Commit graph

90 commits

Author SHA1 Message Date
Eelco Dolstra 0a84137c45 * checkVarDefs: don't check in closed terms, which don't have
undefined variables by definition.  This matters for the
  implementation of "with", which does a call to checkVarDefs to see
  if the body of the with has no undefined variables.  (It can't be
  checked at parse time because you don't know which variables are in
  the "with" attribute set.)  If we check closed terms, then we check
  not just the with body but also the substituted terms, which are
  typically very large.  This is the cause of the poor nix-env
  performance on Nixpkgs lately.  It didn't happen earlier because
  "with" wasn't used very often in the past.

  This fix improves nix-env performance roughly 60x on current Nixpkgs.
  nix-env -qa is down from 29.3s to 0.5s on my laptop, and nix-env -qa
  --out-path is down from 229s to 3.39s.  Not bad for a 1-line fix :-)
2008-02-21 12:01:24 +00:00
Eelco Dolstra 451dbf687f * nix-env now maintains meta info (from the `meta' derivation
attribute) about installed packages in user environments.  Thus, an
  operation like `nix-env -q --description' shows useful information
  not only on available packages but also on installed packages.

* nix-env now passes the entire manifest as an argument to the Nix
  expression of the user environment builder (not just a list of
  paths), so that in particular the user environment builder has
  access to the meta attributes.
  
* New operation `--set-flag' in nix-env to change meta info of
  installed packages.  This will be useful to pass per-package
  policies to the user environment builder (e.g., how to resolve
  collision or whether to disable a package (NIX-80)) or upgrade
  policies in nix-env (e.g., that a package should be "masked", that
  is, left untouched by upgrade actions).  Example:

  $ nix-env --set-flag enabled false ghc-6.4
2007-02-02 01:52:42 +00:00
Eelco Dolstra f23dcdd603 * Canonicalise ASTs in `nix-instantiate --eval': remove position
info, sort attribute sets.
2007-01-13 16:17:07 +00:00
Eelco Dolstra dd300fb48d * Some better error messages. 2006-10-23 16:45:19 +00:00
Eelco Dolstra d7efd76394 * Big cleanup of the semantics of paths, strings, contexts, string
concatenation and string coercion.  This was a big mess (see
  e.g. NIX-67).  Contexts are now folded into strings, so that they
  don't cause evaluation errors when they're not expected.  The
  semantics of paths has been clarified (see nixexpr-ast.def).
  toString() and coerceToString() have been merged.

  Semantic change: paths are now copied to the store when they're in a
  concatenation (and in most other situations - that's the
  formalisation of the meaning of a path).  So

    "foo " + ./bla

  evaluates to "foo /nix/store/hash...-bla", not "foo
  /path/to/current-dir/bla".  This prevents accidental impurities, and
  is more consistent with the treatment of derivation outputs, e.g.,
  `"foo " + bla' where `bla' is a derivation.  (Here `bla' would be
  replaced by the output path of `bla'.)
2006-10-16 15:55:34 +00:00
Eelco Dolstra 7d4567f2cc * Removed URIs from the evaluator (NIX-66). They are now just another
kind of notation for strings.
2006-10-11 21:59:33 +00:00
Eelco Dolstra d20c3011a0 * toFile: added an additional argument to specify the store path
suffix, e.g., `builtins.toFile "builder.sh" "..."'.
* toFile: handle references to other files correctly.
2006-10-03 14:55:54 +00:00
Eelco Dolstra fc195519b5 * Sone missing #includes. 2006-09-05 08:54:48 +00:00
Eelco Dolstra 75068e7d75 * Use a proper namespace.
* Optimise header file usage a bit.
* Compile the parser as C++.
2006-09-04 21:06:23 +00:00
Eelco Dolstra 1854f84e83 * Fix a few warnings. 2006-08-04 17:07:13 +00:00
Eelco Dolstra c11839d7b2 * `nix-instantiate --print-args': print out the valid values for
functions arguments that have a domain.
2006-07-28 14:01:29 +00:00
Eelco Dolstra f4c5531d92 * New language feature: domain checks, which check whether a function
argument has a valid value, i.e., is in a certain domain.  E.g.,

    { foo : [true false]
    , bar : ["a" "b" "c"]
    }: ...

  This previously could be done using assertions, but domain checks
  will allow the buildfarm to automatically extract the configuration
  space from functions.
2006-07-24 16:35:34 +00:00
Eelco Dolstra 57751fdb55 * Refactoring to support domain checks. 2006-07-24 15:16:03 +00:00
Eelco Dolstra 4f3725b167 * Better error messages (especially wrt types). 2006-07-19 15:36:15 +00:00
Eelco Dolstra e10b830251 * Doh! Of couse we cannot memoize across scopes. 2006-07-11 10:29:52 +00:00
Eelco Dolstra 0832956089 * Use the new ATermMap. 2006-05-04 12:21:08 +00:00
Eelco Dolstra 68174bdc7d * Use a linked list of substitutions. This reduces the amount of
copying.
2006-05-02 21:39:02 +00:00
Eelco Dolstra c791e94aee * Removed a bunch of ATreverses. 2006-05-02 17:51:50 +00:00
Eelco Dolstra 11ae2d1e7a * Memory reduction: replaced expensive calls to ATmakeApplList by
ATmakeApplArray, and got rid of ATreverse in substitute().
2006-05-02 14:07:28 +00:00
Eelco Dolstra dc719e6ba5 * Some preliminaries towards NIX-45. 2006-05-02 13:39:55 +00:00
Eelco Dolstra 2b3b6c9b34 * In theory, this should reduce the number of ATermMap
re-allocations.
2006-03-10 16:14:13 +00:00
Eelco Dolstra 5bf939885a * Memoise checkVarDefs since internally produced terms (i.e., not the
result of parsing) can have very heavy sharing, causing exponential
  complexity if we naively recurse into them.  ATerms are graphs, not
  trees!
2005-11-04 15:17:05 +00:00
Eelco Dolstra 0f827cc607 * Prevent repeated wrapping of closed terms
(closed(closed(closed(...)))) since this reduces performance by
  producing bigger terms and killing caching (which incidentally also
  prevents useful infinite recursion detection).
2005-07-19 11:48:05 +00:00
Eelco Dolstra 426593162e * ATermMap needs an assignment operator, otherwise we are screwed. 2005-05-08 10:28:19 +00:00
Eelco Dolstra 96de272b48 * Renamed normalise.cc' -> build.cc', `storeexprs.cc' ->
`derivations.cc', etc.
* Store the SHA-256 content hash of store paths in the database after
  they have been built/added.  This is so that we can check whether
  the store has been messed with (a la `rpm --verify').
* When registering path validity, verify that the closure property
  holds.
2005-01-19 16:39:47 +00:00
Eelco Dolstra cb7ccb528b * string2ATerm -> overloaded toATerm. 2004-11-03 18:12:03 +00:00
Eelco Dolstra a69534fc21 * Drop ATmake / ATMatcher also in handling store expressions. 2004-10-29 11:22:49 +00:00
Eelco Dolstra 5fe9222b36 * Don't use ATmake / ATmatch anymore, nor the ATMatcher class.
Instead we generate data bindings (build and match functions) for
  the constructors specified in `constructors.def'.  In particular
  this removes the conversions between AFuns and strings, and Nix
  expression evaluation now seems 3 to 4 times faster.
2004-10-26 22:54:26 +00:00
Eelco Dolstra 9fa07b376d * String/path concatenation operator (`+'). 2004-10-26 17:01:35 +00:00
Eelco Dolstra 37d7abd694 * New language feature: with expressions.
The expression `with E1; E2' evaluates to E2 with all bindings in
  the attribute set E1 substituted.  E.g.,

    with {x = 123;}; x

  evaluates to 123.  That is, the attribute set E1 is in scope in E2.

  This is particularly useful when importing files containing lots
  definitions.  E.g., instead of

    let {
      inherit (import ./foo.nix) a b c d e f;

      body = ... a ... f ...;
    }

  we can now say

    with import ./foo.nix;

    ... a ... f ...

  I.e., we don't have to say what variables should be brought into scope.
2004-10-25 16:54:56 +00:00
Eelco Dolstra 59b94ee18a * When something goes wrong in the evaluation of a Nix expression,
print a nice backtrace of the stack, rather than vomiting a gigantic
  (and useless) aterm on the screen.  Example:

    error: while evaluating file `.../pkgs/system/test.nix':
    while evaluating attribute `subversion' at `.../pkgs/system/all-packages-generic.nix', line 533:
    while evaluating function at `.../pkgs/applications/version-management/subversion/default.nix', line 1:
    assertion failed at `.../pkgs/applications/version-management/subversion/default.nix', line 13

  Since the Nix expression language is lazy, the trace may be
  misleading.  The purpose is to provide a hint as to the location of
  the problem.
2004-04-05 22:27:41 +00:00
Eelco Dolstra c4ac2a164a * The recent change in nixpkgs of calling `stdenv.mkDerivation'
instead of `derivation' triggered a huge slowdown in the Nix
  expression evaluator.  Total execution time of `nix-env -qa' went up
  by a factor of 60 or so.

  This scalability problem was caused by expressions such as

    (x: y: ... x ...) a b

  where `a' is a large term (say, the one in
  `all-packages-generic.nix').  Then the first beta-reduction would
  produce

    (y: ... a ...) b

  by substituting `a' for `x'.  The second beta-reduction would then
  substitute `b' for `y' into the body `... a ...', which is a large
  term due to `a', and thus causes a large traversal to be performed
  by substitute() in the second reduction.  This is however entirely
  redundant, since `a' cannot contain free variables (since we never
  substitute below a weak head normal form).

  The solution is to wrap substituted terms into a `Closed'
  constructor, i.e.,

    subst(subs, Var(x)) = Closed(e) iff subs[x] = e

  have substitution not descent into closed terms,

    subst(subs, Closed(x)) = Closed(x)

  and otherwise ignore them for evaluation,

    eval(Closed(x)) = eval(x).

* Fix a typo that caused incorrect substitutions to be performed in
  simple lambdas, e.g., `(x: x: x) a' would reduce to `(x: a)'.
2004-03-30 15:05:35 +00:00
Eelco Dolstra db3e644c1c * Added plain lambdas, e.g., `let { id = x: x; const = x: y: x; }'.
`bla:' is now no longer parsed as a URL.

* Re-enabled support for the `args' attribute in derivations to
  specify command line arguments to the builder, e.g.,

    ...
    builder = /usr/bin/python;
    args = ["-c" ./builder.py];
    ...
2004-03-28 20:34:22 +00:00
Eelco Dolstra fbc48a469c * Inherited attributes in recursive attribute sets are in scope of the
non-inherited attributes.
2004-02-16 09:18:35 +00:00
Eelco Dolstra 9b44480612 * Use a map to lookup primops.
* Various performance improvements in the evaluator.
* Do not link against unused (and missing!) libraries (-lsglr, etc.).
2004-02-04 16:03:29 +00:00
Eelco Dolstra c4f7ae4aa5 * Verify that all variables in a Nix expression are defined. 2004-02-03 14:45:34 +00:00
Eelco Dolstra 1c9c0a5a46 * Added syntactic sugar to the construction of attribute sets to
`inherit' variables from the surrounding lexical scope.

  E.g.,

    {stdenv, libfoo}: derivation {
      builder = ./bla;
      inherit stdenv libfoo;
      xyzzy = 1;
    }

  is equivalent to

    {stdenv, libfoo}: derivation {
      builder = ./bla;
      stdenv = stdenv;
      libfoo = libfoo;
      xyzzy = 1;
    }

  Note that for mutually recursive attribute set definitions (`rec
  {...}'), this also works, that is, `rec {inherit x;}' is equivalent
  to `let {fresh = x; body = rec {x = fresh;};}', *not*
  `rec {x = x}'.
2004-02-02 21:39:33 +00:00
Eelco Dolstra 1109ea0680 * Fixed a subtle uninitialised variable bug in ATermMaps copied from
ATermMaps.  Found thanks to Valgrind!
2004-01-21 14:49:32 +00:00
Eelco Dolstra 447089a5f6 * Catch SIGINT to terminate cleanly when the user tries to interrupt
Nix.  This is to prevent Berkeley DB from becoming wedged.

  Unfortunately it is not possible to throw C++ exceptions from a
  signal handler.  In fact, you can't do much of anything except
  change variables of type `volatile sig_atomic_t'.  So we set an
  interrupt flag in the signal handler and check it at various
  strategic locations in the code (by calling checkInterrupt()).
  Since this is unlikely to cover all cases (e.g., (semi-)infinite
  loops), sometimes SIGTERM may now be required to kill Nix.
2004-01-15 20:23:55 +00:00
Eelco Dolstra ac68840e79 * Refactoring: put the Nix expression evaluator in its own library so
that it can be used by multiple programs.
2003-11-19 11:35:41 +00:00
Renamed from src/nix-instantiate/nixexpr.cc (Browse further)