macro into actual hyperlinks in the output. This can be disabled (by passing "none") or done in two modes: * all: links all .Xr cross references as if they were in $out/%N.%S.html. This will lead to broken links of course. * inManDir: only link to files in $out if the man page is found in manDir, use the template defined in linkXrFallback if not. all is the default, since we don't require all man pages to be in manDir, so it would be potentially confusing if the path attribute was used in the pages list. linkXrFallback uses the debian online man viewer by default currently, since it can be decently hyperlinked and debian has a lot of packages. Other options would be: * https://manpages.ubuntu.com/manpages/latest/en/man%S/%N.%S.html * https://man.archlinux.org/man/%N.%S.en * https://man.openbsd.org/%N.%S * https://www.man7.org/linux/man-pages/man%S/%N.%S.html Change-Id: I1363b9dfdda25cb7383c7310b8115c335444bd3d Reviewed-on: https://cl.tvl.fyi/c/depot/+/2597 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2021-03-11 r/2276 feat(users/sterni/htmlman): static site generator for manual pagessterni5-0/+291
htmlman is a very simple nix based static site generator which is intended for rendering HTML representations for man pages plus an index page listing all available pages. For the sake of simplicity (and unlike previous iterations of this piece of code) other documentation artifacts and formats are not supported. Usually web services like GitHub and depot's web interface are pretty good at displaying "normal" documentation artifacts like markdown files, but man pages are usually not rendered — with the additional problem that it's source is virtually unreadable. htmlman should provide a simple static site generator which can be plugged into GitHub actions or the like to automatically generate rendered version of man pages tracked in version control. Change-Id: Ib53292964b3ff84c32d70c5fde257a2edb8c2122 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2596 Tested-by: BuildkiteCI Reviewed-by: tazjin <mail@tazj.in> Reviewed-by: Profpatsch <mail@profpatsch.de> Reviewed-by: sterni <sternenseemann@systemli.org>
2021-03-08 Implement functions, both top-level and anonymousGriffin Smith10-125/+501
Implement both top-level and anonymous functions, but not closures in either case.
2021-03-07 Initial commitGriffin Smith24-0/+2316
2021-03-06 r/2275 fix(tazjin/emacs): Always show newest emails firstVincent Ambo1-0/+1
The default for this seems to have changed in a recent notmuch release. Change-Id: I1542b20c2e3edf72a3472c5277bce313c6df12b8 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2595 Reviewed-by: tazjin <mail@tazj.in> Tested-by: BuildkiteCI
2021-03-06 r/2274 feat(tazjin/rlox): Implement global variable accessVincent Ambo4-6/+48
This also includes a fix for an issue where the identifiers of variables were pushed onto the stack, which is incorrect. Change-Id: Id89b388268efad295f29978d767aa4b33c4ded14 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2594 Reviewed-by: tazjin <mail@tazj.in> Tested-by: BuildkiteCI
2021-03-06 r/2273 feat(tazjin/rlox): Implement global variable definitionVincent Ambo5-15/+80
identifier_str might look a bit overengineered, but we want to reuse this bit of code and it needs a reference to the token from which to pick the identifier. The problem with this is that the token would be owned by self, but the function needs to mutate (the interner), so this implementation is the most straightforward way of acquiring and working with an immutable reference to the token before interning the identifier. Change-Id: I618ce8f789cb59b3a9c5b79a13111ea6d00b2424 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2592 Reviewed-by: tazjin <mail@tazj.in> Tested-by: BuildkiteCI
2021-03-06 r/2272 refactor(tazjin/rlox): Refactor Compiler::consume into a macroVincent Ambo1-18/+21
Making this function a macro instead makes it possible to match arbitrary token kinds, even the ones that carry data, without changing the syntax too much. Change-Id: I5cda9e36d6833bd9c259f7d4d8340db6e783b4e8 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2593 Reviewed-by: tazjin <mail@tazj.in> Tested-by: BuildkiteCI
2021-03-06 r/2271 fix(tazjin/rlox): Resynchronise after panickingVincent Ambo1-1/+32
Change-Id: I60939f7a2c523b6ca1e9782e58c97959da38cfff Reviewed-on: https://cl.tvl.fyi/c/depot/+/2591 Reviewed-by: tazjin <mail@tazj.in> Tested-by: BuildkiteCI
2021-03-05 r/2270 feat(users/sterni/nix/utf8): pure nix utf-8 decodersterni3-0/+332
users.sterni.nix.utf8 implements UTF-8 decoding in pure nix. We implement the decoding as a simple state machine which is fed one byte at a time. Decoding whole strings is possible by subsequently calling step. This is done in decode which uses builtins.foldl' to get around recursion restrictions and a neat trick using builtins.deepSeq puck showed me limiting the size of the thunks in a foldl' (which can also cause a stack overflow). This makes decoding arbitrarily large UTF-8 files into codepoints using nix theoretically possible, but it is not really practical: Decoding a 36KB LaTeX file I had lying around takes ~160s on my laptop. Change-Id: Iab8c973dac89074ec280b4880a7408e0b3d19bc7 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2590 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2021-03-05 r/2269 feat(users/sterni/nix/flow): add switch conditionalsterni1-23/+53
switch would probably otherwise be called match, but has been renamed so it isn't confused with string.match and the enum matching capabilities yants has. It implements the closest to pattern matching nix can come which is still flexible enough to not be painful: Syntactically it works like cond, but is given a value. Instead of booleans it checks passed predicates or equality if simple values are passed. Both types of checks can be mixed. Change-Id: I40f000979cfd469316e15fd58d6c3a80312c1cc4 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2589 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2021-03-05 r/2268 feat(users/sterni/nix/fun): make lrs read left to right completelysterni1-1/+1
Change-Id: I57d290f770bc1d6bd88a46924889b919d68201e3 Reviewed-on: https://cl.tvl.fyi/c/depot/+/2588 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2021-03-05 r/2267 refactor(users/sterni/nix/string): don't calculate length for dropsterni1-4/+5
Since nix ends the substring at the end of the string anyways we can just statically use the largest nix integer as the length of the string. According to my testing this it ever so slightly faster as well. Change-Id: I64566e91c7b223f03dcebe3bc5710696dc4261bc Reviewed-on: https://cl.tvl.fyi/c/depot/+/2587 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>