Commit Graph

37 Commits

Author SHA1 Message Date
Terrence Cole
f6ddf5df97 Bug 759991 - Fix infinite loop in rekeyFront with fully collided Table; r=luke
This hooks up the same path to putNew, because it is slightly more efficient and
fixes an OOM failure introduced in c9024bcb8da0.
2012-06-06 16:40:56 -07:00
Phil Ringnalda
498d1a6e91 Back out ecbe3c75551d (bug 759991) to investigate the effects on TestStartupCache bustage on a CLOSED TREE 2012-06-11 17:38:08 -07:00
Terrence Cole
2c6246852d Bug 759991 - Fix infinite loop in rekeyFront with fully collided Table; r=luke
This hooks up the same path to putNew, because it is slightly more efficient and
fixes an OOM failure introduced in c9024bcb8da0.
2012-06-06 16:40:56 -07:00
Nicholas Nethercote
267c1cb82d Bug 760337 - Add JS_ASSERT(table) where appropriate in HashTable.h. r=luke.
--HG--
extra : rebase_source : ab85f7061f79605afe54a6341612cededef35558
2012-06-04 00:00:24 -07:00
Terrence Cole
373c2da2dd Bug 760342 - Check explicitly for mis-use of HashTable::Enum; r=luke
If a user does removeFront or rekeyFront on an Enum, then continues to use it,
it will fail, potentially randomly. This patch makes the failure explicit in
debug builds.
2012-06-01 15:41:27 -07:00
Matt Brubeck
8f0eb7c918 Back out bug 760342 for Windows build error, bug 755070 for Android crashes, and bug 760458 because it depends on 755070 2012-06-01 17:45:02 -07:00
Terrence Cole
53b40ccbbb Bug 760342 - Check explicitly for mis-use of HashTable::Enum; r=luke
If a user does removeFront or rekeyFront on an Enum, then continues to use it,
it will fail, potentially randomly. This patch makes the failure explicit in
debug builds.
2012-06-01 15:41:27 -07:00
Gervase Markham
ca171eec44 Bug 716478 - update licence to MPL 2. 2012-05-21 12:12:37 +01:00
Terrence Cole
ab1313c6c1 Bug 753479 - Make HashTable::Enum::rekeyFront take a separate Lookup; r=luke
--HG--
extra : rebase_source : 7f8b2cda64f34c60f702182f64e370a73a593934
2012-05-04 14:18:28 -07:00
Terrence Cole
2d725e1799 Bug 752604 - Ergonomics improvements for HashTable::rekeyFront; r=luke
--HG--
extra : rebase_source : 9764f2db38844106ac13f5281d85464f0b2aff80
2012-05-07 12:46:28 -07:00
Tom Schuster
fde4c4d67e Bug 735036 - Check for OOM in debug only method. r=luke 2012-04-24 20:03:48 +02:00
Terrence Cole
ae810f2b36 Bug 744052 - Restructure HashTable::checkOverflow for easier reading; r=luke
When I added this function, I just copied the existing logical structure.  It
makes more sense here to check overflowed and return early, since we can do this
now.
2012-04-10 09:51:48 -07:00
Terrence Cole
49d3da6509 Bug 743885 - Fix OOM handling regression in HashTable; r=luke
Introduced by me in bug 722946.
2012-04-09 19:07:11 -07:00
Daniel Holbert
6fdccfb996 Bug 742438: Fix straggling bool-vs-NULL comparison after HashSet::put return-type change. r=terrence 2012-04-04 11:18:18 -07:00
Terrence Cole
ed16bd1b59 Bug 740153 - Don't expose Entry directly through HashMap; r=luke
HashMap and HashSet only expose Entries through Ptr and Range, except for one
variant of put. They also don't expose the default size they use if you call
init() without an argument. These weirdnesses seem to be interface carried over
from the old jsdhash.

The first oddness is unused and can be removed, the second should simply get
exposed to make it easier to wrap HashMap and HashSet.

--HG--
extra : rebase_source : e1432a9af189229f379158ba0529956ddbf5f234
2012-03-30 11:01:58 -07:00
Terrence Cole
a683dde987 Bug 722946 - Add method to HashTable::Enum for inline rekeying; r=luke
It is possible in several places to have unconstrained client objects as the key
of a HashMap/HashSet. For generational and compacting GC, we will need to move
these objects in memory. This means we will need to remove and re-insert these
keys in a new location, during GC. Because we cannot safely allocate memory for
this, we need a specialized algorithm inside the HashTable itself.
2012-03-14 13:48:59 -07:00
Ehsan Akhgari
aad636cb78 Backout changeset ea6be5f60c42 (bug 722946) for breaking Windows builds 2012-03-14 14:29:58 -04:00
Terrence Cole
6ab75b39c1 Bug 722946 - Add method to HashTable for inline rekeying/removal; r=luke
It is possible in several places to have unconstrained client objects as the key
of a HashMap/HashSet.  For generational and compacting GC, we will need to move
these objects in memory.  This means we will need to remove and re-insert these
keys in a new location, during GC.  Because we cannot safely allocate memory
for this, we need a specialized algorithm inside the HashTable itself.
2012-02-03 15:43:34 -08:00
Jason Orendorff
0af3864536 Add static assertions as requested in bug 726298 comment 2 and 4.
--HG--
extra : rebase_source : 3edf2446076dae08c16b121de18505174c84f370
2012-03-01 12:48:47 -06:00
Nicholas Nethercote
a7ad4ef70e Bug 715453 - Remove computedSize from nsMallocSizeOfFun. r=jlebar,bhackett.
--HG--
extra : rebase_source : a65039a407daab45360a5b375b53cbf1bc05b7f6
2012-01-25 00:52:51 -08:00
Bill McCloskey
1762125a12 Bug 720853 - Add read barrier for atom table (r=luke) 2012-01-24 18:32:56 -08:00
Jeff Walden
c9f5077db7 Bug 714264 - Move declarations around such that it's possible to include jsprvtd.h before Vector.h or HashTable.h. r=luke
--HG--
extra : rebase_source : e981944be5edf5c772df986d0898494b9bc53a3e
2012-01-02 16:34:25 -06:00
Brian Hackett
7e5c4a14dd Backed out changeset 4282a285d381 (bug 707049 backout). 2011-12-31 07:32:04 -07:00
Ms2ger
ed558e4c7e Backout 0d642e4e08cf (bug 707049); requested by evilpie. 2011-12-31 15:13:01 +01:00
Brian Hackett
cda1e2a770 Add dynamic analysis for identifying moving GC hazards, autorooter classes, some VM rooters, bug 707049. r=billm 2011-12-30 20:12:45 -07:00
Jeff Walden
ede1252065 Use < rather than != when comparing a bunch of pointers to sentinel end-pointer values, as a small hedge against things going awry and skipping past the sentinel. No bug, r=firebot
--HG--
extra : rebase_source : 567f70397ec9ef006c55c01f3acde305ab527b56
2011-12-29 15:50:33 -06:00
Hub Figuière
f340716015 Bug 714169 - Fix warning in HashTable.h by using |continue;| instead of |;|. r=jwalden
--HG--
extra : rebase_source : 176f652ee0e2c23b2c995b51814e4e1fd1ae0149
2011-12-29 13:37:04 -08:00
Jeff Walden
87d3f92e61 Bug 708735 - Use <stdint.h> types in JSAPI and throughout SpiderMonkey. Continue to provide the {u,}int{8,16,32,64} and JS{Uint,Int}{8,16,32,64} integer types through a single header, however, for a simpler backout strategy -- and also to ease the transition for embedders. r=timeless on switching the jsd API to use the <stdint.h> types, r=luke, r=dmandelin 2011-12-08 22:54:10 -05:00
Brian Hackett
4fa421bf15 Merge MC -> JM 2011-11-30 12:45:27 -08:00
Brian Hackett
586dafd5c1 Add about:memory reporters for per compartment shape tables, bug 704372. 2011-11-21 20:29:05 -05:00
Nicholas Nethercote
f102556f32 Bug 698968 - Add mallocSizeOf functions and start using them. r=jlebar,bhackett,jfkthame, sr=bz. 2011-11-27 19:03:14 -08:00
Bill McCloskey
8e7a5d934a Bug 641027 - Add snapshot-at-the-beginning write barriers for incremental GC (r=luke,bhackett) 2011-10-25 16:07:42 -07:00
Nicholas Nethercote
b5f0310b81 Bug 699724 - Shrink js::HashTable. r=luke. 2011-11-08 21:02:11 -08:00
Nicholas Nethercote
dbd773b4ff Bug 684800 - Measure slop in TI-related JS memory reporters. r=bhackett. 2011-11-01 21:59:43 -07:00
Nicholas Nethercote
e2861e8677 Bug 697933 - Allow HashTables to have a capacity as small as four. r=luke. 2011-10-30 16:12:48 -07:00
Nicholas Nethercote
95ee61442b Bug 697931 - Remove KidHashes when possible. r=luke. 2011-10-27 23:06:17 -07:00
Luke Wagner
c7d3e6785c Bug 689362 - Create js/public (with the intention that its contents will eventually be the only INSTALLED_HEADERS), move Vector and HashTable to js/public and some dependencies (r=billm,sr=dmandelin)
--HG--
rename : js/src/jshashtable.h => js/public/HashTable.h
rename : js/src/jsvector.h => js/public/Vector.h
rename : js/src/mfbt/InlineMap.h => js/src/ds/InlineMap.h
extra : rebase_source : 818639b74653667b7a572bc1234398c0e4559a60
2011-10-04 15:33:00 -07:00