November 2014 Commit Log

Number of Commits:
13
Number of Active Developers:
2
akozlov 2014-11-27 00:11 Rev.: 13264

- partial merge from sativa branch
* log:branches/sativa@13263:13263

0 lines of code changed in 1 file:

akozlov 2014-11-26 03:56 Rev.: 13243

- partial merge from sativa branch
* log:branches/sativa@13241:13241

370 lines of code changed in 14 files:

westram 2014-11-19 09:13 Rev.: 13204


* +lib/sativa, -bin/sativa
- source-tarball-generation
- svn-ignores

1 lines of code changed in 1 file:

akozlov 2014-11-19 02:34 Rev.: 13203

- partial merge from sativa branch
* log:branches/sativa@13202:13202

0 lines of code changed in 8 files:

westram 2014-11-18 15:36 Rev.: 13197


- delete sub-mergeinfo accidentally introduced with [13191]

0 lines of code changed in 2 files:

westram 2014-11-17 16:42 Rev.: 13191


* reintegrates 'slv' into 'trunk'
* performs #624
* adds:
- log:branches/slv@13152:13190

1010 lines of code changed in 22 files:

westram 2014-11-12 18:51 Rev.: 13163


* accidently activated location checks with gcc>=4.8 for others with [13019]

7 lines of code changed in 1 file:

westram 2014-11-06 17:12 Rev.: 13136


- fix clang compilation (broken by [13134])

3 lines of code changed in 1 file:

westram 2014-11-06 16:07 Rev.: 13135


- disable {{{-Wmaybe-uninitialized}}} in DEBUG mode (too many bogus warnings that do NOT appear in NDEBUG)

4 lines of code changed in 1 file:

westram 2014-11-06 15:51 Rev.: 13134


- fix warnings reported by NDEBUG compilation

16 lines of code changed in 11 files:

westram 2014-11-06 15:50 Rev.: 13133


- enable warnings in NDEBUG and RELEASE

6 lines of code changed in 1 file:

westram 2014-11-06 14:31 Rev.: 13130


* fix undefined behavior in loops over enum values
- only checked for-loops, other manual loops still might exist
- NDEBUG compilation with some gcc versions (e.g. 4.3.3, 4.4.3) optimize out the loop condition:
* in case of the loop added in [13129] the condition {{{at<=MY_AA}}} is regarded as "always true", because standard defines the variable {{{at}}} '''may not''' contain any value greater than MY_AA
* newer gccs do no longer perform such rigid optimizations
* the "bug" is described [https://gcc.gnu.org/bugzilla/show_bug.cgi?id=42810 here]
- as you can see the test {{{TEST_EXPECT(at>=1 && at<=3);}}} is optimized out as well
- this patch shows two different ways to remove undefined behavior:
* add another value (greater than tested loop condition) to enum (done in GEN_graphic.cxx). may easily be broken by removing that (unused) value later .
* use int as loop-variable, then assign to enum variable

11 lines of code changed in 4 files:

westram 2014-11-05 17:09 Rev.: 13129


- check some undefined behavior

19 lines of code changed in 1 file:

October 2014 »

Generated by StatSVN 0.7.0