Age | Commit message (Collapse) | Author | |
---|---|---|---|
2012-04-02 | only include i64 precise code if it will actually be used | Alon Zakai | |
2012-03-29 | windows fixes | Alon Zakai | |
2012-03-26 | use shlex | Alon Zakai | |
2012-03-21 | fix some bugs with preloading of entire directories | Alon Zakai | |
2012-03-21 | properly handle closure compiler return codes | Alon Zakai | |
2012-03-21 | add option to add closure compiler arguments in EMCC_CLOSURE_ARGS env var | Alon Zakai | |
2012-03-20 | back out change to environment vars | Alon Zakai | |
2012-03-20 | back out change to cmake compiler stuff due to openjpeg breakage | Alon Zakai | |
2012-03-20 | Misc fixes for Windows. | Sigmund Vik | |
Most of these changes have to do with how python scripts are invoked. For Linux, 'Popen([EMCC] + args)' works because the first line in emcc is '#!/usr/bin/env python'. On Windows, the python interpreter has to be explicitly invoked, e.g. 'Popen(['python', EMCC] + args)'. Note that there is no harm in explicitly invoking the python interpreter on Linux, so this works on both platforms. For Windows, execvp() behaves differently than on Linux: http://mail.python.org/pipermail/python-list/2002-July/763863.html http://msdn.microsoft.com/en-us/library/3xw6zy53.aspx This causes many strange things to happen as the parent process terminated before its children. In this change the use of execvp() has been replaced with subprocess.call(). This change also fixes some code that assumed that the path separator always is '/', but for Windows it is '\'. And where the path module can be required, we use path.normalize() and path.resolve() to check if a filename is absolute in a platform agnostic manner. | |||
2012-03-15 | add file preloading in browsers | Alon Zakai | |
2012-03-08 | default closure compiler in a way that doesn't require EMSCRIPTEN_ROOT | Alon Zakai | |
2012-02-26 | do not use -std-compile-opts, which is not really meant to be used | Alon Zakai | |
2012-02-11 | keep i64 mode 1 in -O3, it is no longer slow, in fact it is faster now | Alon Zakai | |
2012-02-09 | enforce strict order in multiple libraries generated in the test runner: the ↵ | Alon Zakai | |
order does matter | |||
2012-02-09 | fix for bug with using old full pathnames when loading builds from cache in ↵ | Alon Zakai | |
test runner | |||
2012-02-09 | Merge pull request #228 from ehsan/use_ar_upstream | Alon Zakai | |
Use emar for creating static libraries and ld for linking | |||
2012-02-09 | Cache the result of is_ar | Ehsan Akhgari | |
2012-02-08 | Fix the cmake handling | Ehsan Akhgari | |
This enables us to use our own ar and ranlib for CMake based projects. It also lets us use the CFLAGS environment variable in CMake based projects if needed. It also fixes a bug introduced by using sys.argv instead of args. | |||
2012-02-08 | Add a test for handling of multiply defined symbols | Ehsan Akhgari | |
2012-02-08 | Improve get_library() to create real library files | Ehsan Akhgari | |
Also, enable caching of the libraries generated in tests in the new get_library implementation. | |||
2012-02-08 | Use llvm-ar instead of llvm-link in emar | Ehsan Akhgari | |
This makes the semantics of emar to be the same as the semantics of system ar. | |||
2012-02-06 | fix cmake typo | Alon Zakai | |
2012-02-06 | Use the more Pythonic syntax for substring matches | Ehsan Akhgari | |
2012-02-05 | show warning if using emscripten.py directly | Alon Zakai | |
2012-02-03 | output pretty printed from closure, since we do whitespace compression later ↵ | Alon Zakai | |
separately anyhow; pretty printed makes debugging easiser | |||
2012-02-03 | Merge pull request #212 from ehsan/cmake_support_2 | Alon Zakai | |
Support CMake in emconfigure | |||
2012-02-02 | Support CMake in emconfigure | Ehsan Akhgari | |
This adds a CMAKE_TOOLCHAIN_FILE argument to CMake based build systems so that they don't attempt to find headers, libraries and packages which Emscripten does not support | |||
2012-02-02 | accept empty bitcode files as valid bitcode files | Alon Zakai | |
2012-02-01 | llvm inlining support (disabled by default; enabled in benchmarks) | Alon Zakai | |
2012-02-01 | fix test_poppler and test_iostreamllvm-opts | Alon Zakai | |
2012-02-01 | refactor build checks for standalone and unsafe opts | Alon Zakai | |
2012-01-31 | fix usage of llvm opt level in emcc, and test_emcc fixes | Alon Zakai | |
2012-01-31 | Merge branch 'master' into llvmopts | Alon Zakai | |
2012-01-30 | save build_library logs in each project dir separately | Alon Zakai | |
2012-01-30 | do not do -Ox or LTO when building a linkable module | Alon Zakai | |
2012-01-29 | split LLVM LTO into separate pass | Alon Zakai | |
2012-01-29 | switch to -Ox from -std-compile-opts | Alon Zakai | |
2012-01-29 | enable poppler in o1 and o2, and some test_emcc fixes | Alon Zakai | |
2012-01-28 | Enable the EM_CONFIG environment variable to override ~/.emscripten | Ehsan Akhgari | |
This commit fixes issue #180. | |||
2012-01-23 | simplify llvm opt picking code, and allow up to level 3 | Alon Zakai | |
2012-01-23 | enable unsafe llvm opts in emcc | Alon Zakai | |
2012-01-21 | do not generate annotations when disassembling bitcode | Alon Zakai | |
2012-01-21 | disable pruning of emcc -s opts from test runner - it's too buggy | Alon Zakai | |
2012-01-20 | fix bug with Settings.default values not including -O, so comparisons were ↵ | Alon Zakai | |
wrong in some cases | |||
2012-01-20 | fix bug with missing keys in Settings.defaults | Alon Zakai | |
2012-01-20 | remove dead function eliminator, which is no longer needed | Alon Zakai | |
2012-01-20 | do not add unneeded -s 'es in test runner | Alon Zakai | |
2012-01-18 | LINKABLE options, disabling dead code elimination for shared libraries and ↵ | Alon Zakai | |
main files that have things that shared libraries will access through an extern | |||
2012-01-18 | use llvm internalive and globaldce to remove unneeded code before compiling ↵ | Alon Zakai | |
to js | |||
2012-01-18 | fix emcc bug with detecting bitcode after refactoring of llvm-nm calling code | Alon Zakai | |