Commit graph

11971 commits

Author SHA1 Message Date
Mickey Rose
15f841213a expression_grammar: restrict unquoted strings to single-atom expressions
- fixes #3017
- closes #3214
2016-02-08 17:15:13 +01:00
artemp
2b725dd257 shapeindex - skip 'null' shapes (ref #3288) 2016-02-08 15:51:13 +01:00
bergwerkgis
9e7b67e35d Merge branch 'master' of https://github.com/mapnik/mapnik 2016-02-08 13:55:25 +00:00
bergwerkgis
643391ba20 Windows: props file to override solution settings for /MP 2016-02-08 13:55:08 +00:00
Mickey Rose
8da250a0ee show build status badges for master (svg)
- Travis badge URL was .png
- AppVeyor badge used latest build (any branch)
- [skip ci]
2016-02-07 14:22:44 +01:00
Mickey Rose
ed8c8df738 expression_grammar: use one symbol table for all constants 2016-02-06 18:40:33 +01:00
Mickey Rose
ae39d48b86 test/expressions: add checks for logical operator precedence 2016-02-06 14:26:57 +01:00
Mickey Rose
959d4ded23 image_filter_grammar: rewrite using Nabialek trick 2016-02-06 02:03:55 +01:00
Mickey Rose
401dfacfc5 test/image_filter: prevent SIGSEGV when parsing fails 2016-02-06 01:56:05 +01:00
lightmare
b24f520439 Merge pull request #3239 from lightmare/expr-test
condense expressions tests
2016-02-06 00:37:05 +01:00
Mickey Rose
5764d619f3 expressions test: move expression strings into Catch calls, continue on failure 2016-02-05 17:22:01 +01:00
Mickey Rose
53bb55eade rename expresssions_test.cpp, didn't pass spell check 2016-02-05 17:22:01 +01:00
artemp
cb708d7f5a transform_grammar: correct and simplify rules per bcc7495a87 (commitcomment-15915607)
image_filter: relax `agg-stack-blur` rule to allow `agg-stack-blur,agg-stack-blur(),agg-stack-blur(1),agg-stack-blur(1,1)` syntax + add image filter parsing tests
2016-02-05 16:50:49 +01:00
artemp
bcc7495a87 transform grammar - fix optional seprator logic 2016-02-05 13:08:17 +01:00
bergwerkgis
eeb1507c6d AppVeyor: new try, nr.2 2016-02-05 11:13:09 +00:00
bergwerkgis
9d359ff3df AppVeyor: new try, nr.1 2016-02-05 11:05:12 +00:00
Mickey Rose
7594301d98 test/shapeindex: include loop variables in failure message
[skip ci]
2016-02-04 18:54:21 +01:00
artemp
8c1f71c39e try 8 JOBS 2016-02-04 12:51:20 +01:00
bergwerkgis
9282c60c91 AppVeyor: try again mapnik::value_null - correct <=, >= return values 2016-02-04 10:21:41 +00:00
artemp
7ab93a30f2 mapnik::value_null - correct <=, >= return values ref #3285 2016-02-04 10:06:11 +01:00
Dane Springmeyer
7ad4bf379d stick with clang-3.5 for now
- I have early evidence that mapnik compiled with 3.6 and mapnik-vt compiled with 3.5 == crashes
  - Also the mason deps are still compiled with 3.5
  - TODO: upgrade mapnik, mapnik-vt, and all mason deps to 3.6 all at once.
2016-02-03 11:18:00 -08:00
Dane Springmeyer
52cabfd107 expat is no longer a must have cairo dep since cairo can be built without fontconfig [skip ci] 2016-02-03 10:57:09 -08:00
artemp
35ab8cccdc follow up on previous commit - convert to common_type prior applying / (MSCV C4804) 2016-02-03 12:45:29 +01:00
artemp
376ffba7fe convert value_bool to value_integer before operator / (fixes unsafe use of bool - MSVC C4804) 2016-02-03 12:35:09 +01:00
artemp
c0724a6d95 value comparison - deduct return type automatically (fixes MSVC C4800) 2016-02-03 12:33:15 +01:00
artemp
09163f969d deps - update variant 2016-02-03 11:06:00 +01:00
Dane Springmeyer
1600bfe981 remove dead code in enumeration.hpp 2016-02-02 12:35:19 -08:00
Dane Springmeyer
09f137796d use std namespace for clib functions and fix spelling in comments 2016-02-02 12:35:19 -08:00
Dane Springmeyer
bc08617eb8 Merge pull request #3281 from mapnik/no-miniz
remove miniz support
2016-02-02 12:33:12 -08:00
Dane Springmeyer
834028db7a remove miniz support - library is unused in mapnik and unmaintained since 2013 (https://code.google.com/archive/p/miniz/) 2016-02-02 11:37:26 -08:00
artemp
b427139951 update variant 2016-02-02 12:41:22 +01:00
artemp
13af423046 unit test - use 'parse_image_filters' and avoid instantiating expensive redundant grammar 2016-02-02 11:30:00 +01:00
artemp
39a97865fc image_filter_grammar - move 'BOOST_FUSION_ADAPT_STRUCT` into *_impl.hpp 2016-02-02 11:30:00 +01:00
Dane Springmeyer
e76a1d9a81 [csv] use faster version of csv_utils::parse_line in more places + fix error output 2016-02-01 12:23:26 -08:00
Dane Springmeyer
d59460861e update to latest variant 2016-02-01 11:51:23 -08:00
artemp
33be145be9 variant - update to the latest upstream 2016-02-01 09:22:14 +01:00
Mickey Rose
958fcbf165 travis: try fewer jobs on osx 2016-01-31 15:44:25 +01:00
Mickey Rose
f9248914af travis: add SCons timings, fix comment in Makefile 2016-01-31 15:36:00 +01:00
Mickey Rose
4ef8efd554 remove struct member designators from initialization - fixes #3277
- it's C99 syntax, illegal in C++11 braced-init-list
- interestingly it's accepted in C++ by GCC and Clang,
  rejected by MSVC
2016-01-30 20:44:40 +01:00
Dane Springmeyer
57027eeee3 trim back what we build on OS X to try to save time 2016-01-29 20:04:42 -08:00
Dane Springmeyer
971c37691b travis/osx - parellelize the heavy compile jobs
- We can probably get away with this since OS X
   does not have an OOM killer
 - Will watch to see if it actually helps overall build times...
2016-01-29 19:00:04 -08:00
Dane Springmeyer
7ddb3cc4a2 try more JOBS=10 on osx 2016-01-29 18:32:14 -08:00
Dane Springmeyer
09ce7a0ea2 use xcode7.2 rather than 7.0.1 2016-01-29 18:31:35 -08:00
Mickey Rose
239623f976 remove unused variable [skip ci] 2016-01-30 03:24:07 +01:00
lightmare
e75342ac09 Merge pull request #3275 from lightmare/cleanup-noexcept-and-defaulted
Cleanup defaulted constructors and assignment operators, noexcept specifiers
2016-01-30 03:22:47 +01:00
Mickey Rose
a8d8a0d74f make clang happy - const default initialization
http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_active.html#253
http://stackoverflow.com/questions/7411515/why-does-c-require-a-user-provided-default-constructor-to-default-construct-a
http://stackoverflow.com/questions/21900237/do-i-really-need-to-implement-user-provided-constructor-for-const-objects
2016-01-30 02:03:25 +01:00
Mickey Rose
be5d772d6c attempt to fortify mapnik::value conversion construction
- including correct noexcept specifier
- adding conversion-assignment operator wasn't really necessary,
  but it might be more efficient since it avoids constructing an
  intermediate variant
2016-01-30 02:03:25 +01:00
Mickey Rose
400e05585f refine noexcept specifiers on forwarding conversion constructors 2016-01-30 02:03:25 +01:00
Mickey Rose
b2c14972de remove explicitly-defaulted constructors and assignment operators ...
... where they should be implicitly defined by the compiler
2016-01-30 02:03:25 +01:00
Dane Springmeyer
61ffe096fc latest variant 2016-01-29 16:32:09 -08:00