Howard Hinnant
|
397717b7c8
|
constexpr applied to <array>.
llvm-svn: 160564
|
2012-07-20 19:20:49 +00:00 |
|
Howard Hinnant
|
38b99b025c
|
Add test for self-referencing emplace test.
llvm-svn: 159921
|
2012-07-09 02:47:43 +00:00 |
|
Howard Hinnant
|
4cb38a82a2
|
This commit establishes a new bucket_count policy in the unordered containers: The policy now allows a power-of-2 number of buckets to be requested (and that request honored) by the client. And if the number of buckets is set to a power of 2, then the constraint of the hash to the number of buckets uses & instead of %. If the client does not specify a number of buckets, then the policy remains unchanged: a prime number of buckets is selected. The growth policy is that the number of buckets is roughly doubled when needed. While growing, either the prime, or the power-of-2 strategy will be preserved. There is a small run time cost for putting in this switch. For very cheap hash functions, e.g. identity for int, the cost can be as high as 18%. However with more typical use cases, e.g. strings, the cost is in the noise level. I've measured cases with very cheap hash functions (int) that using a power-of-2 number of buckets can make look up about twice as fast. However I've also noted that a power-of-2 number of buckets is more susceptible to accidental catastrophic collisions. Though I've also noted that accidental catastrophic collisions are also possible when using a prime number of buckets (but seems far less likely). In short, this patch adds an extra tuning knob for those clients trying to get the last bit of performance squeezed out of their hash containers. Casual users of the hash containers will not notice the introduction of this tuning knob. Those clients who swear by power-of-2 hash containers can now opt-in to that strategy. Clients who prefer a prime number of buckets can continue as they have.
llvm-svn: 159836
|
2012-07-06 17:31:14 +00:00 |
|
Howard Hinnant
|
8b805c915a
|
The rules for emplace in map, multimap, unordered_map and unordered_multimap changed a while back and I'm just now updating to these new rules. In a nutshell, you've got to know you're emplacing to a pair and use one of pair's constructors. I made one extension: If you want to emplace the key and default construct the mapped_type, you can just emplace(key), as opposed to emplace(piecewise_construct, forward_as_tuple(key), forward_as_tuple()).
llvm-svn: 157503
|
2012-05-25 22:04:21 +00:00 |
|
Howard Hinnant
|
1196716270
|
unord test fixes by Edward Meewis
llvm-svn: 145707
|
2011-12-02 21:23:14 +00:00 |
|
Howard Hinnant
|
920b56ca88
|
Another installment on debug mode. This addresses list. However this should be considered a temporary state. The API of the debug database and how vector and list use it, is unsatisfactory at the moment. It is both inefficient and overly verbose. I wanted to get this functionality checked in though. In the next day or so I'll refactor what is there in an attempt to streamline things.
llvm-svn: 140660
|
2011-09-27 23:55:03 +00:00 |
|
Howard Hinnant
|
0695db06d7
|
The vector test suite now passes for no-debug, debug-lite and debug-regular
llvm-svn: 139930
|
2011-09-16 18:41:29 +00:00 |
|
Howard Hinnant
|
54976f2619
|
Fixed PR10574: http://llvm.org/bugs/show_bug.cgi?id=10574
llvm-svn: 137522
|
2011-08-12 21:56:02 +00:00 |
|
Howard Hinnant
|
bd0c160078
|
noexcept for <stack>. This completes noexcept for Chapter 23 [containers].
llvm-svn: 132652
|
2011-06-04 22:09:19 +00:00 |
|
Howard Hinnant
|
6971d82668
|
noexcept for <queue>.
llvm-svn: 132650
|
2011-06-04 21:32:33 +00:00 |
|
Howard Hinnant
|
557da866a3
|
noexcept for <unordered_set>.
llvm-svn: 132647
|
2011-06-04 20:18:37 +00:00 |
|
Howard Hinnant
|
3714107ed6
|
noexcept for <unordered_map>.
llvm-svn: 132646
|
2011-06-04 18:54:24 +00:00 |
|
Howard Hinnant
|
0e9f71c13f
|
noexcept for <set>. Plus a few fixes to noexcept for <map>.
llvm-svn: 132640
|
2011-06-04 15:22:34 +00:00 |
|
Howard Hinnant
|
1052ee39cb
|
noexcept for <map>.
llvm-svn: 132639
|
2011-06-04 14:31:57 +00:00 |
|
Howard Hinnant
|
3297ed7065
|
noexcept for <vector>. This also includes installing move_if_noexcept() into vector.
llvm-svn: 132577
|
2011-06-03 19:40:40 +00:00 |
|
Howard Hinnant
|
4590010473
|
noexcept for <list>.
llvm-svn: 132562
|
2011-06-03 17:30:28 +00:00 |
|
Howard Hinnant
|
91a4750733
|
noexcept for <forward_list>.
llvm-svn: 132553
|
2011-06-03 16:20:53 +00:00 |
|
Howard Hinnant
|
801291139c
|
After sleeping on it I've decided that all special members that can be noexcept, should be declared so. The client has the traits to detect and branch on this information, and it is often an important optimization. Give deque() a noexcept. Add test for deque default constructor and deque destructor.
llvm-svn: 132549
|
2011-06-03 15:16:49 +00:00 |
|
Howard Hinnant
|
60954bbbbc
|
Add noexcept tests for deque.
llvm-svn: 132523
|
2011-06-03 01:46:22 +00:00 |
|
Howard Hinnant
|
a7fa0710cb
|
Suppress some warings in the tests.
llvm-svn: 131326
|
2011-05-13 23:59:50 +00:00 |
|
Howard Hinnant
|
e4b11156d7
|
Corrected tests for http://llvm.org/bugs/show_bug.cgi?id=9351
llvm-svn: 128807
|
2011-04-04 01:42:30 +00:00 |
|
Howard Hinnant
|
e4bafbee25
|
... And these wonderful tests.
llvm-svn: 128797
|
2011-04-03 20:06:31 +00:00 |
|
Howard Hinnant
|
2774545736
|
Bug 9096 - list::iterator not default constructible
llvm-svn: 124508
|
2011-01-28 23:46:28 +00:00 |
|
Howard Hinnant
|
601afb30ec
|
LWG 1191
llvm-svn: 119545
|
2010-11-17 19:52:17 +00:00 |
|
Howard Hinnant
|
412dbebe1b
|
license change
llvm-svn: 119395
|
2010-11-16 22:09:02 +00:00 |
|
Howard Hinnant
|
c1ef47986b
|
Fix two test bugs
llvm-svn: 116515
|
2010-10-14 21:08:32 +00:00 |
|
Howard Hinnant
|
1c1620b269
|
Updated by-chapter-summary with weekly test results, and fixed up some bad paths in some tests caused by aligning the test suite with N3126.
llvm-svn: 113403
|
2010-09-08 20:31:42 +00:00 |
|
Howard Hinnant
|
cbf93f3959
|
sync with N3126
llvm-svn: 113101
|
2010-09-05 01:14:30 +00:00 |
|
Howard Hinnant
|
7609c9b665
|
Changed __config to react to all of clang's currently documented has_feature flags, and renamed _LIBCPP_MOVE to _LIBCPP_HAS_NO_RVALUE_REFERENCES to be more consistent with the rest of the libc++'s flags, and with clang's nomenclature.
llvm-svn: 113086
|
2010-09-04 23:28:19 +00:00 |
|
Howard Hinnant
|
8f2f7e7c79
|
Fixing whitespace problems
llvm-svn: 111755
|
2010-08-22 00:15:28 +00:00 |
|
Howard Hinnant
|
3db88036f6
|
US 117
llvm-svn: 111745
|
2010-08-21 20:58:44 +00:00 |
|
Howard Hinnant
|
e57dc14c44
|
LWG 1278
llvm-svn: 111532
|
2010-08-19 17:40:04 +00:00 |
|
Howard Hinnant
|
5b08a8a432
|
Wiped out some non-ascii characters that snuck into the copyright.
llvm-svn: 103516
|
2010-05-11 21:36:01 +00:00 |
|
Howard Hinnant
|
3e519524c1
|
libcxx initial import
llvm-svn: 103490
|
2010-05-11 19:42:16 +00:00 |