Jean Privat [Wed, 2 Jul 2014 17:11:15 +0000 (13:11 -0400)]
tests: test new policy in test_superstring
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 2 Jul 2014 17:09:07 +0000 (13:09 -0400)]
typing: refuse a null value in superstrings
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 2 Jul 2014 06:04:51 +0000 (02:04 -0400)]
Merge: Enable dummy executions of mnit applications
A dummy execution need a stream of input events and a dummy display that is not connected to a physical screen.
`lib/mnit/mnit_injected_events.nit` gives the injection of pseudo-events read from a file.
It can be used, by mixin-refinement, with any existing mnit programs.
Eg. to have a replay of a game, just compile with `-m` and run with some environment variables
~~~
$ cd examples/shoot
$ ../../bin/nitc -o bin/shoot src/shoot_linux.nit -m ../../lib/mnit/mnit_injected_input.nit
$ MNIT_SRAND=0 MNIT_READ_INPUT=../../tests/shoot_null.inputs bin/shoot
~~~
`lib/mnit/mnit_null.nit` gives a new dummy target-platform with no display.
It must be used instead of `mnit_linux` and `mnit_android`.
Eg. for the same run, but without display ; note that the main module is `shoot.nit` instead of `shoot_linux.nit`
~~~
$ cd examples/shoot
$ ../../bin/nitc -o bin/shoot src/shoot.nit -m ../../lib/mnit/mnit_null.nit
$ MNIT_SRAND=0 MNIT_READ_INPUT=../../tests/shoot_null.inputs bin/shoot
~~~
All this is mainly a proof of concept, and not that useful, except maybe for testing, it is why tests.sh has be extended to control the run of a dummy `shoot` game.
Pull-Request: #509
Reviewed-by: Alexis Laferrière <alexis.laf@xymus.net>
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Jean Privat [Wed, 2 Jul 2014 05:56:43 +0000 (01:56 -0400)]
tests: update mnit
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 2 Jul 2014 05:54:26 +0000 (01:54 -0400)]
Merge master to resolve conflict in tests not present yet
Jean Privat [Wed, 2 Jul 2014 05:52:22 +0000 (01:52 -0400)]
Merge: String api
Modified a bit of the API for Strings/Buffers as a reaction to the `appendgate` that occured earlier this week.
Now, copy methods (+,*,to_upper/lower) are deferred to String and in-place equivalents are available for Buffer.
Pull-Request: #516
Reviewed-by: Jean Privat <jean@pryen.org>
Reviewed-by: Alexis Laferrière <alexis.laf@xymus.net>
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Jean Privat [Tue, 1 Jul 2014 19:23:40 +0000 (15:23 -0400)]
Merge: model: fix a compiler crash on access to null constructor signature
Before this patch, compilation process stopped on a null pointer exception.
Because of the compilation workflow, nitg tried to access a super init even
if this init was malformed (type error) and thus not modelized properly.
This commit add one more condition to ensure that msignature is initialized before the test
and raise the correct error instead of crashing.
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Pull-Request: #538
Reviewed-by: Jean Privat <jean@pryen.org>
Jean Privat [Mon, 16 Jun 2014 15:38:50 +0000 (11:38 -0400)]
tests_full: add mnit_null variants, and the .inputs file for shoot
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Mon, 16 Jun 2014 15:37:34 +0000 (11:37 -0400)]
tests: teach tests.sh to use MNIT_* environment variables
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Mon, 16 Jun 2014 15:30:28 +0000 (11:30 -0400)]
shoot: add a `null` target that uses mnit::mnit_null
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Mon, 16 Jun 2014 15:29:10 +0000 (11:29 -0400)]
lib/mnit: introduce `mnit_null` to execute mnit without a display
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Tue, 1 Jul 2014 19:05:24 +0000 (15:05 -0400)]
lib/mnit: introduce `mnit_injected_input` to inject events into mnit applications
Signed-off-by: Jean Privat <jean@pryen.org>
Alexandre Terrasa [Tue, 1 Jul 2014 16:17:17 +0000 (12:17 -0400)]
model: fix a compiler crash on access to null constructor signature
Before this patch, compilation process stopped on a null pointer exception.
Because of the compilation workflow, nitg tried to access a super init even
if this init was malformed (type error) and thus not modelized properly.
This commit add one more condition to ensure that msignature is initialized before the test
and raise the correct error instead of crashing.
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Jean Privat [Tue, 1 Jul 2014 12:52:33 +0000 (08:52 -0400)]
Merge: java: intro pop_from_local_frame
A very useful features when using the JNI reference stack to clean up local references.
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Pull-Request: #533
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Jean Privat [Tue, 1 Jul 2014 12:51:01 +0000 (08:51 -0400)]
Merge: Improvements to Pep8analysis Web version
* Support C++ in the emscripten platform
* The emscripten platoform support for the `--release` option reduces the file size of 50%
* pep8analysis error messages are now friendlier
* Do not spend as much time trying to find a fixed point (which is much faster!)
* Use Google Analytics to collect data and events
* Report fatal crash of the tool to the user
* Notice the user when the tool is working
* Does not lauch the full main for each analysis, only call a single Nit method
* Launch computation async using web workers when supported
The stack size of web workers are limited in Google Chrome under Linux, you are better off using Firefox to test big programs. Maybe they should be deactivated by default under this OS and browser.
Pull-Request: #530
Reviewed-by: Jean Privat <jean@pryen.org>
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Alexis Laferrière [Fri, 27 Jun 2014 18:13:17 +0000 (14:13 -0400)]
java: intro pop_from_local_frame
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 18:43:14 +0000 (14:43 -0400)]
pep8analysis: do work async using workers
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:34:06 +0000 (12:34 -0400)]
pep8analysis: Web interface launches only the analysis (not the full main)
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:15:57 +0000 (12:15 -0400)]
pep8analysis: report fatal crash of the tool
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:14:08 +0000 (12:14 -0400)]
pep8analysis: fix use of #consolePanel and its code block
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:13:21 +0000 (12:13 -0400)]
pep8analysis: add "Working..." panel to Web interface
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 00:38:44 +0000 (20:38 -0400)]
pep8analysis: add some google analytics event tracking
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Mon, 16 Jun 2014 20:20:03 +0000 (16:20 -0400)]
pep8analysis: set larger maximum memory
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:34:43 +0000 (12:34 -0400)]
pep8analysis: minor correction in text
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 00:38:20 +0000 (20:38 -0400)]
pep8analysis: minor indentation fix
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 18:40:50 +0000 (14:40 -0400)]
pep8analysis: clean up include of Google Analytics
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 00:53:48 +0000 (20:53 -0400)]
pep8analysis: remove duplicated code
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Tue, 17 Jun 2014 20:40:50 +0000 (16:40 -0400)]
pep8analysis: improve error messages
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 18:40:39 +0000 (14:40 -0400)]
pep8analysis: short format of CFG contains code only
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:21:44 +0000 (12:21 -0400)]
pep8analysis: make `manager` a top level method
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:17:47 +0000 (12:17 -0400)]
pep8analysis: suggest garbage collection between phases
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:17:12 +0000 (12:17 -0400)]
pep8analysis: do not crash on expected errors
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:09:18 +0000 (12:09 -0400)]
pep8analysis: reduce the max amount of looping to find fixed point
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:32:52 +0000 (12:32 -0400)]
emscripten: minify code when release
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Jean Privat [Thu, 26 Jun 2014 19:51:50 +0000 (15:51 -0400)]
Merge: No old style accessors
Old-style accessors are those in the form
~~~
readable writable var _foo: Foo
~~~
The plan is to remove them from the grammar (and the AST and modelize_property) in future PR.
This PR only update the exisiting code to remove the remaining occurences by:
* using new-style attributes, or
* keep old style attributes with explicit getters/setters if using new-style attributes has a too big impact on the performances (or is too complex to update)
Because of performance implications, old-style attribute may remains in the language to implements some intern parts.
Pull-Request: #528
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Alexis Laferrière [Thu, 26 Jun 2014 16:32:16 +0000 (12:32 -0400)]
emscripten: use the macro CXX to support C++
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:11:18 +0000 (12:11 -0400)]
emscripten: add GC controls
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Mon, 16 Jun 2014 20:20:21 +0000 (16:20 -0400)]
emscripten: enable dynamic memory size
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexis Laferrière [Thu, 26 Jun 2014 16:12:11 +0000 (12:12 -0400)]
FFI with C++: use a customizable macro to compile C++ code
Signed-off-by: Alexis Laferrière <alexis.laf@xymus.net>
Jean Privat [Thu, 26 Jun 2014 17:16:08 +0000 (13:16 -0400)]
Merge: private DFAState
nitcc genrates a bunch of subclasses of DFAState, but did not mark them private altrought they are not useful for the programme and polute the auto-documentation.
Pull-Request: #529
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Jean Privat [Thu, 26 Jun 2014 17:16:00 +0000 (13:16 -0400)]
Merge: PNaCl: Added 'create_thread' function in the pnacl lib.
Added 'create_thread' function to enable thread creation after a exit.
Pull-Request: #519
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Reviewed-by: Alexis Laferrière <alexis.laf@xymus.net>
Reviewed-by: Jean Privat <jean@pryen.org>
Jean Privat [Thu, 26 Jun 2014 17:15:50 +0000 (13:15 -0400)]
Merge: Concerns sorter
Just replace some sorters by an MEntity sorter.
Pull-Request: #526
Reviewed-by: Jean Privat <jean@pryen.org>
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Alexandre Terrasa [Thu, 19 Jun 2014 20:03:34 +0000 (16:03 -0400)]
model_utils: replace many sorters by MEntityNameSorter
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Alexandre Terrasa [Wed, 18 Jun 2014 18:24:51 +0000 (14:24 -0400)]
model_utils: add MEntityNameSorter
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Jean Privat [Thu, 26 Jun 2014 10:05:47 +0000 (06:05 -0400)]
modelize_property: remove old-style accesors
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Thu, 26 Jun 2014 10:04:06 +0000 (06:04 -0400)]
pep8analysis: remove old-style accessors in the AST
Note: the transformation was done with a script.
The parser cannot seem to be regenerated.
A future move could be to use nitcc instead.
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Thu, 26 Jun 2014 02:13:50 +0000 (22:13 -0400)]
tests: remove remaining old-style accessors
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Thu, 26 Jun 2014 09:18:48 +0000 (05:18 -0400)]
lib/json: regenerate json_lexer with private DFAStates
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Thu, 26 Jun 2014 09:17:36 +0000 (05:17 -0400)]
nitcc: DFAStates are private
One does not want to see them in the documentation.
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Thu, 26 Jun 2014 02:42:24 +0000 (22:42 -0400)]
Merge: ordered_tree: make OrderedTree implement Collection
OrderedTree now implement major Collection services
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Pull-Request: #524
Reviewed-by: Alexis Laferrière <alexis.laf@xymus.net>
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Reviewed-by: Jean Privat <jean@pryen.org>
Jean Privat [Thu, 26 Jun 2014 02:06:23 +0000 (22:06 -0400)]
parser: remove generated old-style accesors.
ANodes use new style attributes (with two accesors)
Other use old-style attributes with two explicit getter and setters.
Using new-style attributes has a too big impact on performance because of
the redefined setter
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Thu, 26 Jun 2014 02:02:30 +0000 (22:02 -0400)]
parser: remove some remaining old style accessors
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Thu, 26 Jun 2014 01:10:11 +0000 (21:10 -0400)]
Merge: names and concerns in model
First commit adapt the MEntity implementation to it's documentation by moving up the `name` property.
Second commit introduce a new kind of MEntity: MConcern that will be used in documentation tools as a common interface for MProject, MGroup and MModule.
Pull-Request: #522
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Reviewed-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 25 Jun 2014 23:49:29 +0000 (19:49 -0400)]
Merge: Model utils
This pull request needs #522 to be merged.
Introduce some usefull services in model_utils for further uses in documentation tools.
Pull-Request: #523
Reviewed-by: Jean Privat <jean@pryen.org>
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Reviewed-by: Alexis Laferrière <alexis.laf@xymus.net>
Alexandre Terrasa [Wed, 25 Jun 2014 23:12:26 +0000 (19:12 -0400)]
ordered_tree: add some tests
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Djomanix [Wed, 25 Jun 2014 23:09:19 +0000 (19:09 -0400)]
Updated 'create_thread' function comments.
Signed-off-by: Djomanix <johan.kayser@viacesi.fr>
Alexandre Terrasa [Wed, 25 Jun 2014 22:55:42 +0000 (18:55 -0400)]
ordered_tree: make OrderedTree a subclass of Collection and implements missing methods
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Jean Privat [Wed, 25 Jun 2014 22:07:11 +0000 (18:07 -0400)]
Merge: Rope api
Small modifications of the API that will be required for future pull requests such as Splay ropes and Bufferized ropes.
Pull-Request: #518
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Reviewed-by: Jean Privat <jean@pryen.org>
Alexandre Terrasa [Thu, 19 Jun 2014 19:52:51 +0000 (15:52 -0400)]
model: introduce MConcern
Each MProject, MGroup and MModules represents a documentable concern.
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Alexandre Terrasa [Wed, 25 Jun 2014 21:32:34 +0000 (17:32 -0400)]
model: introduce `name` in MEntity according to it's documentation
Documentation was:
# A named and possibly documented entity in the model.
So I added the name...
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Jean Privat [Wed, 25 Jun 2014 19:56:36 +0000 (15:56 -0400)]
Merge: Removed sockets from 'debugger.nit' to put them in a new module
Removed sockets from 'debugger.nit' to put them in 'debugger_socket.nit', because it caused compilation problems for the interpreter in PNaCl.
Pull-Request: #520
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Reviewed-by: Alexis Laferrière <alexis.laf@xymus.net>
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Reviewed-by: Jean Privat <jean@pryen.org>
Alexandre Terrasa [Wed, 18 Jun 2014 19:14:08 +0000 (15:14 -0400)]
model_utils: allow intro and redef mclasses filtering
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Alexandre Terrasa [Wed, 25 Jun 2014 15:24:43 +0000 (11:24 -0400)]
model_utils: collect mmodules in mproject and mgroups
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Alexandre Terrasa [Wed, 25 Jun 2014 15:24:07 +0000 (11:24 -0400)]
model_utils: extract mclasses and mclassdefs from MProject and Mgroup
Signed-off-by: Alexandre Terrasa <alexandre@moz-code.org>
Lucas Bajolet [Mon, 23 Jun 2014 15:38:08 +0000 (11:38 -0400)]
lib/ropes_debug: Adapted new leaves for to_dot operation.
Signed-off-by: Lucas Bajolet <r4pass@hotmail.com>
Lucas Bajolet [Mon, 23 Jun 2014 15:12:35 +0000 (11:12 -0400)]
lib/standard/ropes: Added to_leaf method on RopeNode.
Signed-off-by: Lucas Bajolet <r4pass@hotmail.com>
Djomanix [Wed, 25 Jun 2014 18:30:36 +0000 (14:30 -0400)]
Added an import for 'debugger_socket'.
Signed-off-by: Djomanix <johan.kayser@viacesi.fr>
Jean Privat [Wed, 25 Jun 2014 13:12:50 +0000 (09:12 -0400)]
mailcap: fix Julien Pagès
Signed-off-by: Jean Privat <jean@pryen.org>
Djomanix [Mon, 23 Jun 2014 20:06:03 +0000 (16:06 -0400)]
Put sockets in 'debugger_socket.nit
Signed-off-by: Djomanix <johan.kayser@viacesi.fr>
Djomanix [Mon, 23 Jun 2014 20:05:11 +0000 (16:05 -0400)]
Removed sockets from 'debugger.nit' to put them in 'debugger_socket.nit
Signed-off-by: Djomanix <johan.kayser@viacesi.fr>
Jean Privat [Mon, 23 Jun 2014 19:30:53 +0000 (15:30 -0400)]
Merge: Nit for mac
Some minor, but needed, tweaks to allow the compilation of Nit on Mac OS X
A special page is also created: http://nitlanguage.org/on_mac/
Currenlty, only 31 tests does not pass, most of them are platform or FFI related (some SDK are missing).
Pull-Request: #517
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Jean Privat [Mon, 23 Jun 2014 19:30:45 +0000 (15:30 -0400)]
Merge: Tests emscripten
The produced .js files are then executed by nodejs.
All base tests that do not import kernel are skipped because
since the module lib/emscripten.nit imports the kernel module,
an automatic double definition of classes `Object` and `Sys` occurs.
nodejs cannot perform correctly synchronized input, thus `stdin` and all
file reading does not work although they work in a browser.
As a workaround, a special module `emscripten_nodejs.nit` is added that just
print "NOT_YET_IMPLEMENTED" on unsupported services on synchronized input,
thus marking the test as skipped.
Currently, 99 (7% of the 1390 tests) still cause some other kind of failures;
they will be investigated later.
The whole testing of all the tests takes 68m on my machine.
ccache in unusable and the option --jcache of emcc does not bring any boost.
So it is unlikely that `testall.sh` will include this engine.
Pull-Request: #514
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Jean Privat [Mon, 23 Jun 2014 19:29:49 +0000 (15:29 -0400)]
Merge: Bench scripts
Reunited common functions for benches in a new shell file.
Signed-off-by: Lucas Bajolet <r4pass@hotmail.com>
Pull-Request: #515
Reviewed-by: Jean Privat <jean@pryen.org>
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Djomanix [Mon, 23 Jun 2014 18:28:09 +0000 (14:28 -0400)]
Added 'create_thread' function to enable thread creation after a exit.
Signed-off-by: Djomanix <johan.kayser@viacesi.fr>
Lucas Bajolet [Mon, 23 Jun 2014 13:50:40 +0000 (09:50 -0400)]
tests: Updated to fit new API.
Signed-off-by: Lucas Bajolet <r4pass@hotmail.com>
Lucas Bajolet [Mon, 23 Jun 2014 13:48:53 +0000 (09:48 -0400)]
lib/standard/text: Updated API for String/Buffer, a few methods disappear from Text and are relocated a level below.
Signed-off-by: Lucas Bajolet <r4pass@hotmail.com>
Jean Privat [Sat, 21 Jun 2014 02:31:07 +0000 (22:31 -0400)]
mkcsrc: clear OLDNITCOPT in src/Makefile
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Fri, 20 Jun 2014 19:27:12 +0000 (15:27 -0400)]
c_src: update Makefile to compile on MacOSX
Add a make variable OLDNITCOPT to give options for the first step
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Fri, 20 Jun 2014 20:03:09 +0000 (16:03 -0400)]
tests: update some tests to pass on Mac
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Fri, 20 Jun 2014 19:39:16 +0000 (15:39 -0400)]
Makefile: use `--no-stacktrace` on c_src/nitg to deal with the -lunwind issue on MacOSX
Jean Privat [Fri, 20 Jun 2014 19:28:23 +0000 (15:28 -0400)]
tests.sh: force LC_ALL that cause issues with MacOSX
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Fri, 20 Jun 2014 19:25:05 +0000 (15:25 -0400)]
global: compiler remove a useless `const` that cause C warnings
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Fri, 20 Jun 2014 19:19:33 +0000 (15:19 -0400)]
abstract_compiler: remove -lunwind if target is Darwin (macosx)
Jean Privat [Fri, 20 Jun 2014 19:18:01 +0000 (15:18 -0400)]
abstract_compiler: add -Qunused-argument if clang
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Fri, 20 Jun 2014 19:13:20 +0000 (15:13 -0400)]
abstract_compiler: add some flags to teach clang to be more silent
gcc also accepts these flags, so it is OK.
Signed-off-by: Jean Privat <jean@pryen.org>
Lucas Bajolet [Wed, 18 Jun 2014 19:38:06 +0000 (15:38 -0400)]
benchmarks: Reunited common functions for benches in a new shell file.
Signed-off-by: Lucas Bajolet <r4pass@hotmail.com>
Lucas Bajolet [Thu, 19 Jun 2014 17:19:07 +0000 (13:19 -0400)]
lib/standard/ropes: Concat nodes cannot be constructed without children anymore.
Signed-off-by: Lucas Bajolet <r4pass@hotmail.com>
Lucas Bajolet [Thu, 19 Jun 2014 16:13:49 +0000 (12:13 -0400)]
lib/standard/ropes: API change, Leaves are now abstract for future flexibility.
Signed-off-by: Lucas Bajolet <r4pass@hotmail.com>
Jean Privat [Thu, 19 Jun 2014 15:01:24 +0000 (11:01 -0400)]
Merge: Tests nitg --semi-global
Some work and cleanup on tests.sh and cie.
Also add a new engine `nitg-sg` that is `nitg --semi-global`
Pull-Request: #513
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Reviewed-by: Alexandre Terrasa <alexandre@moz-code.org>
Jean Privat [Wed, 18 Jun 2014 20:01:21 +0000 (16:01 -0400)]
tests.sh: new engine `emscripten` for `nitg -m emscripten`
The produced .js files are then executed by nodejs.
All base tests that do not import kernel are skipped because
since the module lib/emscripten.nit imports the kernel module,
an automatic double definition of classes `Object` and `Sys` occurs.
nodejs cannot perform correctly synchronized input, thus `stdin` and all
file reading does not work although they work in a browser.
As a workaround, a special module `emscripten_nodejs.nit` is added that just
print "NOT_YET_IMPLEMENTED" on unsupported services on synchronized input,
thus marking the test as skipped.
Currently, 99 (7% of the 1390 tests) still cause some other kind of failures;
they will be investigated later.
The whole testing of all the tests takes 68m on my machine.
ccache in unusable and the option --jcache of emcc does not bring any boost.
So it is unlikely that `testall.sh` will include this engine.
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 18 Jun 2014 16:51:25 +0000 (12:51 -0400)]
tests: testall.sh include nitg-sg
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 18 Jun 2014 16:42:05 +0000 (12:42 -0400)]
tests.sh: distinguish the generated file from the executed file
Some some future engines that does not produce directly executable files
could still use the tests.sh framework.
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 18 Jun 2014 14:23:17 +0000 (10:23 -0400)]
tests: add engine `nitg-sg` for --semi-global
Some platforms starts using it, so better check that we do
not broke it.
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 18 Jun 2014 15:19:53 +0000 (11:19 -0400)]
tests: clean some .skip files (use some UNDEFINED .res)
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 18 Jun 2014 15:05:30 +0000 (11:05 -0400)]
tests: teach a res file to contains `UNDEFINED` and always match
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 18 Jun 2014 14:20:57 +0000 (10:20 -0400)]
tests: add tests for tests.sh so we can test the test script
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 18 Jun 2014 14:17:25 +0000 (10:17 -0400)]
tests.sh: soso are always detected as failed tests
A `soso` is when a test fails only because errors and warnings
differs.
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 18 Jun 2014 13:59:21 +0000 (09:59 -0400)]
tests.sh: less hackish way to handle fixme directories for savdirs
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 18 Jun 2014 13:57:00 +0000 (09:57 -0400)]
tests.sh: remove tap output
Badly implemented and unused now since junit-xml-output
Signed-off-by: Jean Privat <jean@pryen.org>
Jean Privat [Wed, 18 Jun 2014 15:25:32 +0000 (11:25 -0400)]
tests: fix base_inline
Signed-off-by: Jean Privat <jean@pryen.org>
nitlanguage / nit.git / log