| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Change-Id: Idd4ae78ef55891d89b72b5c2f3afc8b697b4b26e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7189
Reviewed-by: grfn <grfn@gws.fyi>
Reviewed-by: Adam Joseph <adam@westernsemico.com>
Tested-by: BuildkiteCI
Autosubmit: tazjin <tazjin@tvl.su>
|
|
This commit adds a markdown document which explains how the
thread-local VM infrastructure works, in case it is useful in the
future.
Change-Id: Id10e32a9e3c5fa38a15d4bec9800f7234c59234a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7193
Tested-by: BuildkiteCI
Autosubmit: Adam Joseph <adam@westernsemico.com>
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
This is a follow up to cl/7191 which neglected to adjust the
status.tvl.su.nix module and re-enable it.
Change-Id: Icc1917004cd50e5eab61a29bc68b393ba9bd6325
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7226
Tested-by: BuildkiteCI
Reviewed-by: lukegb <lukegb@tvl.fyi>
Reviewed-by: tazjin <tazjin@tvl.su>
Reviewed-by: grfn <grfn@gws.fyi>
|
|
Change-Id: Ib2f76310d24912a44efcabb0dd9eaf9a5d6cc6a9
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7225
Reviewed-by: grfn <grfn@gws.fyi>
Tested-by: BuildkiteCI
|
|
Profpatsch originally implemented an advisory checker from scratch in
Rust. We now ended up just using cargo-audit for the global checks
exposed via CI and the custom implementation is unused. To clean up
//tools/rust-crates-advisory a bit, we can move the unused parts to his
user directory.
Change-Id: Iacbd27c163edd07c804220fd1b3569c23aebd3e7
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7171
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
|
|
Change-Id: Ib6f4fd5817fb5415cff5ea1d8c75c8c9a08d56b4
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7185
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
We need a user in the context when we ask the groups backend to look up
groups by name, so for now if we don't have a _real_ user in the context
(such as during change indexing), then populate the context with the
anonymous user just for the duration of the groups backend calls.
Change-Id: If961d84fe57443cb95deb59628802658585ed1cb
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7172
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
Adds initial placeholders for builtins.{derivation,
unsafeDiscardStringContext}.
Change-Id: I67a126c9b9f9f4f11e2256e69b9a32ebd9eb1b0e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7187
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
|
|
This is more generally useful than just inside the VM, until it is
stabilised in Rust itself.
Change-Id: Id9aa3d5b533ff38e3d2c6b85ad484394fdd05dcf
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7186
Tested-by: BuildkiteCI
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: grfn <grfn@gws.fyi>
Reviewed-by: Adam Joseph <adam@westernsemico.com>
|
|
Reference the non-deprecated version of the grafana http port option in
the proxyPass for the nginx config of mugwump.
Change-Id: Ic7f370c7f7a451fe95a046d491d7b1cdf5f728cd
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7200
Reviewed-by: grfn <grfn@gws.fyi>
Autosubmit: grfn <grfn@gws.fyi>
Tested-by: BuildkiteCI
|
|
This fixes a mistake I made in d978b556e6.
Change-Id: I88db697105a7149e9785f6aface03bff68566d2b
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7085
Reviewed-by: grfn <grfn@gws.fyi>
Tested-by: BuildkiteCI
|
|
Scope_depth and with_stack_depth were being reset to zero for nested
function abstractions. Fortunately nothing depends on them being
computed correctly in these cases, but it sure was confusing.
Change-Id: I59980b6a5aff043f60079f97211220b0086eb97d
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7091
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
It is very confusing that this opcode is called DataLocalIdx, but it
carries a StackIdx rather than a LocalIdx. It seems like this
really ought to be called DataStackIdx, but maybe I've
misunderstood; if so please explain it to me.
Change-Id: I91f6ffa759412beef0b91d3c19ec0d873fe51b99
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7088
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
Uncomment and update the grafana config for whitby based on the new
config format that nixos accepts. I've validated this locally by
visually inspecting the resulting `ini` file, but not actually run it
yet.
Change-Id: I12d78ae48146e1b01bd2a4152276d4c6b16c1a3d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7191
Autosubmit: grfn <grfn@gws.fyi>
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
Lots of deprecations were made in the new nixos version for the grafana
config - this updates all of those settings in mugwump's system config
Change-Id: I69cdc9d2d59702c38d6334a4d27a04bef4e8c132
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7190
Reviewed-by: grfn <grfn@gws.fyi>
Autosubmit: grfn <grfn@gws.fyi>
Tested-by: BuildkiteCI
|
|
I've elected to split the check-all-our-lock-files script into two new
scripts: One very simple script which generates the report by invoking
lock-file-report on the fake lock file for //third_party/rust-crates and
all lock files in depot, and one which executes this and adds it as a
buildkite annotation if there are any warnings (which is reported by the
report generating script using a non zero exit code).
The latter script could become the basis for generalizing buildkite
annotations, a slight attempt at making it easily reusable in the future
has been made. So far we expect a report generating script to exit non
zero if a report should be made and to print commonmark to stdout. In
the future we may want to use a JSON format for generating the report,
allowing us to filter it by buildkite target (using the drvmap to
exclude certain reports, potentially).
Change-Id: I1df9e440509d69adff5b8e6304105a45dc62c018
Reviewed-on: https://cl.tvl.fyi/c/depot/+/5260
Reviewed-by: kn <klemens@posteo.de>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
I think migrating the execline scripts over to bash makes sense:
1. Ever since nixpkgs-fmt, execline scripts in depot have become a huge
pain to write and edit and I can't think of a satisfying solution to
this problem.
2. The scripts here require remembering things across loop cycles (i. e.
the status variable) which is not possible in pure execline. As a a
workaround we used to read the entire report into memory first and
check if it was empty (tying us to the argv limit for the report
length).
Change-Id: I954b08b982ef947f9014a685676d2b83a2aec4d2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/5259
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: I8c72405880a9342eb502d92c1e0087f5bb17e03c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7087
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
This implements builtins.split, and passes eval-okay-regex-split.nix
(which is moved out of notyetpassing).
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: Ieb0975da2058966c697ee0e2f5b3f26ccabfae57
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7143
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
|
|
This was broken in my blog for way too long.
Change-Id: I03c45c666d67006a4608a4b19d6167ab692e321d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/5905
Reviewed-by: wpcarro <wpcarro@gmail.com>
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: wpcarro <wpcarro@gmail.com>
Tested-by: BuildkiteCI
|
|
We have to be careful implementing `builtins.groupBy`, since the
list may contain thunks, and tvix's to_xxx() functions do not work
on thunks.
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: I182b6fc2d4296f864ed16744ef70b153e8e6978a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7039
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
The impl Display for NixAttrs needs to wrap double quotes around any
keys which are not valid Nix identifiers. This commit does that,
and adds a test (which fails prior to this commit and passes after
this commit).
Change-Id: Ie31ce91e8637cb27073f23f115db81feefdc6424
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7084
Autosubmit: Adam Joseph <adam@westernsemico.com>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
The variable name `local_idx` is used here for a StackIdx, which invites
confusion.
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: I2e22db90acdc0d29586ee5b72ea18d42d93badcb
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7086
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
Change-Id: I9ae66d54cc557e0f4dec7bd3b38022274b16263f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7169
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
The DNS server on the network of this boat refuses to deliver *any*
results to my laptop, only to my phone, so I had to do some trickery
to figure out the address of the WiFi login portal.
Change-Id: I7934c570be79d6191feb999e6860428623cfd88f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7168
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
If self.depth > other.depth then self is deeper than other, so self
is *below* other, not above it. Let's just inline the function.
Change-Id: I8dda3d90cbc86c8a6fa01bc4a5e506a2e403bd20
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7090
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
It isn't possible to implement PartialEq properly for Value, because
any sensible implementation needs to force() thunks, which cannot be
done without a `&mut VM`.
The existing derive(PartialEq) has false negatives, which caused the
bug which cl/7142 fixed. Fortunately that bug was easy to find, but
a silent false negative deep within the bowels of nixpkgs could be a
real nightmare to hunt down.
Let's just remove the PartialEq impl for Value, and the other
derive(PartialEq)'s that depend on it.
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: Iacd3726fefc7fc1edadcd7e9b586e04cf8466775
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7144
Reviewed-by: kanepyork <rikingcoding@gmail.com>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
I think we should bring this into $PATH too.
Change-Id: Ie31ac558355b7c4ed9dcd3dd60e1b03f141d1178
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7166
Autosubmit: flokli <flokli@flokli.de>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
Relatively minor bump, but makes sure we use a proper channel again
instead of a (not so) random commit on staging-next. Should include the
odd build fix or two as well.
Change-Id: Iefe7afdf173828edcb3e88263d27aba13ded691b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7167
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
If more of these issues crop up, we should consider making a patch for
our Nix fork that ignores the --experimental-features flag.
Change-Id: If146804f2c9f9375f2035985a6fd99f72b0fdcc7
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7165
Autosubmit: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
* //ops/machines/whitby: Disable grafana, since the grafana module was
changed upstream in a way that our configuration no longer works.
Since the OpenSSL security update is relatively pressing, adapting the
grafana configuration beforehand is not a hard requirement. See
https://github.com/NixOS/nixpkgs/pull/191768.
* //tools/depotfmt: keep Go at version 1.18 to forgo a reformat of the
tree.
* //nix/buildGo: keep Go at version 1.18, as 1.19 changed the CLI
interface (?) in a way that breaks buildGo.
* //3p/overlays/tvl: drop upstreamed tdlib upgrade.
* //3p/overlays/tvl: patch buf to work around breakage due to git 2.38.1
TODO items for Go are tracked in b/215.
Change-Id: Ie08fef49cf3db12e6b5225a8b992a990ddc5b642
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7141
Tested-by: BuildkiteCI
Autosubmit: sterni <sternenseemann@systemli.org>
Reviewed-by: grfn <grfn@gws.fyi>
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
In ripple, this is used for the internal on-disk format, but it's not
suitable for remote consumption.
Change-Id: I327361a2254566ac9216e23eaed36dba8fdd283b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7127
Tested-by: BuildkiteCI
Reviewed-by: edef <edef@edef.eu>
|
|
This is
0d4906cbabb183caa96e763671810fb39bd0c935:ripple/fossil/src/store.proto,
from https://src.unfathomable.blue.
It is not pulled in as a subtree, as some of the commits in there break
index-pack:
```
remote: error: object 2f487c3bf7cd8efd64f1d217edac732db98ff1c0: badDateOverflow: invalid author/committer line - date causes integer overflow
remote: fatal: fsck error in packed object
error: remote unpack failed: index-pack abnormal exit
```
Co-authored-by: edef <edef@unfathomable.blue>
Change-Id: I3369044090a3192e2322775a335887c37536a942
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7126
Tested-by: BuildkiteCI
Reviewed-by: edef <edef@edef.eu>
|
|
... until I get my self-hosting situation dealt with.
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: I44764862f754286249b90278f3932c1470e8214c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7146
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
|
|
Thank you, tpw_rules from #nixos IRC :)
Change-Id: Iebf07add10d96b38f318afaa175cfc0e35991df9
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7145
Reviewed-by: wpcarro <wpcarro@gmail.com>
Autosubmit: wpcarro <wpcarro@gmail.com>
Tested-by: BuildkiteCI
|
|
Example Usage:
```nix
{ pkgs, ... }:
(pkgs.writeText "foo" "bar").overrideAttrs (_: {
meta.ci.extraSteps.bar = {
label = "testing soft_fail";
softFail = [
{ exit_status = 253; }
{ exit_status = 251; }
];
command = pkgs.writeShellScript "fail" ''
# Trigger a soft failure by exiting 253 or 251
exit 253
'';
};
})
```
Buildkite Documentation: https://buildkite.com/changelog/56-command-steps-can-now-be-made-to-soft-fail
Change-Id: Idfce15a8d9cde568aa24ebd03956ada399bba5d2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7076
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
The current implementation of nix_eq will force one level of thunks
and then switch to the (non-forcing) rust Eq::eq() method. This
gives incorrect results for lists-of-thunks.
This commit changes nix_eq() to be recursive.
A regression test (which fails prior to this commit) is included.
This fix also causes nix_tests/eval-okay-fromjson.nix to pass, so it
is moved out of notyetpassing.
Change-Id: I655fd7a5294208a7b39df8e2c3c12a8b9768292f
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7142
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
We're getting close to the finish line, folks.
I went through the list of builtins and there are only 33 that
remain unimplemented. I've marked them, and indicated which are
ready to be implemented vs which are waiting for other things.
We can delete this column from the table once everything is
implemented.
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: Idfaef93283536288b12e59aef5c3e1cd139bd133
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7140
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
I believe that the currentTime, findFile, hashFile, pathExists,
readDir, path (unless ?sha256), and readFile builtins are impure.
This commit marks them as such in docs/builtins.md.
Change-Id: Ib1b59fe643dde73cb2b00050b4ef9d3401ad22eb
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7139
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
Change-Id: I59a0756940d1e5360a2ab4e886cf0bc9af7b8901
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7133
Autosubmit: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
This is a bit tricky because the comparator can throw errors, so we
need to propagate them out if they exist and try to avoid sorting
forever by returning a reasonable ordering in this case (as
short-circuiting is not available).
Co-Authored-By: Vincent Ambo <tazjin@tvl.su>
Change-Id: Icae1d30f43ec1ae64b2ba51e73ee467605686792
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7072
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
|
|
Lists are compared lexicographically in C++ nix as of [0], and our
updated nix test suites depend on this. This implements comparison of
list values in `Value::nix_cmp` using a very similar algorithm to what
C++ does - similarly to there, this requires passing in the VM so we can
force thunks in the list elements as we go.
[0]: https://github.com/NixOS/nix/commit/09471d2680292af48b2788108de56a8da755d661#
Change-Id: I5d8bb07f90647a1fec83f775243e21af856afbb1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7070
Autosubmit: grfn <grfn@gws.fyi>
Reviewed-by: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
|
|
CL/7034 looks great, except that for a length-N target string it
will perform N deep copies of each of the from and to-lists. Let's
use references instead of clones.
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: Icd341213a9f0e728f9c8453cec6d23af5e1dea91
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7095
Reviewed-by: wpcarro <wpcarro@gmail.com>
Reviewed-by: j4m3s <james.landrein@gmail.com>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
Change-Id: I93dcdaeb101364ee2273bcaeb19acb57cf6b9e7d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7034
Autosubmit: j4m3s <james.landrein@gmail.com>
Reviewed-by: Adam Joseph <adam@westernsemico.com>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
I assumed that AttrsRep::KV represented attrsets with a single
attribute as a Key-Value pair. That is not the case. Let's warn
other people about this.
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: Ie3d2765fcc1ab705c153ab94ffe77bbd6d4ab39e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7093
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: I460230863de853ca5248733bc977d4780b216f36
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7096
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
|
|
:)
Change-Id: Ib6a5990551cb0c86d0fc7f9a4121e3a6613ec6a1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7125
Reviewed-by: wpcarro <wpcarro@gmail.com>
Autosubmit: wpcarro <wpcarro@gmail.com>
Tested-by: BuildkiteCI
|
|
Rustc uses wasm32-unknown-unknown, which is rejected by config.sub,
for wasm-in-the-browser environments. Rustc should be using
wasm32-unknown-none, which config.sub accepts. Hopefully the rustc
people will change their triple before stabilising this triple. In
the meantime, we fix it here in order to unbreak tvixbolt.
https://doc.rust-lang.org/beta/nightly-rustc/rustc_target/spec/wasm32_unknown_unknown/index.html
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: I941fd8d6f3db4e249901772fd79321ad88cd9cc6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7107
Reviewed-by: tazjin <tazjin@tvl.su>
Reviewed-by: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
|
|
This commit contains two search-and-replace renames which are broken
out from I04131501029772f30e28da8281d864427685097f in order to
reduce the noise in that CL:
- `is_thunk -> is_suspended_thunk`, since there are now
OpThunkClosure and OpThunkSuspended
- `compile_lambda_or_thunk` -> `compile_lambda_or_suspension`
Change-Id: I7cc5bbb75ef6605e3428c7be27e812f41a10c127
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7037
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
CL/6867 added support for builtins.import, which required a cyclic
reference import->globals->builtins->import. This was implemented
using a RefCell, which makes it possible to mutate the builtins
during evaluation. The commit message for CL/6867 expressed a
desire to eliminate this possibility:
This opens up a potentially dangerous footgun in which we could
mutate the builtins at runtime leading to different compiler
invocations seeing different builtins, so it'd be nice to have
some kind of "finalised" status for them or some such, but I'm not
sure how to represent that atm.
This CL replaces the RefCell with Rc::new_cyclic(), making the
globals/builtins immutable once again. At VM runtime (once opcodes
start executing) everything is the same as before this CL, except
that the Rc<RefCell<>> introduced by CL/6867 is turned into an
rc::Weak<>.
The function passed to Rc::new_cyclic works very similarly to
overlays in nixpkgs: a function takes its own result as an argument.
However instead of laziness "breaking the cycle", Rust's
Rc::new_cyclic() instead uses an rc::Weak. This is done to prevent
memory leaks rather than divergence.
This CL also resolves the following TODO from CL/6867:
// TODO: encapsulate this import weirdness in builtins
The main disadvantage of this CL is the fact that the VM now must
ensure that it holds a strong reference to the globals while a
program is executing; failure to do so will cause a panic when the
weak reference in the builtins is upgrade()d.
In theory it should be possible to create strong reference cycles
the same way Rc::new_cyclic() creates weak cycles, but these cycles
would cause a permanent memory leak -- without either an rc::Weak or
RefCell there is no way to break the cycle. At some point we will
have to implement some form of cycle collection; whatever library we
choose for that purpose is likely to provide an "immutable strong
reference cycle" primitive similar to Rc::new_cyclic(), and we
should be able to simply drop it in.
Signed-off-by: Adam Joseph <adam@westernsemico.com>
Change-Id: I34bb5821628eb97e426bdb880b02e2097402adb7
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7097
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
