Completes the switch from Meson to CMake for the core build system in
Nix.
Meson was added originally because someone else had already done the
work for integrating it in Nix and it was an upgrade from the previous
setup.
However over time it became clear that Meson is not quite mature
enough for projects like Nix that have occasionally peculiar
configuration constraints.
Some issues encountered with Meson (some of these are due to the Meson
setup in Nix):
* Difficulty with generating correct compile_commands.json for
external tools like clangd
* Difficulty linking to libc++ when using clang
* Ugly shell invocations for certain parts of the build system (I want
these to be gone!!!)
This CMake setup mimics the Meson configuration, but there are some
differences (some temporary):
* headers are now included separately for each library (see a previous
commit that changes includes appropriately)
* autoheaders-style configuration is currently hardcoded. Before
blindly copying this I want to evaluate how much of it actually exists
for portability concerns that I don't have (such as support for OS
X).
* Nix is built with libc++ by default.
* [libstore] SQL schema is now inlined via a generated header, not an
included string literal
Abseil is still built as part of this build, rather than an external
dependency, because it chokes on differently configured compiler
invocations.
Note that because of the move to libc++ an unwanted behaviour is
introduced: glog log messages no longer have a body. I have yet to
debug what is going on there.
Previously all includes were anchored in one global mess of header
files. This moves the includes into filesystem "namespaces" (if you
will) for each sub-package of Nix.
Note: This commit does not introduce the relevant build system changes.
Because modules are not called via the default depot setup (for now
...), this introduces a dummy module that stores the depot tree itself
in the module configurations.
This makes it possible to write modules that use packages from the
depot.
This function was a custom (and inefficient in the case of
single-character delimiters) string splitter which was used all over
the codebase. Abseil provides an appropriate replacement function.
Replaces these functions with corresponding functions from Abseil,
namely absl::StripAsciiWhitespace and absl::SimpleAtoi.
In the course of doing this some minor things I encountered along the
way were also refactored.
This also changes the signatures of the various custom readFile
functions to use absl::string_view types.
It is considered bad form to use things from includes in headers, as
these directives propagate to everywhere else and can make it
confusing.
types.hh (which is includes almost literally everywhere) had some of
these directives, which this commit removes.
Suppose I have a path /nix/store/[hash]-[name]/a/a/a/a/a/[...]/a,
long enough that everything after "/nix/store/" is longer than 4096
(MAX_PATH) bytes.
Nix will happily allow such a path to be inserted into the store,
because it doesn't look at all the nested structure. It just cares
about the /nix/store/[hash]-[name] part. But, when the path is deleted,
we encounter a problem. Nix will move the path to /nix/store/trash, but
then when it's trying to recursively delete the trash directory, it will
at some point try to unlink
/nix/store/trash/[hash]-[name]/a/a/a/a/a/[...]/a. This will fail,
because the path is too long. After this has failed, any store deletion
operation will never work again, because Nix needs to delete the trash
directory before recreating it to move new things to it. (I assume this
is because otherwise a path being deleted could already exist in the
trash, and then moving it would fail.)
This means that if I can trick somebody into just fetching a tarball
containing a path of the right length, they won't be able to delete
store paths or garbage collect ever again, until the offending path is
manually removed from /nix/store/trash. (And even fixing this manually
is quite difficult if you don't understand the issue, because the
absolute path that Nix says it failed to remove is also too long for
rm(1).)
This patch fixes the issue by making Nix's recursive delete operation
use unlinkat(2). This function takes a relative path and a directory
file descriptor. We ensure that the relative path is always just the
name of the directory entry, and therefore its length will never exceed
255 bytes. This means that it will never even come close to AX_PATH,
and Nix will therefore be able to handle removing arbitrarily deep
directory hierachies.
Since the directory file descriptor is used for recursion after being
used in readDirectory, I made a variant of readDirectory that takes an
already open directory stream, to avoid the directory being opened
multiple times. As we have seen from this issue, the less we have to
interact with paths, the better, and so it's good to reuse file
descriptors where possible.
I left _deletePath as succeeding even if the parent directory doesn't
exist, even though that feels wrong to me, because without that early
return, the linux-sandbox test failed.
Reported-by: Alyssa Ross <hi@alyssa.is>
Thanks-to: Puck Meerburg <puck@puckipedia.com>
Tested-by: Puck Meerburg <puck@puckipedia.com>
Reviewed-by: Puck Meerburg <puck@puckipedia.com>
(cherry picked from commit c05e20daa1abb3446e378331697938b78af2b3d7)
Replaces the previous implementations which performed sorting with one
that instead walks through the map (which is already sorted) and
yields values from it.
This fixes a handful of language tests because the previous
implementation did not actually yield useful values on the new implementation.
In the change to the backing structure of attribute sets, the
requirement to manually balance the capacity of the structure went
away.
This is a) because Abseil's data structures manage this on their own,
and b) because the new Bindings class is allocated using `new (GC)`
rather than writing into a predefined memory area.
As part of this change functions related to the capacity were
deprecated and set to 0 values, which in turn caused the creation of
new attribute sets to return the same (mutable!) default value in
various cases, leading to "side effects" that caused evaluation
failures.
FWIW, I'm not sure if this optimisation had noticeable performance
impact, but while untangling libexpr it definitely doesn't help trying
to follow what it's doing - so bye, bye!
