Commit graph

403 commits

Author SHA1 Message Date
Eelco Dolstra
54945a2950 * Refactoring: move parseExprFromFile() and parseExprFromString() into
the EvalState class.
2011-08-06 13:02:55 +00:00
Eelco Dolstra
c8606664ab * Don't allow derivations with fixed and non-fixed outputs. 2011-07-20 18:26:00 +00:00
Eelco Dolstra
b2027f70d9 * Fix a huuuuge security hole in the Nix daemon. It didn't check that
derivations added to the store by clients have "correct" output
  paths (meaning that the output paths are computed by hashing the
  derivation according to a certain algorithm).  This means that a
  malicious user could craft a special .drv file to build *any*
  desired path in the store with any desired contents (so long as the
  path doesn't already exist).  Then the attacker just needs to wait
  for a victim to come along and install the compromised path.

  For instance, if Alice (the attacker) knows that the latest Firefox
  derivation in Nixpkgs produces the path

    /nix/store/1a5nyfd4ajxbyy97r1fslhgrv70gj8a7-firefox-5.0.1

  then (provided this path doesn't already exist) she can craft a .drv
  file that creates that path (i.e., has it as one of its outputs),
  add it to the store using "nix-store --add", and build it with
  "nix-store -r".  So the fake .drv could write a Trojan to the
  Firefox path.  Then, if user Bob (the victim) comes along and does

    $ nix-env -i firefox
    $ firefox

  he executes the Trojan injected by Alice.

  The fix is to have the Nix daemon verify that derivation outputs are
  correct (in addValidPath()).  This required some refactoring to move
  the hash computation code to libstore.
2011-07-20 18:10:47 +00:00
Eelco Dolstra
d329c3ea9d * Support multiple outputs. A derivation can declare multiple outputs
by setting the ‘outputs’ attribute.  For example:

    stdenv.mkDerivation {
      name = "aterm-2.5";

      src = ...;

      outputs = [ "out" "tools" "dev" ];

      configureFlags = "--bindir=$(tools)/bin --includedir=$(dev)/include";
    }

  This derivation creates three outputs, named like this:

    /nix/store/gcnqgllbh01p3d448q8q6pzn2nc2gpyl-aterm-2.5
    /nix/store/gjf1sgirwfnrlr0bdxyrwzpw2r304j02-aterm-2.5-tools
    /nix/store/hp6108bqfgxvza25nnxfs7kj88xi2vdx-aterm-2.5-dev

  That is, the symbolic name of the output is suffixed to the store
  path (except for the ‘out’ output).  Each path is passed to the
  builder through the corresponding environment variable, e.g.,
  ${tools}.

  The main reason for multiple outputs is to allow parts of a package
  to be distributed and garbage-collected separately.  For instance,
  most packages depend on Glibc for its libraries, but don't need its
  header files.  If these are separated into different store paths,
  then a package that depends on the Glibc libraries only causes the
  libraries and not the headers to be downloaded.

  The main problem with multiple outputs is that if one output exists
  while the others have been garbage-collected (or never downloaded in
  the first place), and we want to rebuild the other outputs, then
  this isn't possible because we can't clobber a valid output (it
  might be in active use).  This currently gives an error message
  like:

    error: derivation `/nix/store/1s9zw4c8qydpjyrayxamx2z7zzp5pcgh-aterm-2.5.drv' is blocked by its output paths

  There are two solutions: 1) Do the build in a chroot.  Then we don't
  need to overwrite the existing path.  2) Use hash rewriting (see the
  ASE-2005 paper).  Scary but it should work.

  This is not finished yet.  There is not yet an easy way to refer to
  non-default outputs in Nix expressions.  Also, mutually recursive
  outputs aren't detected yet and cause the garbage collector to
  crash.
2011-07-18 23:31:03 +00:00
Eelco Dolstra
d5d4dcd4c9 * Allow attribute names to be strings. Based on the
allow-arbitrary-strinsg-in-names patch by Marc Weber.
2011-07-13 15:53:24 +00:00
Eelco Dolstra
0a623a10c7 * Allow a default value in attribute selection by writing
x.y.z or default

  (as originally proposed in
  https://mail.cs.uu.nl/pipermail/nix-dev/2009-September/002989.html).

  For instance, an expression like

    stdenv.lib.attrByPath ["features" "ckSched"] false args

  can now be written as

    args.features.ckSched or false
2011-07-13 12:19:57 +00:00
Eelco Dolstra
2b9e29b1c8 * Change the right-hand side of the ‘.’ operator from an attribute to
an attribute path.  This is a refactoring to support default values.
2011-07-06 12:28:57 +00:00
Eelco Dolstra
5637037802 * In the ‘?’ operator, allow attribute paths. For instance, you can
write ‘attrs ? a.b’ to test whether ‘attrs’ has an attribute ‘a’
  containing an attribute ‘b’.  This is more convenient than ‘attrs ?
  a && attrs.a ? b’.

  Slight change in the semantics: it's no longer an error if the
  left-hand side of ‘?’ is not an attribute set.  In that case it just
  returns false.  So, ‘null ? foo’ no longer throws an error.
2011-07-06 10:58:17 +00:00
Eelco Dolstra
538b7caab0 * Don't allocate a big initial GC address space on machines with
little RAM.  Even if the memory isn't actually used, it can cause
  problems with the overcommit heuristics in the kernel.  So use a VM
  space of 25% of RAM, up to 384 MB.
2011-02-10 14:31:04 +00:00
Eelco Dolstra
5a6b039802 * Don't call GC_expand_hp unless we're actually using the garbage
collector.
2011-02-09 22:59:50 +00:00
Ludovic Courtès
f1462c208b Use $BDW_GC_LIBS instead of a custom variable. 2011-02-09 21:23:13 +00:00
Eelco Dolstra
9db190eb31 * builtins.substring: if "start" is beyond the end of the string,
return the empty string.
2011-01-14 12:47:10 +00:00
Eelco Dolstra
812fae424e * DrvInfo contains pointers to expressions, so DrvInfos should be
traced by the garbage collector.  Otherwise "nix-env -u" can crash
  randomly.
2010-11-25 13:47:34 +00:00
Eelco Dolstra
0c4828ea05 * new(UseGC) is inexplicably slower than GC_MALLOC, so prefer the
latter.
2010-10-28 12:50:01 +00:00
Eelco Dolstra
e11e6fb1c6 * Handle out of memory condition. 2010-10-28 12:29:40 +00:00
Eelco Dolstra
43535499f3 * When allocating an attribute set, reserve enough space for all
elements.  This prevents the vector from having to resize itself.
2010-10-24 20:09:37 +00:00
Eelco Dolstra
e0b7fb8f27 * Keep attribute sets in sorted order to speed up attribute lookups.
* Simplify the representation of attributes in the AST.
* Change the behaviour of listToAttrs() in case of duplicate names.
2010-10-24 19:52:33 +00:00
Eelco Dolstra
2dc6d50941 * Don't create thunks for variable lookups (if possible). This
significantly reduces the number of values allocated (e.g. from 8.7m
  to 4.9m for the Bittorrent test).
2010-10-24 14:20:02 +00:00
Eelco Dolstra
0b305c534f * Store attribute sets as a vector instead of a map (i.e. a red-black
tree).  This saves a lot of memory.  The vector should be sorted so
  that names can be looked up using binary search, but this is not the
  case yet.  (Surprisingly, looking up attributes using linear search
  doesn't have a big impact on performance.)

  Memory consumption for

    $ nix-instantiate /etc/nixos/nixos/tests -A bittorrent.test --readonly-mode

  on x86_64-linux with GC enabled is now 185 MiB (compared to 946
  MiB on the trunk).
2010-10-24 00:41:29 +00:00
Eelco Dolstra
a247d20604 * Fix compiling without Boehm.
* Fix the stats.
2010-10-23 22:58:24 +00:00
Eelco Dolstra
b2ba62170c * Optimise string constants by putting them in the symbol table. 2010-10-23 21:11:59 +00:00
Eelco Dolstra
8ac06726b9 * Make Value smaller by not storing redundant PrimOp info.
* Clear pointers in Values after overwriting them to make sure that no
  objects are kept alive unnecessarily.
2010-10-23 20:07:47 +00:00
Eelco Dolstra
3f66cfb96b * Remove allocValues(). 2010-10-23 18:18:07 +00:00
Eelco Dolstra
4dee289550 * In environments, store pointers to values rather than values. This
improves GC effectiveness a bit more (because a live value doesn't
  keep other values in the environment plus the parent environments
  alive), and removes the need for copy nodes.
2010-10-22 15:51:52 +00:00
Eelco Dolstra
41c45a9b31 * Store Value nodes outside of attribute sets. I.e., Attr now stores
a pointer to a Value, rather than the Value directly.  This improves
  the effectiveness of garbage collection a lot: if the Value is
  stored inside the set directly, then any live pointer to the Value
  causes all other attributes in the set to be live as well.
2010-10-22 14:47:42 +00:00
Eelco Dolstra
64c3325b0b * Make building against the Boehm GC a configure option. 2010-10-22 13:39:15 +00:00
Eelco Dolstra
76feaf016a * Keep some more stats. 2010-10-20 15:48:00 +00:00
Eelco Dolstra
e879a0371b * Use the Boehm garbage collector to reclaim unused memory in the Nix
expression evaluator.
2010-10-20 11:38:30 +00:00
Eelco Dolstra
705868a8a9 * Make sure that config.h is included before the system headers,
because it defines _FILE_OFFSET_BITS.  Without this, on 
  OpenSolaris the system headers define it to be 32, and then 
  the 32-bit stat() ends up being called with a 64-bit "struct 
  stat", or vice versa.

  This also ensures that we get 64-bit file sizes everywhere.

* Remove the redundant call to stat() in parseExprFromFile().
  The file cannot be a symlink because that's the exit condition
  of the loop before.
2010-10-04 17:55:38 +00:00
Eelco Dolstra
95f4f2cf61 * If std::tr1::unordered_set is unavailable, use std::set. 2010-10-04 16:16:19 +00:00
Eelco Dolstra
4aa9245083 * Hack needed for GCC 4.3.2 on OpenSolaris. 2010-10-04 10:51:16 +00:00
Eelco Dolstra
df50916e46 * Oops - "null" was displayed as "true". 2010-08-27 12:10:56 +00:00
Eelco Dolstra
6d6200f37a * Optimisation in the // operator: if one of the sets is empty, return
the other set.
2010-08-02 16:31:05 +00:00
Eelco Dolstra
7af6a2fd71 * intersectAttrs: optimise for the case where the second set is larger
than the first set.  (That's usually the case with callPackage.)
2010-08-02 11:54:44 +00:00
Eelco Dolstra
f16fe2af8d * builtins.toXML: propagate the string context. This is a regression
from the old ATerm-based evaluator.
2010-06-10 10:29:50 +00:00
Eelco Dolstra
1ab67cf437 2010-06-02 09:43:04 +00:00
Eelco Dolstra
89865da76d * Turn build errors during evaluation into EvalErrors. 2010-06-01 11:19:32 +00:00
Eelco Dolstra
93cd5a4a13 * The << operator on values should be const. 2010-05-18 10:36:37 +00:00
Eelco Dolstra
b2235d81d1 * Restore the __overrides feature that was lost somewhere in the
fast-eval branch.
2010-05-15 08:10:12 +00:00
Eelco Dolstra
bd25ac2260 * Print attributes in sorted order. 2010-05-12 12:15:49 +00:00
Eelco Dolstra
81a4b4e49b * Implemented tryEval, the last missing primop in the fast-eval
branch.  Also added a test for tryEval.
2010-05-12 11:23:44 +00:00
Eelco Dolstra
1a8eb6e3ec 2010-05-07 15:26:33 +00:00
Eelco Dolstra
83dfa89870 * Sync with the trunk. 2010-05-07 14:46:47 +00:00
Eelco Dolstra
01e58adce0 * Store position info for inherited attributes. 2010-05-07 12:43:57 +00:00
Eelco Dolstra
83d7b89660 * Updated addErrorContext. 2010-05-07 12:33:14 +00:00
Eelco Dolstra
e2d5e40f4f * Keep track of the source positions of attributes. 2010-05-07 12:11:05 +00:00
Eelco Dolstra
84ce7ac76f * Store attribute positions in the AST and report duplicate attribute
errors with position info.
* For all positions, use the position of the first character of the
  first token, rather than the last character of the first token plus
  one.
2010-05-06 16:46:48 +00:00
Eelco Dolstra
2be6118f4c * Don't need the test program anymore. 2010-04-23 09:09:32 +00:00
Eelco Dolstra
0bc468f195 * Simplify the implementation of `with'. This gives a 7% speedup in
evaluating the NixOS system configuration.
2010-04-22 15:08:09 +00:00
Eelco Dolstra
ebade9ff8b * Check for duplicate attribute names / function arguments. `make
check' now succeeds :-)
* An attribute set such as `{ foo = { enable = true; };
  foo.port = 23; }' now parses.  It was previously rejected, but I'm
  too lazy to implement the check.  (The only reason to reject it is
  that the reverse, `{ foo.port = 23; foo = { enable = true; }; }', is
  rejected, which is kind of ugly.)
2010-04-22 11:02:24 +00:00