Previously the memory would occasionally be collected during eval since
the GC doesn't consider the member variable as alive / doesn't scan the
region of memory where the pointer lives.
By using the traceable_allocator<T> allocator provided by Boehm GC we
can ensure the memory isn't collected. It should be properly freed when
SourceExprCommand goes out of scope.
Backported from:
d2c371927e
Change-Id: I1f7c745dbc66c7164bee50f4d9b0d437dbc7dd51
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1251
Tested-by: BuildkiteCI
Reviewed-by: isomer <isomer@tvl.fyi>
This is a blocker to enabling the linter.
Change-Id: I9f8d3cc19c7539086f53474a505362230fc56c04
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1245
Tested-by: BuildkiteCI
Reviewed-by: tazjin <mail@tazj.in>
This is a bad idea, it shouldn't exist, nixpkgs doesn't use it.
Change-Id: Ic4d1b936d8f059d5c40f0567af165b02427d7e36
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1241
Tested-by: BuildkiteCI
Reviewed-by: glittershark <grfn@gws.fyi>
These are the queries that are handled in the confusing case statement
in the old daemon implementation, because they have very similar
structure.
Change-Id: Ie7143354f66cef4336dff8072ede9a56271a7e89
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1228
Tested-by: BuildkiteCI
Reviewed-by: glittershark <grfn@gws.fyi>
What it does I will never know. It takes a path and returns a path.
Something happens in the middle.
Change-Id: I499a9df700e5b954c9aaf6d694753ff34e773dfd
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1210
Tested-by: BuildkiteCI
Reviewed-by: tazjin <mail@tazj.in>
These files were always implicitly depended upon by worker-protocol.hh, but just so happened also always be included in places where the other ones were already present.
We are likely getting rid of this file sooner rather than later, but in
the meantime this will cause clangd to shut up while I'm editing the
file.
Change-Id: I0d765f8b86828b6612e5483417e452221ea3c5b1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1207
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Add QueryValidPaths and QuerySubstitutablePaths, both of which filter a
list of paths based on a set of criteria.
Change-Id: I6aa4647efe82b82dc9582a311643d5f9b6d521d5
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1203
Tested-by: BuildkiteCI
Reviewed-by: tazjin <mail@tazj.in>
Replaces the previous uses of the (ordered!) std::map and std::set
with absl::flat_hash_{map|set}.
After some careful reading it seems that there is actually no
ordering dependency on these types, and the (drop-in) replacements
perform slightly better.
Overall this is not fixing a bottleneck, just a driveby thing.
Change-Id: Ided695dc75676bd58515aa9382df0be0a09c565e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1220
Reviewed-by: glittershark <grfn@gws.fyi>
Reviewed-by: isomer <isomer@tvl.fyi>
Tested-by: BuildkiteCI
Several definitions of functions declared in eval.hh were previously
implemented in parser.y, this moves them over to parser.cc.
While this still isn't a reasonable place to keep them, the long-term
fix is more likely to be that eval.hh needs to be split up.
Before we get to that point however, this already gives us the ability
to use tooling with this code.
Change-Id: If06fb655325fe281564047ffab0a0a640428a0ee
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1219
Tested-by: BuildkiteCI
Reviewed-by: isomer <isomer@tvl.fyi>
Moves several of the static helper functions into a new parser.cc
file.
Once the rest of the code is usefully extracted, these will be moved
to a private namespace.
Change-Id: I0d7b53dcefe31bb5c6bad3ad7f5fcb48276bf799
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1218
Tested-by: BuildkiteCI
Reviewed-by: isomer <isomer@tvl.fyi>
First step (of many?) towards extracting all the inline code from the
Yacc file and keeping it somewhere more accessible instead.
Note that none of this code has previously been touched by a linter or
formatter, pretty much ever, so as it is extracted it also undergoes
similar changes to the whole codebase after the initial fork.
Change-Id: If3b7181f22e3b3fd8c58dfa9befa7ee2896ea06d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1217
Tested-by: BuildkiteCI
Reviewed-by: isomer <isomer@tvl.fyi>
Instead of manually iterating over the two bindings to be combined,
this adds a new static method on the Bindings class which merges two
attribute sets by calling the range insertion operator over them.
In some anecdotal tests, this can lead to a ~10% speed bump -
depending on the specific operation.
Change-Id: I5dea03b0589a83a789d3a8a0fc81d0d9e6598371
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1216
Tested-by: BuildkiteCI
Reviewed-by: glittershark <grfn@gws.fyi>
Changes the derivation name & README overview to say "Tvix" instead of
"tazjix".
The previous name was mostly intended as a joke, and a way for me to
distinguish output paths. It's certainly not the intention to have a
portmanteau with my name here, especially now that several people are
contributing to the fork.
Change-Id: Icface5484d52355111eca23b2f6bd3b9e5567275
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1212
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
They're not mentioned anywhere on the client or the server.
Change-Id: Ia78c8bde49326f6bf69b4aeded083cfd3235131f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1188
Tested-by: BuildkiteCI
Reviewed-by: tazjin <mail@tazj.in>
On the chopping block this time:
* requests/responses that return one or more store paths, and contain
nothing else, have been changed to use the same types (StorePath and
StorePaths, respectively)
* QuerySubstitutablePathInfos has been added. It should be noted that
legacy Nix has two versions of this call, one that only queries a
single info (deprecated) and one that queries multiple. We have only
implemented the latter.
* QueryDerivationOutputs has been added.
Change-Id: Iccc9041e7064e141cf593467ecdcc327581c4056
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1186
Tested-by: BuildkiteCI
Reviewed-by: glittershark <grfn@gws.fyi>
Slight performance optimisation of nix::printString by copying chunks
of the input string which do not need escaping as contiguous blocks.
Paired-With: Perry Lorier <isomer@tvl.fyi>
Change-Id: I48bad90c8f2831ae4524c814a12b1982989922f9
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1184
Tested-by: BuildkiteCI
Reviewed-by: edef <edef@edef.eu>
Reviewed-by: glittershark <grfn@gws.fyi>
A significant fraction of all created attribute sets are empty; hence
this is an easy optimisation to make.
Paired-With: Perry Lorier <isomer@tvl.fyi>
Change-Id: I0884194d04c1ee95b2b239a253515f2152bc0856
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1179
Tested-by: BuildkiteCI
Reviewed-by: glittershark <grfn@gws.fyi>
Add tools.perf-flamegraph, which collects the base case execution of
perf piped through stackcollapse-perf and flamegraph to flamegraph the
execution of an external command via perf.
Change-Id: I671fe254dc374b6cd7deca2d3bdea266164de025
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1176
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Change-Id: If580b76b7b7170ce7840d3a97c3a547c29405a6f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1174
Tested-by: BuildkiteCI
Reviewed-by: Kane York <rikingcoding@gmail.com>
This reverts parts of the CLs splitting the backing implementation for
Bindings and moves back to only the BTreeMap-backed implementation.
Our evaluation has indicated that the Vector-backed implementation
does not match the performance of the plain array used upstream, and
in my view the complexity introduced by it is not worth the relatively
small (single-digit percentage) performance increase with a
pivot-point close to the number of attributes yielded by
stdenv.mkDerivation.
Going forward we will trial implementations of attribute sets backed
by HAMTs, and investigate other mechanisms of speeding up the language.
Some changes from the previous CLs are retained, for example the
removal of insert_or_assign and the passing of capacity.
Change-Id: I6eb4b075b453949583360755055c21a29d7ff642
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1172
Reviewed-by: glittershark <grfn@gws.fyi>
Tested-by: BuildkiteCI
Since one of the two implementations essentially uses the same shape
as the upstream Bindings, we backport their merge sort implementation
to ensure that we're doing the same thing semantically.
Paired-With: Luke Granger-Brown <git@lukegb.com>
Paired-With: Vincent Ambo <mail@tazj.in>
Paired-With: Perry Lorier <isomer@tvl.fyi>
Change-Id: I0d865897991eec0c4dd84d9bd0415cd1ca437792
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1162
Tested-by: BuildkiteCI
Reviewed-by: lukegb <lukegb@tvl.fyi>
Reviewed-by: glittershark <grfn@gws.fyi>
We accidentally returned the incremented iterator in the
post-increment, this fixes it.
Paired-With: Luke Granger-Brown <git@lukegb.com>
Paired-With: Vincent Ambo <mail@tazj.in>
Paired-With: Perry Lorier <isomer@tvl.fyi>
Change-Id: I36c79eb56359bb12a78ad3489e7d7d2eb2053510
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1140
Tested-by: BuildkiteCI
Reviewed-by: lukegb <lukegb@tvl.fyi>
Reviewed-by: glittershark <grfn@gws.fyi>
This function in never called, so let's just remove it
Paired-With: Luke Granger-Brown <git@lukegb.com>
Paired-With: Vincent Ambo <mail@tazj.in>
Change-Id: I79125866254d90dd0842bc86830d2103ac313cb6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1125
Tested-by: BuildkiteCI
Reviewed-by: lukegb <lukegb@tvl.fyi>
Reviewed-by: isomer <isomer@tvl.fyi>
To aid in making the decision of where to (currently just statically)
use a vector or btree as the backing implementation, add an extra
constructor argument to Bindings::NewGC for a capacity, and use
a (currently hardcoded at 32, for no good reason other than it felt like
a reasonable number) pivot to switch between our possible backing
implementations. Then, update all the call sites where it feels
reasonable that we know the capacity statically to *pass* that capacity
to the constructor.
Paired-With: Luke Granger-Brown <git@lukegb.com>
Paired-With: Vincent Ambo <mail@tazj.in>
Paired-With: Perry Lorier <isomer@tvl.fyi>
Change-Id: I1858c161301a1cd0e83aeeb9a58839378869e71d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1124
Tested-by: BuildkiteCI
Reviewed-by: lukegb <lukegb@tvl.fyi>
Reviewed-by: isomer <isomer@tvl.fyi>
Add an alternative impl of the now-abstract Bindings base class that is
backed by a std::vector, somewhat similar but stylistically a little
superior to the array-backed implementation in upstream nix. The
underlying iterator type in BindingsIterator is now backed by a
std::variant that we std::visit an overload over in order to implement
the various bits of the iterator interface.
Paired-With: Luke Granger-Brown <git@lukegb.com>
Paired-With: Vincent Ambo <mail@tazj.in>
Paired-With: Perry Lorier <isomer@tvl.fyi>
Change-Id: I7fbd1f4d5c449e2f9b82102a701b0bacd5e80672
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1123
Tested-by: BuildkiteCI
Reviewed-by: tazjin <mail@tazj.in>
To pave the way for the thing we want to do eventually which is use a
linear-time array for bindings (aka attribute sets) that are statically
known to be small enough to get a performance benefit from doing so,
make the Bindings class abstract, and define a BTreeBindings class that
inherits from it and is (currently always) returned from the static
initializer. The idea is that we'll have an ArrayBindings class as well
later that we can dispatch to conditionally based on an optional
"capacity" parameter or something like that.
There was some difficulty here in getting the iterator to work - the
approach we settled on ended up making a concrete BindingsIterator class
which will wrap a std::variant of either a btree iterator or something
else later, but right now just wraps a btree iterator.
Paired-With: Luke Granger-Brown <git@lukegb.com>
Paired-With: Vincent Ambo <mail@tazj.in>
Paired-With: Perry Lorier <isomer@tvl.fyi>
Change-Id: Ie02ca5a1c55e8ebf99ab1e957110bd9284278907
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1121
Tested-by: BuildkiteCI
Reviewed-by: isomer <isomer@tvl.fyi>
Having a default constructor for this causes a variety of annoying
situations across the codebase in which this is initialised to an
unexpected value, leading to constant guarding against those
conditions.
It turns out there's actually no intrinsic reason that this default
constructor needs to exist. The biggest one was addressed in CL/1138
and this commit cleans up the remaining bits.
Change-Id: I4a847f50bc90e72f028598196592a7d8730a4e01
Reviewed-on: https://cl.tvl.fyi/c/depot/+/1139
Reviewed-by: isomer <isomer@tvl.fyi>
Tested-by: BuildkiteCI
