More bindings that existed in the default i3 config but I had never
explicitly specified.
Change-Id: I57de0d3221afac299da9a09224564571037f67fa
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2245
Tested-by: BuildkiteCI
Reviewed-by: glittershark <grfn@gws.fyi>
An ec2 node I'm using as a remote dev box
Change-Id: I7d81371ecdc11d6c1b5bc06d1b4f55de534d25ad
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2244
Tested-by: BuildkiteCI
Reviewed-by: glittershark <grfn@gws.fyi>
These take precedence over cabal old-style aliases, since those aren't
used anymore
Change-Id: Iad1593bffb35d8f7b9c7df16c7c2da09a6ebb906
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2220
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
This is a step towards making the completing-read framework more
easily interchangeable (I'm eyeing selectrum).
Change-Id: I7a066e212a5384136defbba8f11ef9ed57abf22e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2240
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
I'm too tired for part 2.
Change-Id: Ic7058344806466276e3792e9ff9bbf660a18f672
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2239
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Solve time is like 36 hours, that must be a record!
Change-Id: I3713f033d83e6179a5d5fa7513952ee3864a6164
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2238
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This lets the parser collect multiple errors instead of returning
after the first one, with some optimistic synchronisation after
encountering something that looks wonky.
Change-Id: Ie9d0ce8de9dcc7a3d1e7aa2abe15f74cab0ab96b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2236
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
... with the exception of parenthesised expressions, because error
threading is not implemented yet.
Change-Id: I8d455d85e647548d5b71cbfd3d078f4970dab7fb
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2232
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This changes the structure of the output, too, where all AoC solutions
now end up in a big folder with `bin/day$n` executables.
Change-Id: I77928f4129489d06779b50059835925652688c9c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2231
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
... found this in the same place as hanebuschtag.txt
Change-Id: Iff5f4867b6c1d7685edae8a190489f3efb4890bc
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2229
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This old, legendary file should never be lost.
Change-Id: I2ae1607da24d684199c4136a880f56528d45a6e8
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2228
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Something I changed - perhaps removing lib.mkOptionDefault - caused
the *default* i3 keybindings (things like switching workspaces and
toggling the split layout which I had not added myself) not to get
added. This adds them back.
Change-Id: I301a346eabb4f7fc4499b60a4c26956fb1e08b0d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2218
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
This actually appears to be working this time!
Change-Id: I3e10eb7e1621a050e024b8b53313f13d44a999ae
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2217
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
This is a whole pile of things suggested by the internet to fix the
weird text flickering issues I've been seeing. upon first look it seems
like one of the kernel params (or all of them, or some combination of
them) fixed the issue.
Change-Id: Idc98902b46d4cba3bab367f6e22fb9ad10b26a26
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2216
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
Running docker is silly if I can't access it
Change-Id: I476915dacd44fac1ce4c533a84849fa6175d8107
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2215
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
At some point I should just convert this into an option that gets set
per-machine...but whatever.
Change-Id: I745fc5126469e887f8657e990d14a7e8b5085330
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2214
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
... mostly some AST boilerplate and a first top-level rule, plus
boilerplate similar to that set up in the Scanner.
Change-Id: I605d1de23c47a3b3702ab4f62cd3371bc3988c7d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2194
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Accidentally removed this everywhere when I disabled it for Darwin
Change-Id: Ia3480f1fbc6431a27da8c1de03bf0a66808f46b9
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2208
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
rustup and rust-analyzer for now, but this will likely also have shell
aliases and the like
Change-Id: I7838e537a72600410205e018a0c86be1493a9ffe
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2207
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
To avoid it clogging up git status in the depot
Change-Id: I102126dedb427d632679ee091aced6971495b8cc
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2206
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
My new work laptop, a dell XPS 13.
Change-Id: Ieab06622c9b280182025edfa63adf649e5fc70d8
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2205
Tested-by: BuildkiteCI
Reviewed-by: glittershark <grfn@gws.fyi>
Reviewed-by: lukegb <lukegb@tvl.fyi>
I want to use this as a prefix key instead
Change-Id: Idc865eb2edd1cbeff0b8a849232d98272c21ca8d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2204
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
... and show them to users, very crudely.
Change-Id: If4491b14db1124313f6ab7e5fbfdce9fea501d11
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2193
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Note that Lox does not support escapes, and I don't care about that.
Change-Id: Ie848cbc1164c4b005b15e29aad8fe723aaa68d1b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2190
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
The "swap-window" command swaps two windows, the 'source' and
'destination' window, exchanging the indices they occupy in the window
list. In at least tmux 2.8 (which is what's in Debian Stable), if the
window under focus is the source window, then focus follows that window
(i.e. the focus changes to the current window's new index atomically
with it being moved to that index).
In more recent tmux versions, this is not the case, and the focus
remains on the old index, so that the replacement window comes under
focus. The former behaviour is still possible, but must be explicitly
requested.
Change-Id: Ieff606dfc2624b869d3bb6e1344dd4d0c6301857
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2169
Reviewed-by: multi <depot@in-addr.xyz>
Tested-by: BuildkiteCI
... they're just noisy at the moment. This isn't complete because it
doesn't thread through scanner errors.
Change-Id: I0f75d2b20fa3f57be1af5d1d8aa8059856855825
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2162
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This makes it easier to work with the Unicode issue. The original
string representation can be discarded.
Change-Id: I740be4cb9654679ea7950f3899c5c709b1e7a739
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2160
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This was locked away in the urbint-specific module, but I use it
elsewhere.
Change-Id: Ifced2196dc22a9dbed74a18d4e1fed9488eb0e26
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2152
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
Update everything in home-manager to properly work on darwin (including
adding dobharchu as a top-level attribute from
//users/glittershark/home) and also fix font faces and sizes in emacs
config
Change-Id: Ica889dd212876030d5c2a916a71d8b614e6964f1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2147
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
This adds my standard inputrc to my home-manager configuration, via the
latter's readline module, and adds nix-shell detection to my bash prompt
string.
Additionally, nix-shell is wrapped in a shell function, so that entering
a nix-shell environment automatically runs bash as a subprocess where
appropriate, as the default environment spawned by nix-shell is not to
my liking.
Change-Id: Id5b1b9415a185ad9920f268c66de32d6ccc0b452
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2146
Reviewed-by: multi <depot@in-addr.xyz>
Tested-by: BuildkiteCI
This was only done to get an upgraded version, which is no longer
necessary with the nixpkgs bump but was also causing incompatibilities
with opengl
Change-Id: Ic398b4ac6caf24a1d40b7a4917b5053fba12d97a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2145
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
... still not that interesting, but at this point slightly divergent
from the book:
The book embraces mutability for interpreter state, initially for
tracking whether an error condition has occured.
I avoid this by instead defining an error type and collecting the
error values, to be handled later on.
Notes: So far nothing special, but this is just the beginning of the
book. I like the style it is written in and it has pointed to some
interesting resources, such as a 1965 paper titled "The Next 700
Languages".
Change-Id: I030b38438fec9eb55372bf547af225138908230a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2144
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
... as well as a Nix derivation, because why not.
Change-Id: Iaf2591ab72676fe0732c3f807b3aa0cff13fb4ef
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2143
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This is going to be the first of two interpreters from "Crafting
Interpreters".
Change-Id: I354ddd2357444648d0245f35d92176dd176525d8
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2142
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Mugwump is too unstable for such an important internet service
Change-Id: Ic714200ce5ce51f366777f538b4a6f443f010960
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2124
Tested-by: BuildkiteCI
Reviewed-by: tazjin <mail@tazj.in>
I previously removed my local package set from my HM config while the
latter was being made readTree compatible. Now that both the HM config
and the local package set can be built with readTree, I can re-enable
the locally-overridden htop package.
Change-Id: I77e20248c010bc7027e0b0a3164ec48d6ec29f31
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2132
Tested-by: BuildkiteCI
Reviewed-by: multi <depot@in-addr.xyz>
This adds readTree configuration for accessing my local package set,
and also adds these packages to the CI configuration.
Change-Id: Icd2d16e85859343902e73a466f3c6ba8d781537f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2131
Tested-by: BuildkiteCI
Reviewed-by: multi <depot@in-addr.xyz>
This reverts 1478317d149539d74fa4bad8414658fb7119ea07.
Using depot.depotPath in my home-manager configuration results in my
NIX_PATH and home-manager config file path being pointed at a copy of
the depot in the nix store, which makes building from my local depot
checkout a bit cumbersome.
Change-Id: Ib687d3e8147cb32df071d3c19a5300294ea62c0c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2130
Tested-by: BuildkiteCI
Reviewed-by: multi <depot@in-addr.xyz>
Add a new home-manager-compatible configuration file which loads the
common config attrset used by the readTree machinery into a structure
which the home-manager command line tool understands.
Garbage-collect the old home-manager configuration file used on whitby,
and update the HOME_MANAGER_CONFIG path to point at the new shim config.
Instead of having a per-environment HM configuration (not that I have
more than one environment), there's now a single configuration which
evaluates to an attrset of configurations, which can be loaded and built
using "home-manager build -A $attr".
Change-Id: Id8b35dc89aabffedf1a4dadfa0d3d4b914e4e2e7
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2129
Tested-by: BuildkiteCI
Reviewed-by: multi <depot@in-addr.xyz>
My home-manager config is not currently readTree compatible, which means
that it's not built by CI. This constructs a house of cards around
home-manager to make this buildable in CI.
Change-Id: I80480f24ff47347f46d708edbbf34d59fa76adac
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2123
Tested-by: BuildkiteCI
Reviewed-by: multi <depot@in-addr.xyz>
The depot knows where it is, not because it knows where it isn't, but
because it does an "import ./." at the top level and then makes this
path value available in the attrset passed to the rest of the tree.
My home-manager config on whitby previously involved manual
specification of the depot checkout location on whitby, however this
isn't necessary when the depot can already magically tell us where it
is.
Change-Id: I68577c8ef2cda6ba5bc46cf8f4821aac021c8066
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2122
Tested-by: BuildkiteCI
Reviewed-by: multi <depot@in-addr.xyz>
"let pkgs = import <nixpkgs> {}; in pkgs.home-manager.src" evaluates to
the source derivation for home-manager, however home-manager's configuration
machinery expects to be passed the store path of this derivation instead of the
derivation object itself.
Change-Id: I6b0ba3efaff9d080900349529576443192b058e9
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2121
Reviewed-by: multi <depot@in-addr.xyz>
Tested-by: BuildkiteCI
Included fixes for random breakage:
* 3p/awscli: pick from the stable channel; it is broken on unstable
* 3p/googletest: bumped version & removed patches that nixpkgs applies
* 3p/lisp/cffi: bumped library version for SBCL compat
* 3p/nix: fix libsystemd attribute
* 3p/nix: reformatted (clang-format handling of ternaries changed)
* glittershark/home: Use home-manager from nixkpgs
* glittershark/kernel: bumped linux-ck patch hash
* glittershark/kernel: removed "patch patch"
* multi/whitby: Use home-manager from nixpkgs
* tazjin/frog: drop Sourcetrail (it doesn't build currently)
Note that in addition to these changes, some previous CLs updated the
versions of git and cgit which was necessary for this channel bump,
but which could not be done in the same commit due to the nature of
the subtree merges.
Change-Id: If2563e8a68e2750c4b913a976ff7b93b42e8b7f3
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2110
Tested-by: BuildkiteCI
Reviewed-by: multi <depot@in-addr.xyz>
Reviewed-by: glittershark <grfn@gws.fyi>
Previously changed kernel versions would not cachebust the patch
download, because it would still be using the same SHA hash.
Forcing a different store path (by adding the version to the name)
also forces a redownload of the patch (and in turn cause the hash to
mismatch), avoiding this as a silent cause of failures in channel
updates.
Change-Id: I81a136ee2401126795cf042b0aadf2a1e7a707b4
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2114
Tested-by: BuildkiteCI
Reviewed-by: glittershark <grfn@gws.fyi>
- htop has moved upstreams, which has been producing new releases, so
update the derivation to pull from the new repository on GitHub.
- All of the patches I have locally have been merged upstream, so drop
them from the depot.
- Pull from a reasonably recent git commit instead of from a numbered
release, as the ZFS ARC stats and CPU meter columnation patches
haven't made it into a release yet.
Change-Id: I66ad4c035df07709abf4f75a9d4e1486920091d0
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2105
Reviewed-by: multi <depot@in-addr.xyz>
Tested-by: BuildkiteCI
for some reason installing it directly via nix doesn't work atm, so I
have this hack here
Change-Id: I45093633c35e756988078eb136c6e7bc3c532eea
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2078
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
This machine is significantly faster. Also, drop nixbuild, since the
transfer speed is too slow to make it worth it.
Change-Id: Ic14ef96e03a81dc429e4b4fec961c891dbb4b2b9
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2066
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
My old macbook, which I still use sometimes
Change-Id: I3ad080083b67b6cfab4cd31a4dce0a80a7227bd5
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2065
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
When posting a link to a PR created during an org headline, include the
name of the repository the PR was opened to in the link text.
Change-Id: I6c564aee3b098d3c6f96c7d7d609aa2638bc98e1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2063
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
There's just not enough juice in this machine to run more than one.
Change-Id: I6e6afc86337ca023e718023e4789fc29b6d8e175
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2059
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
Don't enable whitby+nixbuild as remote builders on every machine (eg not
mugwump), only chupacabra
Change-Id: I8aa8f20d76da4ec0d8caa64ef04697b7e76cbc03
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2058
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
Eventually this should be generalized, but for now this is good enough +
nice to have
Change-Id: Icca815b651cfb6f8f0cd2d6a1f64e56c63d2fef5
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2057
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
Install some packages and enable the necessary services + udev stuff to
make yubikeys usable
Change-Id: I8aee8a8b06895880c8195f02fb57b1216a5fdffc
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2049
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
In this case mostly so I can have it on mugwump
Change-Id: Ifa24caf607b30c1d034f4a9e7044ece88fcee38e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2048
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
Since buildkite is running on there, it'll be nice to be able to
download things. Obviously if this laptop ever becomes a laptop again
this'll have to go away (or just become the external domain)
Change-Id: I5fc49c061dbf79f8d523244bcf822e8d96fa6d42
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2047
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
I accidentally dropped this when reconfiguring things around to get
mugwump working, and when I rebuilt my x session turned off!
Change-Id: I252c90b6f4d796fef1f8183739fcc8dbfdd0fbf4
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2046
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
Having SSL on all the vhosts in nginx breaks the prometheus scraper with
the default config, since because it's targeting a different domain the
cert validation fails. It's pointing at localhost, so it's fine to just
have it not validate.
Change-Id: I1cbddc73335d4fa060115c253d69e27059a3113f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2045
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
Add a couple of buildkite agents, based off of the config we're using
for whitby (thanks!) for building my own projects that are closed
source.
Change-Id: I2c73538595002fdf4116f534dc9a5806f17e0558
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2044
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
The default size is a little bit large
Change-Id: I9f7096cdf9f9ba8433e6ead6124bdc04fc88877b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2043
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
The one I had downloaded before was empty, accidentally
Change-Id: I37c2ce8c556aa3f728ee675ccf14ace8416dcca4
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2042
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
