I’m not 100% sure this is wanted since it kind of makes everything
have to know about ca even if they don’t really want to. But it also
make things easier in dealing with looking up ca.
Originally, the test was only checking for different “real” storeDir.
That’s an easy case to handle, but the much harder one is if different
virtual store dirs are used. To do this, we need the SubstitutionGoal
to know about the ca, so it can recalculate the path to copy it over.
An important note here is that the store path passed to copyStorePath
needs to be one for srcStore - so that queryPathInfo works properly.
This also adds an error message when the store path from queryPathInfo
is different from the one we requested.
Substituters can substitute from one store dir to another with a
little bit of help. The store api just needs to have a CA so it can
recompute the store path based on the new store dir. We can only do
this for fixed output derivations with no references, though.
Most functions now take a StorePath argument rather than a Path (which
is just an alias for std::string). The StorePath constructor ensures
that the path is syntactically correct (i.e. it looks like
<store-dir>/<base32-hash>-<name>). Similarly, functions like
buildPaths() now take a StorePathWithOutputs, rather than abusing Path
by adding a '!<outputs>' suffix.
Note that the StorePath type is implemented in Rust. This involves
some hackery to allow Rust values to be used directly in C++, via a
helper type whose destructor calls the Rust type's drop()
function. The main issue is the dynamic nature of C++ move semantics:
after we have moved a Rust value, we should not call the drop function
on the original value. So when we move a value, we set the original
value to bitwise zero, and the destructor only calls drop() if the
value is not bitwise zero. This should be sufficient for most types.
Also lots of minor cleanups to the C++ API to make it more modern
(e.g. using std::optional and std::string_view in some places).
Without this, substitute info is fetched sequentially, which is
superslow. In the old UI (e.g. nix-build), we call printMissing(),
which calls queryMissing(), thereby preheating the binary cache
cache. But the new UI doesn't do that.
The fact that queryPathInfo() is synchronous meant that we needed a
thread for every concurrent binary cache lookup, even though they end
up being handled by the same download thread. Requiring hundreds of
threads is not a good idea. So now there is an asynchronous version of
queryPathInfo() that takes a callback function to process the
result. Similarly, enqueueDownload() now takes a callback rather than
returning a future.
Thus, a command like
nix path-info --store https://cache.nixos.org/ -r /nix/store/slljrzwmpygy1daay14kjszsr9xix063-nixos-16.09beta231.dccf8c5
that returns 4941 paths now takes 1.87s using only 2 threads (the main
thread and the downloader thread). (This is with a prewarmed
CloudFront.)
For one particular NixOS configuration, this cut the runtime of
"nix-store -r --dry-run" from 6m51s to 3.4s. It also fixes a bug in
the size calculation that was causing certain paths to be counted
twice, e.g. before:
these paths will be fetched (1249.98 MiB download, 2995.74 MiB unpacked):
and after:
these paths will be fetched (1219.56 MiB download, 2862.17 MiB unpacked):
Caching path info is generally useful. For instance, it speeds up "nix
path-info -rS /run/current-system" (i.e. showing the closure sizes of
all paths in the closure of the current system) from 5.6s to 0.15s.
This also eliminates some APIs like Store::queryDeriver() and
Store::queryReferences().
Also, move a few free-standing functions into StoreAPI and Derivation.
Also, introduce a non-nullable smart pointer, ref<T>, which is just a
wrapper around std::shared_ptr ensuring that the pointer is never
null. (For reference-counted values, this is better than passing a
"T&", because the latter doesn't maintain the refcount. Usually, the
caller will have a shared_ptr keeping the value alive, but that's not
always the case, e.g., when passing a reference to a std::thread via
std::bind.)
Not substituting builds with "preferLocalBuild = true" was a bad idea,
because it didn't take the cost of dependencies into account. For
instance, if we can't substitute a fetchgit call, then we have to
download/build git and all its dependencies.
Partially reverts 5558652709 and adds a
new derivation attribute "allowSubstitutes" to specify whether a
derivation may be substituted.
In particular this means that "trivial" derivations such as writeText
are not substituted, reducing the number of GET requests to the binary
cache by about 200 on a typical NixOS configuration.
So if a path is not garbage solely because it's reachable from a root
due to the gc-keep-outputs or gc-keep-derivations settings, ‘nix-store
-q --roots’ now shows that root.
If a derivation has multiple outputs, then we only want to download
those outputs that are actuallty needed. So if we do "nix-build -A
openssl.man", then only the "man" output should be downloaded.
Likewise if another package depends on ${openssl.man}.
The tricky part is that different derivations can depend on different
outputs of a given derivation, so we may need to restart the
corresponding derivation goal if that happens.
For example, given a derivation with outputs "out", "man" and "bin":
$ nix-build -A pkg
produces ./result pointing to the "out" output;
$ nix-build -A pkg.man
produces ./result-man pointing to the "man" output;
$ nix-build -A pkg.all
produces ./result, ./result-man and ./result-bin;
$ nix-build -A pkg.all -A pkg2
produces ./result, ./result-man, ./result-bin and ./result-2.
Getting substitute information using the binary cache substituter has
non-trivial latency overhead. A package or NixOS system configuration
can have hundreds of dependencies, and in the worst case (when the
local info cache is empty) we have to do a separate HTTP request for
each of these. If the ping time to the server is t, getting N info
files will take tN seconds; e.g., with a ping time of 0.1s to
nixos.org, sequentially downloading 1000 info files (a typical NixOS
config) will take at least 100 seconds.
To fix this problem, the binary cache substituter can now perform
requests in parallel. This required changing the substituter
interface to support a function querySubstitutablePathInfos() that
queries multiple paths at the same time, and rewriting queryMissing()
to take advantage of parallelism. (Due to local caching,
parallelising queryMissing() is sufficient for most use cases, since
it's almost always called before building a derivation and thus fills
the local info cache.)
For example, parallelism speeds up querying all 1056 paths in a
particular NixOS system configuration from 116s to 2.6s. It works so
well because the eccentricity of the top-level derivation in the
dependency graph is only 9. So we only need 10 round-trips (when
using an unlimited number of parallel connections) to get everything.
Currently we do a maximum of 150 parallel connections to the server.
Thus it's important that the binary cache server (e.g. nixos.org) has
a high connection limit. Alternatively we could use HTTP pipelining,
but WWW::Curl doesn't support it and libcurl has a hard-coded limit of
5 requests per pipeline.
Jörg Thalheim