Activity
From 11/03/2009 to 12/02/2009
12/02/2009
- 11:58 PM Revision 6c1741ea32d310b5ce1d466eb76fc269c55dcbf5: core/js: asserts.diag should stringify its arg...
- core/js: asserts.diag should stringify its argument
- 11:44 PM Revision bf7e88fd606df1fc99263e89d8f8cd3d14391cc4: core: exception now accepts a boost::format ar...
- core: exception now accepts a boost::format arg. (#44) Got fed up enough of to finally go and fix it.
-
11:07 PM Feature #178: Improve native_function_base constructor
- Currently it doesn't have access to the underlying @object@ in the constructor, which makes things a bit messy. See for example source:src/core/modules.cpp@b9b48725#L87 which cant be in the constructor as there is no object to set properties on...
-
- Kris Kowal suggested on irc the following API: ...
-
10:44 PM Feature #119: Write JS docs for XML plugin
- But now for the Arabica XML plugin!
-
- Getting hold of a file relative to the current module in a generic way involves doing some nasty regexen on module.uri. So lets give a nicer way to do this: ...
-
- commit:da90c1bf837a336445c4a83fe2332270311eace7
-
- commit:3cae7518dd930ff9dc3c01a7abf2ff19ee61ac23
-
- core: create<NativeFunction> now passes name and arity to the constructor of NativeFunction by default (fixes #176)
-
- core: add is_native/get_native for native functions (fixes #131)
-
- create<NativeFunction>(_name = "x", _arity = 1) should construct NativeFunction(1, std::string("x")).
-
- core: make is_native/get_native overloadable
-
- ChangeLog: #116 deserves a ChangeLog entry
-
- commit:84f5b2dc7872f71f350c10b6593d93cb2c8352f1
-
- shell/interpreter: look for modules in $HOME/.flusspferd/modules and $FLUSSPFERD_USER_MODULES (only for installed flusspferd, fixes #116)
-
- build/install: forgot to install util/load plugin
-
- commit:ed6f35a1576ca5540b4c68ff085f6625dcf629c0
-
- js/util: add load module (fixes #34)
-
- core: remove unnecessary include
-
- core: remove many of the superfluous flusspferd::gc calls that were inserted for debugging purposes
-
- Unfortunately not urgent enough. :-)
-
- commit:06362f1c2d232f52f8195a34be8d1d1ca249d524
-
- core: object::constructor/prototype/parent now work on const objects (fixes #175)
-
- Right now they are const-only, unnecessarily.
-
- commit:8da0f6d67d1c10d45f913ebfa2113cee3c84f460
-
- core: add object::constructor (fixes #172)
-
- Hmmm. I wonder if it is possible to make the catch/exception marshalling magic use the convert framework...
-
- In JS you can throw strings, numbers, random objects. Pretty much any value in fact. Our C++ bindings currently convert anything that isn't explicitly an exception subclass to a string. We should probably catch @value &@ and report the pending ...
-
- Made a start on this using arabica. Its largely complete now. Needs DOMException wrapping/exposing and marshalling when it gets thrown, and need to port the W3C DOM test suites to CommonJS format.
-
- We moved the old libxml2 based module out of master as it wasn't working. We really need XML and HTML parsing + DOM. With Tests. and without memory leaks
-
- https://developer.mozilla.org/en/SpiderMonkey/JSAPI_Reference/JS_GetConstructor Should have a method on object.
12/01/2009
-
- ChangeLog: add entry for #128
-
- commit:db7998e801a30e102537174acacf6e405f5506d8
-
- core: add instance_of (fixes #128)
-
- Is there any drawback to this? If not, we should go for it.
-
-
- commit:cbd68f8ef5a1eff53493f78c02e20317ddac888b
-
- core/getopt: fix #169
-
- Regression test: put it in tests/js/getopt.t.js ...
-
- Our current @require('test')@ is very similar to the ratified http://wiki.commonjs.org/wiki/Unit_Testing/1.0 Need to make the relevant changes to it so it complies.
-
07:19 PM Revision 816b53d5d0114f8e0b42dfbe7c5992c441a022b3: Merge branch 'master' of git://github.com/rued...
- Merge branch 'master' of git://github.com/ruediger/flusspferd
11/28/2009
-
- add info about modern compiler to README
11/25/2009
-
- Probably not something that we want to implement soon.
-
- Definitely a bug. I think we already noticed this some time and then forgot it again. Oh well.
-
- You forgot to set the proper target version. :-)
-
09:17 PM Bug #170 (Closed): Flusspferd doesn't work with Boost 1.41
-
- Boost.Fusion API changed in 1.41. vector0 => vector0<>. Fix #170
-
- The Boost.Fusion API changed in Boost 1.41 (not in the changelog though). vector0 is now vector0<>.
11/24/2009
-
- I expect @getopt@ aliases to work by being transparently converted to the canonical argument name, but the current behaviour is to stay as the aliased argument name in the returned hash: ...
11/23/2009
-
- Added note to getopt docs example.
11/22/2009
-
- "SWIG":http://www.swig.org/ is an interface generator. It generates interfaces to C and C++ libraries for other languages, like Python, Perl and so on. It would be great if we could provide a backend that generates interfaces for Flusspferd.
-
- Working on it, see branch implement_167
11/21/2009
-
- feature #167: implement obvious bugs, and add (failing) test
-
- feature #167: allow expressions as functions (untested so far)
-
- In a lot of cases, I wish to just expose a method of an underlying C++ object I am wrapping. Take the following example: ...
- 01:50 AM Revision 3d546c7513dc24c89bc6d956c1751c8be3fac170: core/js: add an asserts.throwsOk
- core/js: add an asserts.throwsOk
11/18/2009
- 07:41 PM Revision a20c8e6c44ca2474fc650b11d51c20c4ff25a106: Merge branch 'pdoc'
- Merge branch 'pdoc'
11/17/2009
- 11:52 PM Revision 217e7efa0c3f555701ac36cfe899285b701da04b: Message for when vendor/pdoc-template is missing
- Message for when vendor/pdoc-template is missing
- 11:29 PM Revision 3ec54f204bc61adb8086e71fea31fc9e6269eadf: Better error message for when gems isn't insta...
- Better error message for when gems isn't installed
- 07:24 PM Revision 57ca9e18d58fb5db5a345b99a73af0035daec977: Remove jsdocs - not used/needed anymore
- Remove jsdocs - not used/needed anymore
- 07:23 PM Revision 3770cf9d79db11c8601ac8a7bb9f499ec0ad9867: util/docs.sh now build pdocs, not old jsdocs
- util/docs.sh now build pdocs, not old jsdocs
-
- This will allow easier access to built in classes in modules without having to go grubbing about in the (possibly) polluted scope. This will be even more useful once "Spidermonkey bug 480850 - JSAPI should support creating objects given a cons...
11/16/2009
-
- Just fixed this by moving calls to INT_TO_JSVAL out of the header. It wasn't actually an ABI issue (JS_Assert is there in release builds.) It was a problem to do with default symbol resolution on OSX.
-
- This seems unlikely to happen for 0.9
-
- Rüdiger Sonderfeld wrote: > I don't like 2. Since it would add part of the API exclusively to flusspferd.hpp. You than just couldn't replace #include <flusspferd.hpp> by a finer selection of headers you'd also have to add the aliases yourself. ...
-
- Removing target version on rejected bug.
-
- Removing target version - its been rejected so not related to a release.
- 10:34 PM Revision 15c22379bef296b5dc0c5e9311ce824860c0696b: core/sm: move call to INT_TO_JSVAL out of head...
- core/sm: move call to INT_TO_JSVAL out of header file INT_TO_JSVAL is actually a function which calls back JS_Assert under a debug build. This caused link issues on OSX (and probably with MSVC too)
-
- If we can, we should do this API-breaking thing in 0.9.
- 09:25 PM Revision e369a8395a1833932295c37b95a786f14473a113: more fixes / cleanups in the name of compiling...
- more fixes / cleanups in the name of compiling on OS X It seems that Apple's gcc doesn't like the fn(root(...)) pattern. Issue raised to work around this: #165
-
- Idea 2 would not work because each copy has a different gcptr. But generally I like this, because it would make a lot of things easier.
-
- So the changes that Aristid landed as part of #147 seemed to cause compilation problems on OSX - related to a probably bug in Apple's build of GCC. Ignoring the fact that there is a bug in that gcc, the problem was that roots are not currently cop...
-
- core: fixes / cleanups in the name of compiling on OS X
-
- core: conservatively root getter/setter functions
11/15/2009
-
- ChangeLog: clean up
-
- core: remove deprecated exception::empty
- 12:03 AM Revision a96674be13e4dd8c567d0e8d055caefda24142a8: Fix up broken pdoc references
- Fix up broken pdoc references
11/14/2009
- 11:35 PM Revision c194b79826c0e83afff8c5a6c78ac7c5b392b49f: Add the pdoc templates as a submodule
- Add the pdoc templates as a submodule
- 11:22 PM Revision 2cca3d51902be85bd5a32013d2cf7f6e02f2897d: Update vendor/pdoc to fix issue with borken links
- Update vendor/pdoc to fix issue with borken links
-
- ChangeLog: set_properties/define_properties
-
- test: improve object::define_properties test
-
- commit:273a43ff2ef623606a2323835c5bbbf3814f58de Complete with docs and tests. * "define_properties":http://flusspferd.org/docs/classflusspferd_1_1object.html#405119caada76b2f999ebc1a71454dd5 * "set_properties":http://flusspferd.org/docs/cla...
-
- object::set_properties / object::define_properties (fixes #151)
- 06:37 PM Revision d1884f0bfc9cf01c9e92767cfe92b55a905f8a01: Convert GMP docs to pdoc format and expand the...
- Convert GMP docs to pdoc format and expand the copy.
-
- commit:96a358d7fdd0d180e7da390ae1744d70ba44a66d
-
- core/module: implement toString/toSource for exports object (fixes #153)
-
- Pending the conversion overhaul, I'd like to see roughly the following API: ...
-
- So the GMP plugin is functionally complete, but the API could be improved: * Deprecate the functions with '_' in the name (these are to mirror the C api) – for now this means just don't document them * Remove @asserts@ in API functions: @Integ...
-
- doc: remove obsolete module
-
- doc: fix documentation typo
-
- Implemented, documented and tested. Yay! commit:0395659a5d5e694a5b6269b4aa74b3d6f9c070bc
-
- Merge branch 'new_create' (fixes #147) Conflicts: test/test_object.cpp
-
- new_create/doc: two more examples
-
- new_create/doc: specify result type
-
- new_create/doc: native_function
-
- new_create/doc: native_object
-
- new_create/doc: stub docs directly on the named parameters
-
- new_create/doc: function
-
- new_create/doc: array
-
- new_create/doc: header files and parameters for create<object>
-
- new_create: more documentation and a fresh idea for how to document it
- 02:10 AM Revision 9f74fd5a397d1c6a0279dc8a31050f7cb501b942: Tidy up the pdoc runner script PDoc is now st...
- Tidy up the pdoc runner script PDoc is now stored as a git module, and the code to give nice behaviour when its not there was shamelessly stolen from Prototype itself. Still to do: get the templates somewhere too.
- 02:08 AM Revision 4844321ab2dd32679eeb01427549cd5e9d5a8e73: PDocs for JSON object
- PDocs for JSON object
- 02:08 AM Revision 3c5215a8970ff5e3e1d541a655691c5b5402763a: Get all .js files (in main source) parsable vi...
- Get all .js files (in main source) parsable via PDoc
-
- new_create: adapt rest of c++ tutorials
-
- new_create: fix tutorial examples
11/13/2009
- 11:46 PM Revision 3e14527a032b72a50c1b8da8fa78cd0e497de2ff: Removed deletion of global XML object and disa...
- Removed deletion of global XML object and disabling of E4X Signed-off-by: Ash Berlin <ash_github@firemirror.com>
-
- new_create: document common parameters
-
- new_create/doc: a bit of text and some examples
-
- new_create: add stub documentation
-
- Superseded by #147
-
- commit:91980b97e1cfd475c6f377c3addc7b1243446d52 commit:3a42f2e9c8598903e5b2844665995973751bf5a2
-
- ChangeLog: add changelog entry for extendLeft/Right (#160)
-
- core/binary: add extendLeft/Right (fixes #160)
-
- Need docs, but otherwise pretty good looking.
-
- new_create: get rid of legacy function creation stuff
-
- new_create: fix test
-
- new_create: fix invocation of create<method>
-
- new_create: fix first crash
-
- new_create: remove create_native_{function,method} (now with CRASHING)
-
- new_create: function: support other member functions
-
- new_create: method: support void (T::*)(call_context&)
-
- new_create: support arbitrary function prototypes (except member functions)
-
- new_create: function signature deduction
-
- new_create: add new test for function creation
11/12/2009
-
- Sometimes, local_root_scope does not seem to work correctly in combination with GC Zeal. Find the problem and isolate the test case. If possible, fix it or report to Spidermonkey folks. (10% because we have implicitly observed the problem al...
-
-
- test: tiny fix in test
-
- new_create: so that's the local_root_scope+GC Zeal bug...?
-
- Merge branch 'master' into new_create Conflicts: include/flusspferd/class_description.hpp include/flusspferd/create.hpp src/core/binary.cpp src/core/getopt.cpp src/core/load_core.cpp src/core/modules.cpp src/core/security.cpp
-
- commit:21c9aa3cd86abb22b3c3c7a1752aff4397748c71 Please see if there are still problems somewhere. Also, we should find out if there are problems with local_root_scope.
-
- fix_162: all tests successful
-
- fix_162: enough to start the interactive shell
-
- fix_162: fix property_iterator constructor
-
- fix_162: rooting fix in create_native_function (very subtle!)
-
- fix_162: next step: load_properties_functions
-
- fix_162: more careful rooting in modules.cpp
-
- fix_162: next step: load_core
-
- Merge branch 'fixes_from_propattrs3' into fix_162
-
- fix_162: this tiny change fixes the crash in test_value. WTF
-
- core: harden object::*
-
- fix_162: a few changes to help debugging of bug #162
-
- Since GC Zeal was enabled in commit:8e86bb74573c8608e456caacb056373d9690fe70, Flusspferd crashes.
- 12:56 PM Revision dbb6ca1cf16f9d9563b7c20c9531a65623b12983: Add note about how require is available
- Add note about how require is available
11/11/2009
-
- Speaking with jorendorff in #jsapi, it turns out that when you create a property with a getter/setter you should use the JSPROP_SHARED flag to avoid spidermonkey needlessly creating an extra slot. Its not much memory, but it makes sense to do and ...
11/10/2009
- 03:43 AM Revision 6e4314dafbbde54d50067134af46f6245aa37233: Convert require docs to pdoc format
- Convert require docs to pdoc format
- 03:31 AM Revision 8feec8b02da7c53758a27cc7f149cb033e916791: Docs for the system module
- Docs for the system module
- 03:31 AM Revision dffe228ce353e33576b225222ad98331c7193e30: JS docs for getopt
- JS docs for getopt
11/09/2009
- 04:44 PM Revision 715136176c34a06f92651628e767cd928d5f1efe: Docs for flusspferd module
- Docs for flusspferd module
11/08/2009
- 10:54 PM Revision 8985e813639bc279a58b20b8f2470866fac31bf0: Add JS api docs for encodings module
- Add JS api docs for encodings module
- 01:43 AM Revision 52b494075af2bc0d46fc794f1700604c4f83d163: Binary JS api docs
- Binary JS api docs
- 01:43 AM Revision caf9a8dcd5d74df50225420f9e689a100d6230a8: Convert SQLite3 docs to pdoc format
- Convert SQLite3 docs to pdoc format
- 01:43 AM Revision 960a5e28b90f0bb5b37326dec3f170b551f5b321: Update JS API docs for fs-base module.
- Update JS API docs for fs-base module.
-
- We have them as @prepend@ and @append@ which are better names to my mind, but we should support the standard. We need to add the methods as aliases in the class definition and in the new pdocs.
11/07/2009
-
- Aristid Breitkreuz wrote: > Hello Aaron, > > What does "it crashes" mean? > > Please provide the output of the program. > > > Aristid ...
-
- Hello Aaron, What does "it crashes" mean? Please provide the output of the program. Aristid
-
02:49 PM Bug #159: Flusspferd crashes when calling js function returned from require() within a c++ callback
- How to reproduce: http://gist.github.com/228710 Originated from here: http://groups.google.com/group/flusspferd/browse_thread/thread/ff07c6da747b3d3c
11/06/2009
-
- docs: Ubuntu 9.04 -> 9.10
11/05/2009
-
- doc: mention that ebuilds are not in gentoo repo
Also available in: Atom