Commit graph

16094 commits

Author SHA1 Message Date
dependabot[bot] e0a7643086
Bump cachix/install-nix-action from 24 to 25
Bumps [cachix/install-nix-action](https://github.com/cachix/install-nix-action) from 24 to 25.
- [Release notes](https://github.com/cachix/install-nix-action/releases)
- [Commits](https://github.com/cachix/install-nix-action/compare/v24...v25)

---
updated-dependencies:
- dependency-name: cachix/install-nix-action
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
2024-01-15 22:35:12 +00:00
pennae 09a1128d9e don't repeatedly look up ast internal symbols
these symbols are used a *lot*, so it makes sense to cache them. this
mostly increases clarity of the code (however clear one may wish to call
the parser desugaring here), but it also provides a small performance
benefit.
2024-01-15 16:52:18 +01:00
pennae b596cc9e79 decouple parser and EvalState
there's no reason the parser itself should be doing semantic analysis
like bindVars. split this bit apart (retaining the previous name in
EvalState) and have the parser really do *only* parsing, decoupled from
EvalState.
2024-01-15 16:52:18 +01:00
pennae e1aa585964 slim down parser.y
most EvalState and Expr members defined here could be elsewhere, where
they'd be easier to maintain (not being embedded in a file with arcane
syntax) and *somewhat* more faithfully placed according to the path of
the file they're defined in.
2024-01-15 16:52:18 +01:00
pennae 835a6c7bcf rename ParserState::{makeCurPos -> at}
most instances of this being used do not refer to the "current"
position, sometimes not even to one reasonably close by. it could also
be called `makePos` instead, but `at` seems clear in context.
2024-01-15 16:52:18 +01:00
pennae 0076056164 move ParseData to own header, rename to ParserState
ParserState better describes what this struct really is. the parser
really does modify its state (most notably position and symbol tables),
so calling it that rather than obliquely "data" (which implies being
input only) makes sense.
2024-01-15 16:52:18 +01:00
pennae 1b09b80afa make parser utility functions members of ParseData
all of them need access to parser state in some way. make them members
to allow this without fussing so much.
2024-01-15 16:52:18 +01:00
pennae e8d9de967f simplify parse error reporting
since nix doesn't use the bison `error` terminal anywhere any invocation
of yyerror will immediately cause a failure. since we're *already*
leaking tons of memory whatever little bit bison allocates internally
doesn't much matter any more, and we'll be replacing the parser soon anyway.

coincidentally this now also matches the error behavior of URIs when
they are disabled or ~/ paths in pure eval mode, duplicate attr
detection etc.
2024-01-15 16:52:18 +01:00
pennae f07388bf98 remove ParserFormals
this is a proper subset of Formals anyway, so let's just use those and
avoid the extra allocations and moves.
2024-01-15 16:52:18 +01:00
John Ericson 8a2da82220
Merge pull request #9759 from obsidiansystems/abs-path-string-view
`absPath`: just take a `std::string_view`
2024-01-15 10:22:20 -05:00
Robert Hensing eeaa188973
Merge pull request #9748 from obsidiansystems/no-rapidcheck-headers
Simplify RapidCheck configure
2024-01-15 14:52:14 +01:00
John Ericson beed00c04e absPath: just take a std::string_view
1. Slightly more efficient

2. Easier to call

Co-authored-by: Cole Helbling <cole.e.helbling@outlook.com>
2024-01-15 08:21:54 -05:00
John Ericson 9b9ecdee34 Simplify RapidCheck configure
No more `RAPIDCHECK_HEADERS`!
2024-01-15 08:05:05 -05:00
John Ericson dd42a4e3e9 flake.lock: Update
Flake lock file updates:

• Updated input 'nixpkgs':
    'github:NixOS/nixpkgs/86501af7f1d51915e6c335f90f2cab73d7704ef3' (2024-01-11)
  → 'github:NixOS/nixpkgs/a1982c92d8980a0114372973cbdfe0a307f1bdea' (2024-01-12)
2024-01-15 08:04:46 -05:00
Robert Hensing 268c49264a
Merge pull request #9763 from L-as/avoid-unnecessary-copy
Avoid unnecessary copy of goal log
2024-01-15 08:32:03 +01:00
John Ericson 8d4b67912d
Merge pull request #9768 from obsidiansystems/nix-profile-test-no-outlink
`tests/functional/nix-profile.sh`: Add missing `--no-link`
2024-01-14 16:11:42 -05:00
John Ericson 03a6ca9b25 tests/functional/nix-profile.sh: Add missing --no-link
Otherwise we get a stray `tests/functional/result`, which can cause
spurious failures later.

(I got a failure because the test temp dir effecting the store dir
changed. This caused a test later because Nix didn't want to remove the
old `result` because it wasn't pointing inside the new Nix store.)
2024-01-14 15:27:44 -05:00
John Ericson 7ab0a4c39d
Merge pull request #9757 from obsidiansystems/parser-dodge-windows-macros
Avoid Windows macros in the parser and lexer
2024-01-14 13:36:13 -05:00
Las Safin f61d951909
Avoid unnecessary copy of goal log
The data was (accidentally?) copied into a std::string,
even though the string is immediately converted into a std::string_view.
The code has been changed to construct a std::string_view directly,
such that one copy less happens.
2024-01-13 19:32:37 +00:00
Cole Helbling bbcd9fcfc1 Arbitrarily bring back some nix-daemon calls
This means that both `nix daemon` and `nix-daemon` will be (somewhat)
tested.
2024-01-13 11:27:04 -08:00
Robert Hensing 7e5fa5c25c doc/glossary: Define package and package attribute set
A small step towards https://github.com/NixOS/nix/issues/6507

I believe this incomplete definition is one that can be agreed on.
It would be nice to define more, but considering that the issue
also proposes changes to the design, I believe we should hold off
on those.

As for the wording, we're dealing with some very general and vague
terms, that have to be treated with exactly the right amount of
vagueness to be effective.

I start out with a fairly abstract definition of package.
1. to establish a baseline so we know what we're talking about
2. so that we can go in and clarify that we have an extra, Nix-specific
   definition.

"Software" is notoriously ill-defined, so it makes a great qualifier
for package, which we don't really want to pin down either, because
that would just get us lost in discussion.
We can come back to this after we've done 6057 and a few years in a
desert cave.

Then comes the "package attribute set" definition.
I can already hear Valentin say "That's not even Nix's responsibility!"
and on some days I might even agree.
However, in our current reality, we have `nix-env`, `nix-build` and
`nix profile`, which query the `outputName` attribute - among others -
which just don't exist in the derivation.

For those who can't believe what they're reading:

    $ nix-build --expr 'with import ./. {}; bind // {outputName = "lib";}' --no-out-link
    this path will be fetched (1.16 MiB download, 3.72 MiB unpacked):
      /nix/store/rfk6klfx3z972gavxlw6iypnj6j806ma-bind-9.18.21-lib
    copying path '/nix/store/rfk6klfx3z972gavxlw6iypnj6j806ma-bind-9.18.21-lib' from 'https://cache.nixos.org'...
    /nix/store/rfk6klfx3z972gavxlw6iypnj6j806ma-bind-9.18.21-lib

and let me tell you that bind is not a library.

So anyway, that's also proof of why calling this a "derivation attrset" would be wrong, despite the type attribute.
2024-01-13 20:00:06 +01:00
Robert Hensing e838ac98d4 doc/glossary: Nix expression can be language expression 2024-01-13 19:42:05 +01:00
Qyriad cbd5553d57
doc: provide context in glossary definitions (#9378) 2024-01-13 12:20:08 +01:00
John Ericson e739a5002d Avoid Windows macros in the parser and lexer
`FLOAT`, `INT`, and `IN` are identifers taken by macros.

The name `IN_KW` is chosen to match `OR_KW`, which is presumably named
that way for the same reason of dodging macros.
2024-01-12 19:51:36 -05:00
John Ericson 381df7b9c9
Merge pull request #9756 from obsidiansystems/readline-on-windows
Newer Nixpkgs, get `readline` on Windows
2024-01-12 19:49:50 -05:00
John Ericson dd7e7b0a30 Newer Nixpkgs, get readline on Windows
Now `nix repl` an, in principle, work on that platform too.

Flake lock file updates:

• Updated input 'nixpkgs':
    'github:NixOS/nixpkgs/2c9c58e98243930f8cb70387934daa4bc8b00373' (2023-12-31)
  → 'github:NixOS/nixpkgs/86501af7f1d51915e6c335f90f2cab73d7704ef3' (2024-01-11)
2024-01-12 19:19:21 -05:00
John Ericson 3bc1f8f22f
Merge pull request #9752 from 9999years/hacking-filter-tests
Document unit tests in hacking.md
2024-01-12 19:00:19 -05:00
Robert Hensing 15f7bdaf27 CanonPath: Add rel_c_str()
Defensively because isRoot() is also defensive.
2024-01-12 22:55:37 +01:00
Rebecca Turner f73e50144f
Clarify ambiguity in nix-instantiate man page 2024-01-12 13:53:45 -08:00
Rebecca Turner 2d96c7a51f
Remove outdated reference to y in nix-instantiate man page 2024-01-12 11:27:31 -08:00
Rebecca Turner b29be1ff57
Document unit tests in hacking.md 2024-01-12 10:47:49 -08:00
John Ericson c58da62a06
Merge pull request #9737 from obsidiansystems/sys-error-split
Separate `SystemError` from `SysError`
2024-01-12 12:41:36 -05:00
Mel Zuser 1996105e91 added test for empty substring special case 2024-01-12 09:40:21 -08:00
John Ericson 6208ca7209 Separate SystemError from SysError
Most of this is a `catch SysError` -> `catch SystemError` sed. This
is a rather pure-churn change I would like to get out of the way. **The
intersting part is `src/libutil/error.hh`.**

On Unix, we will only throw the `SysError` concrete class, which has
the same constructors that `SystemError` used to have.

On Windows, we will throw `WinError` *and* `SysError`. `WinError`
(which will be created in a later PR), will use a `DWORD` instead of
`int` error value, and `GetLastError()`, which is the Windows equivalent
of the `errno` machinery. Windows will *also* use `SysError` because
Window's "libc" (MSVCRT) implements the POSIX interface, and we use it
too.

As the docs describe, while we *throw* one of the 3 choices above (2
concrete classes or the alias), we should always *catch* `SystemError`.
This ensures no matter how the implementation changes for Windows (e.g.
between `SysError` and `WinError`) the catching logic stays the same
and stays correct.

Co-Authored-By volth <volth@volth.com>
Co-Authored-By Eugene Butler <eugene@eugene4.com>
2024-01-12 12:00:33 -05:00
Mel Zuser 25c889baac Fix performance of builtins.substring for empty substrings
When returning a 0-length substring, avoid calling coerceToString,
since it returns a string_view with the string's length, which is
expensive to compute for large strings.
2024-01-12 08:37:04 -08:00
Robert Hensing 274d887fee fetchTree/git: Cache export-ignore filter 2024-01-12 17:18:56 +01:00
Robert Hensing d80c582b78 libfetchers: Add CachingFilteringInputAccessor
Co-authored-by: Eelco Dolstra <edolstra@gmail.com>
2024-01-12 17:16:59 +01:00
Eelco Dolstra 52f949bbf5
Merge pull request #9656 from edolstra/nix-profile-stable-names
Make profile element names stable
2024-01-12 16:59:36 +01:00
Eelco Dolstra 72560f7bbe Add profile migration test 2024-01-12 16:33:15 +01:00
Eelco Dolstra 8c7e2ed77c Update release notes 2024-01-12 16:21:07 +01:00
Robert Hensing f68ad5acbb fetchTree/git: Don't expose exportIgnore attr 2024-01-12 16:05:36 +01:00
Robert Hensing 469cf263c7 Format 2024-01-12 15:31:53 +01:00
Robert Hensing 692e9197bc fetchTree: Disallow combination of submodules and exportIgnore for now 2024-01-12 15:31:53 +01:00
Robert Hensing 71d08af15b rl-next: Add *general* note about git fetcher reimpl 2024-01-12 15:31:53 +01:00
Robert Hensing 99bd12f0b1 fetchGit/fetchTree: Improve exportIgnore, submodule interaction
Also fingerprint and some preparatory improvements.

Testing is still not up to scratch because lots of logic is duplicated
between the workdir and commit cases.
2024-01-12 15:31:53 +01:00
Robert Hensing 1bbe837184 fetchTree: Add isFetchGit exportIgnore 2024-01-12 15:31:53 +01:00
Robert Hensing 7774eff10e libfetchers/git: Move workdir accessor into GitRepo::getAccessor 2024-01-12 15:31:53 +01:00
Robert Hensing 8024b954d7 fetchTree: Recommend against exportIgnore 2024-01-12 15:31:53 +01:00
Robert Hensing 467c62a96e GitRepoImpl: Move exportIgnore into a filtering accessor 2024-01-12 15:31:53 +01:00
Robert Hensing cd5e752fa7 GitRepoImpl::getSubmodules: Access getSubmoduleRev without cast
This will be needed because the accessor will be wrapped, and therefore
not be an instance of GitInputAccessor anymore.
2024-01-12 15:31:53 +01:00