Commit graph

2461 commits

Author SHA1 Message Date
Eelco Dolstra d7ca6f44eb * Update some comments. 2010-12-13 13:19:46 +00:00
Eelco Dolstra 542fc69062 * When doing a query (e.g. `nix-store -r --dry-run'), don't make a lot
of expensive calls to `nix-store --check-validity'.
2010-12-13 08:39:10 +00:00
Eelco Dolstra 4d57776813 * Use SQLite 3.7.4. 2010-12-10 11:45:56 +00:00
Eelco Dolstra e4720b1a79 * Ignore the result of sqlite3_reset(). 2010-12-08 18:19:15 +00:00
Eelco Dolstra 7d0444e244 * Bad things happen when a filehandle has the same name as a module. 2010-12-07 12:33:42 +00:00
Eelco Dolstra 8062d3af30 * `nix-store --verify --check-contents': don't hold the global GC lock
while checking the contents, since this operation can take a very
  long time to finish.  Also, fill in missing narSize fields in the DB
  while doing this.
2010-12-06 15:29:38 +00:00
Eelco Dolstra de79d23f76 * Retry a transaction if SQLite returns SQLITE_BUSY. This can happen
even with a very long busy timeout, because SQLITE_BUSY is also
  returned to resolve deadlocks.  This should get rid of random
  "database is locked" errors.  This is kind of hard to test though.
* Fix a horrible bug in deleteFromStore(): deletePathWrapped() should
  be called after committing the transaction, not before, because the
  commit might not succeed.
2010-12-05 18:23:19 +00:00
Eelco Dolstra 365f3028dd * Use CamelCase for the Perl modules. 2010-12-05 17:50:29 +00:00
Eelco Dolstra f42a505ab7 * Add a script `nix-generate-patches'.
* Fix the binary patching test.
2010-12-05 17:36:02 +00:00
Eelco Dolstra 77f7a6d591 * Quick hack around one of the pathological cases in bsdiff
(namely, when there is a long region of 0s).  On one 86 MiB
  case, this cut patch generation time from 44m to 2m.
2010-11-29 17:07:04 +00:00
Eelco Dolstra d3bba0c2d8 * Move the patch generator into a module. 2010-11-29 15:26:28 +00:00
Eelco Dolstra 9737a7eba0 * Don't generate patches if the system type differs. 2010-11-29 09:26:05 +00:00
Eelco Dolstra bd48fd97f6 * Impose a configurable time limit on patch generation. This is
necessary because bsdiff has some pathological worst-case behaviour,
  e.g. O(n^2) on files consisting only of 0s.
2010-11-29 09:15:45 +00:00
Eelco Dolstra 703e5a2ce2 * Sync with the trunk. 2010-11-25 14:09:13 +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 d92ccbf1ac * Test whether sequences of patches work. 2010-11-17 18:01:37 +00:00
Eelco Dolstra 1a211d812f * Oops. 2010-11-17 17:54:49 +00:00
Eelco Dolstra a4f0365b2d * When checking whether a patch is applicable, for efficiency, use
`nix-store -q --hash' to get the hash of the base path rather than
  `nix-hash'.  However, only do this for estimating the size of a
  download, not for the actual substitution, because sometimes the
  contents of store paths are modified (which they shouldn't, of
  course).
2010-11-17 17:41:59 +00:00
Eelco Dolstra 3d38a49840 * In the download size indication, take binary patches into account.
Hopefully this doesn't slow things down too much.
2010-11-17 17:32:25 +00:00
Eelco Dolstra f69626ed3e 2010-11-17 15:45:09 +00:00
Eelco Dolstra a07c68f05e * Finally, a test for the binary patch functionality. 2010-11-17 15:30:07 +00:00
Eelco Dolstra bf658f016f * Test that download sizes are shown correctly. 2010-11-17 14:39:02 +00:00
Eelco Dolstra 1e24cbaba3 * Fix the test. 2010-11-17 14:35:14 +00:00
Eelco Dolstra bdf089f463 * Before a build, show the disk space that the downloaded store paths
will approximately require.
2010-11-17 14:31:42 +00:00
Eelco Dolstra 06699d4219 * Store the NAR size in the manifest. 2010-11-17 12:57:52 +00:00
Eelco Dolstra 5693b8a7e2 * nix-push: no need to compute the NAR hash, since the Nix database
already has it (`nix-store -q --hash').
2010-11-17 12:51:54 +00:00
Eelco Dolstra e60c962fb8 * Add an operation `nix-store -q --size'. 2010-11-17 12:40:52 +00:00
Eelco Dolstra 1db6259076 * Implement RemoteStore::queryPathInfo(). 2010-11-17 12:08:01 +00:00
Eelco Dolstra a3883cbd28 * Store the size of a store path in the database (to be precise, the
size of the NAR serialisation of the path, i.e., `nix-store --dump
  PATH').  This is useful for Hydra.
2010-11-16 17:11:46 +00:00
Eelco Dolstra fb9368b5a0 * Sync with the trunk. 2010-11-16 12:49:47 +00:00
Eelco Dolstra 4aced7f8d0 * Merge the GC branch. 2010-10-29 15:04:39 +00:00
Eelco Dolstra 26def5392f * Document Boehm GC support. 2010-10-29 14:44:02 +00:00
Eelco Dolstra 3d71c8013e * Use pkgconfig to locate the Boehm GC (as suggested by Ludo), if
--enable-gc is given.
2010-10-29 14:00:47 +00:00
Eelco Dolstra 14fbf85380 * Set libgc's initial heap size to 384 MiB to prevent garbage
collection in most cases (and therefore its performance overhead).
2010-10-29 13:11:50 +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 8a788e38ac * Install config.h. 2010-10-26 10:47:02 +00:00
Eelco Dolstra 11ccd44e95 * We need Bison 2.4 now. 2010-10-24 21:48:59 +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 02934b1200 * Regression test for listToAttr's behaviour if an attribute name
occurs multiple times.
2010-10-23 22:55:30 +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 cf7e645a48 * Regression test for __overrides. 2010-10-22 15:15:12 +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