summaryrefslogtreecommitdiff
path: root/giscanner
Commit message (Collapse)AuthorAgeFilesLines
* When handling errors according to errno, catch both IOError and OSErrorSimon McVittie2017-06-192-16/+10
| | | | | | | | | | | | | | | | | | | | | | | | Different Python versions are not completely consistent about the error that is raised and its class hierarchy: Python 3.5.3rc1 (default, Jan 3 2017, 04:40:57) >>> try: open('/foo') ... except Exception as e: print(e.__class__.__mro__) (<class 'FileNotFoundError'>, <class 'OSError'>, <class 'Exception'>, <class 'BaseException'>, <class 'object'>) Python 2.7.13 (default, Dec 18 2016, 20:19:42) >>> try: open('/foo') ... except Exception as e: print e.__class__.__mro (<type 'exceptions.IOError'>, <type 'exceptions.EnvironmentError'>, <type 'exceptions.StandardError'>, <type 'exceptions.Exception'>, <type 'exceptions.BaseException'>, <type 'object'>) This can lead to a race condition during cache cleaning, where two processes both try to delete the same file, and the one that loses the race fails. Signed-off-by: Simon McVittie <smcv@collabora.com> Reviewed-by: Iain Lane <laney@ubuntu.com> Reviewed-by: Colin Walters <walters@verbum.org> Bug: https://bugzilla.gnome.org/show_bug.cgi?id=772173
* giscanner/utils.py: Fix on non-UNIX-like platformsChun-wei Fan2017-06-151-1/+1
| | | | | | | | | | Commit 85b1d70b93211415d975deff6050f1401743e72d changed the null device from the platform-independent os.devnull to the UNIX /dev/null, which broke introspection builds on non-UNIX-like platforms, as /dev/null is a notion that does not exist on those platforms. Fix this by changing back to os.devnull, but still opening it as writable. https://bugzilla.gnome.org/show_bug.cgi?id=781312
* scanner: Fix non-libtool linker flags on DarwinPhilip Chimento2017-05-171-1/+1
| | | | | | | | | Darwin's linker doesn't like "-rpath=path"; instead pass "-rpath path", which works on more linkers than the version with the = sign does. Regressed in commit 5d4cd25292b8ed2c7a821ebe19fc5ab5d447db1a. https://bugzilla.gnome.org/show_bug.cgi?id=781525
* giscanner: Do not use distutils for linkingChun-wei Fan2017-05-032-115/+86
| | | | | | | | | | | | | | | | | | | | | | | | ...the dumper program for all cases. It turned out that using distutils for linking is more troublesome than useful as we need to ensure that the paths specified by -L need to come before the standard library search paths, and distutil's ccompiler.add_library_path() and ccompiler.add_runtime_library_path() does not work for all of its supported compilers (Visual Studio is an example). Instead, we go back to constructing our linker command line manually as we did before (and as we now do in the libtool case), but with some enhancements: -Use '-libpath:' on Visual Studio builds, which corresponds to the -L flag on GCC/CLang. -Extend LIB/PATH (Windows/Visual Studio) or LD_LIBRARY_PATH (other compilers/envs), which is necessary as we resolve the libraries that are passed into g-ir-scanner, at least on Windows. -Don't attempt to link to or resolve m.lib on Visual Studio builds, as the math functions are in the standard CRT .lib/.dll, and there is no such thing as m.lib https://bugzilla.gnome.org/show_bug.cgi?id=781525
* giscanner: Update location of temp sources and objectsChun-wei Fan2017-05-012-3/+4
| | | | | | | | | | | | | | For preprocessing, when we create the temp file for preprocessing, make the temp file be stored in CWD instead of the system's temp directory; and when we compile the dumper program, set the output_dir to be the root directory (<drive letter>:\ on Windows and / otherwise). This is because distutils insists on using the full path from the root directory to compile sources, so that if we set the output_dir as we now do we will get $(abs_srcdir)/$(tmpdir)/<$(abs_srcdir)_minus_rootdir>/$(tmpdir)/<target_gir_file_name>.[o|obj]. https://bugzilla.gnome.org/show_bug.cgi?id=781525
* scannerparser.y: Really remove temp .h file on WindowsChun-wei Fan2017-05-011-1/+2
| | | | | | | | | | | | | In gi_source_scanner_parse_macros(), a temporary .h file is generated during the process of parsing the macros, and they aren't being deleted on Windows. In turns out that the g_unlink() call in that function failed because Windows does not allow one to unlink/delete files while they are open, and that the g_unlink() call is done way early (for some reason). Fix this by calling fclose() on the fmacros FILE* *after* we are done with fmacros, and then finally call g_unlink() on the temp .h file. https://bugzilla.gnome.org/show_bug.cgi?id=781525
* Fix 'Bad file descriptor' error when checking libtool versionSam Thursfield2017-04-151-1/+1
| | | | | | | | | | | | Running g-ir-scanner on Fedora 25 (libtool 2.4.6) produces the following message on stderr, twice: /usr/bin/libtool: line 2460: printf: write error: Bad file descriptor This is because we were redirecting stdout to /dev/null incorrectly -- we were opening /dev/null for reading rather than for writing. https://bugzilla.gnome.org/show_bug.cgi?id=781312
* g-ir-scanner: fix wrong order of -L options in libtool invocationSebastian Geiger2017-03-261-2/+4
| | | | | | | This patch fixes an issue in dumper.py, which causes a wrong oder of -L options to be passed to libtool. This in turn can cause build failures if old library files are present in the linked library directories. https://bugzilla.gnome.org/show_bug.cgi?id=778507
* giscanner: Use shlex.split() for environment variablesNirbheek Chauhan2017-03-212-7/+6
| | | | | | | | | | | | | | str.split() does not handle quoting, so if you have spaces in your CFLAGS, it will be split incorrectly. For instance: CFLAGS="'-I/opt/some dir' -DFOO=bar" >>> os.environ['CFLAGS'].split() ["'-I/opt/some", "dir'", '-DFOO=bar'] >>> shlex.split(os.environ['CFLAGS']) ['-I/opt/some dir', '-DFOO=bar'] https://bugzilla.gnome.org/show_bug.cgi?id=778971
* typedefs: fix type information.Mathieu Duponchelle2017-03-142-7/+2
| | | | | | | | | | | We now reuse _create_type_from_base instead of reimplementing it in a semi-broken way in transformer, and use complete_ctype when writing it out if available. This incidentally allows us to remove a test where the comment acknowledged the non-optimal behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779959
* scanner: Add a way to specify extra libraries to link againstThibault Saunier2016-12-103-3/+7
| | | | | | | | | | | | Currently the only way to get a library to link against the dumper program is through '--link' which implies that library will be defined in the 'libraries' field of the .gir. When using libtool, we link against dependencies of that library as the .la defines that, but when using --no-libtool that won't happen and the user needs to be able to define to what other libraries the program needs to be link against, and this is what the new --extra-library argument is about. https://bugzilla.gnome.org/show_bug.cgi?id=774625
* ast: Avoid a crash with anonymous structsScott D Phillips2016-05-081-1/+2
| | | | | | | We enabled anonymous unions earlier, this is a bugfix to enable anonymous structs. https://bugzilla.gnome.org/show_bug.cgi?id=766011
* ccompiler.py: MSVC builds: Fix running on Python 3.xChun-wei Fan2016-03-171-1/+1
| | | | | When resolving libraries, open the temp file generated by dumpbin with 'r' mode rather than 'rb', since this is a text file.
* ccompiler.py: Win32: Use tempfile for resolving libsChun-wei Fan2016-03-161-21/+36
| | | | | | | | | | | | When using the NMake Makefile within the Visual Studio projects, as we are using dumpbin to resolve the .lib file that we link to, the outputs of dumpbin is captured by the Visual Studio output panel, which causes it not to be processed by proc.communicate(). This is due to dumpbin being an integrated component of Visual Studio. In order to remedy this, we need to use a temp file, and use the /out:<tempfile> flag, and look for the DLL that is linked to by the .lib that we pass in. https://bugzilla.gnome.org/show_bug.cgi?id=763739
* Do not segfault on comments with invalid UTF-8Pascal Terjan2016-03-071-0/+6
|
* giscanner/msvccompiler.py: Fix exception handlingChun-wei Fan2016-02-191-1/+1
| | | | | | | | | | Fix the syntax where we handle the exception during preprocessing, so that we won't trigger a NameError when preprocessing fails, and so the error would become clearer to people. Reported by Cosimo Lupo. https://bugzilla.gnome.org/show_bug.cgi?id=757126
* scanner: Assume gcc use the same path separator as pythonNicola Fontana2016-02-161-1/+1
| | | | | | | resolve_windows_libs() uses `gcc -print-search-dirs`. When we are cross compiling with MinGW on linux, gcc uses ':' as path separator, not ';'. https://bugzilla.gnome.org/show_bug.cgi?id=761981
* Allow overriding of dlltool nameNicola Fontana2016-02-161-2/+1
| | | | | | | | Allow to override it with the DLLTOOL environment variable, leaving "dlltool" as a fallback when not defined so backward compatibility is ensured. https://bugzilla.gnome.org/show_bug.cgi?id=761984
* scanner: Remove leading $SHELL in libtool callNicola Fontana2016-02-161-1/+2
| | | | | | | | | | | | Avoid crashes due to double shell in subprocess.Popen() on linux, e.g.: ``` >>> import subprocess >>> subprocess.Popen(['/bin/sh', '/bin/sh']) /bin/sh: /bin/sh: cannot execute binary file ``` https://bugzilla.gnome.org/show_bug.cgi?id=761982
* Use a launcher for executing g-ir-scannerNicola Fontana2016-02-161-0/+6
| | | | | | | | In cross-compilation the build system is unable to execute the compiled binary. Give the possibility to use a launcher, e.g. wine for MinGW or qemu for different CPU type. https://bugzilla.gnome.org/show_bug.cgi?id=696773
* Revert "giscanner: Disable warnings arising from -D_FORTIFY_SOURCE -O0"Philip Withnall2016-02-061-6/+0
| | | | | | | | | This reverts commit 0ff7ca94a608663649defc72021062de098853a8. As reported by Ting-Wei Lan, this breaks builds with Clang, which doesn’t support -Wno-cpp. https://bugzilla.gnome.org/show_bug.cgi?id=757934#c5
* giscanner: Disable warnings arising from -D_FORTIFY_SOURCE -O0Philip Withnall2016-01-301-0/+6
| | | | | | | | | | | | | | | Using distutils.ccompiler means that we are forced to use the CFLAGS from the system’s Python installation, which may contain -D_FORTIFY_SOURCE. The user’s environment-provided CFLAGS may contain -O0 (because they are a developer). These two flags cause a warning when used together. Silence that warning by passing -Wno-cpp to disable warnings from #warning preprocessor statements in the generated C code. It doesn’t seem to be possible to selectively undefine _FORTIFY_SOURCE or to stop using the compiler flags from distutils.sysconfig, unfortunately. https://bugzilla.gnome.org/show_bug.cgi?id=757934
* giscanner: fix tests with python3.5Marc-Antoine Perennou2016-01-181-1/+2
| | | | | | https://bugzilla.gnome.org/show_bug.cgi?id=760682 Signed-off-by: Marc-Antoine Perennou <Marc-Antoine@Perennou.com>
* MingGW Builds: Don't load msvcrt for MinGW64Alexey Pavlov2016-01-131-1/+1
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=733535
* Win32: Fix giscannermodule.c for Python 3.5+Chun-wei Fan2016-01-081-7/+14
| | | | | | | | | | | | Python 3.5+ is now built with Visual Studio 2015, which now has a refactored set of CRT DLLs where CRT functions are placed in, so we need to acquire the _get_osfhandle() from the correct CRT DLL so that we can pass around file descriptors correctly between different CRT versions. This patch will enable Windows builds of introspection files using Python 3.5+ to work correctly. https://bugzilla.gnome.org/show_bug.cgi?id=759531
* scanner: fix line number countingDieter Verfaillie2015-12-271-13/+13
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=752047
* tests: Add more regression tests for (not nullable)Philip Withnall2015-11-061-0/+1
| | | | | | | | | | | | On bug #719966, the question arose of whether parameters annotated with (element-type) but not (not nullable) are regarded as nullable or non-nullable. Add some new unit tests to Regress-1.0.gir to check the behaviour is as expected: annotating a parameter with (element-type) implicitly makes it non-nullable (unless also annotated with (nullable)). https://bugzilla.gnome.org/show_bug.cgi?id=757678
* g-ir-scanner: Set sensible permissions on the output fileMikhail Zabaluev2015-11-011-0/+8
| | | | | | | | If the file is new, set the permissions to 0644. If the file with the name specified as --output exists, copy its metadata onto the temporary output file. https://bugzilla.gnome.org/show_bug.cgi?id=757442
* Don't lose output silently when overwrite of the output file is deniedMikhail Zabaluev2015-11-011-1/+0
| | | | | | | If the move resulted in EPERM, discard the temporary output file, but raise the error so that the tool exits with a nonzero status. https://bugzilla.gnome.org/show_bug.cgi?id=757442
* giscanner/transformer.py: Fix --identifier-filter on Python 3.xChun-wei Fan2015-10-281-1/+2
| | | | | | | We need to encode the identifier-filter command to bytes, and decode its outputs to ascii in order for it to work on Python 3.x. https://bugzilla.gnome.org/show_bug.cgi?id=757126
* giscanner/dumper.py: Fix for Python 3.x for MSVC 2010+Chun-wei Fan2015-10-281-1/+2
| | | | | | | | sys.exc_clear(), which was called when we need to ignore the error when we can't embed manifests to the dumper program on MSVC 2010 and later, is only available/needed for Python 2.7.x, so only use it there. https://bugzilla.gnome.org/show_bug.cgi?id=757126
* giscanner/ccompiler.py: Make Windows builds work with Python 3.xChun-wei Fan2015-10-281-2/+2
| | | | | | | | | | Windows builds check the library that was passed into the introspection scanner to deduce the correct DLL that is to be used by the .gir/.typelib file, but this was not updated for Python 2.x/3.x compatibility, as the outputs of a subprocess must be decoded. Fix this for both MSVC and MinGW-based builds. https://bugzilla.gnome.org/show_bug.cgi?id=757126
* giscanner/msvccompiler.py: Make it compatible with Python 3.xChun-wei Fan2015-10-281-2/+2
| | | | | | This is checked for Python 2.7.x compatibility as well. https://bugzilla.gnome.org/show_bug.cgi?id=757126
* scanner: Fix non-libtool linker flags on DarwinPhilip Chimento2015-10-201-3/+4
| | | | | | | | | Darwin's linker doesn't like "-rpath=path"; instead pass "-rpath path", which works on more linkers than the version with the = sign does. Additionally, Darwin's linker has no equivalent for "--no-as-needed" (it seems to do the right thing by default?) https://bugzilla.gnome.org/show_bug.cgi?id=625195
* Add support for function typedefs without pointerBen Iofel2015-10-191-2/+10
| | | | | | | | | | g-ir-scanner now supports something like this: typedef void my_callback(int); Notice how my_callback is not a pointer. https://bugzilla.gnome.org/show_bug.cgi?id=755645
* giscanner: Fix python2/3 compatibilityDamien Grassart2015-10-193-8/+8
| | | | | | | This allows building in both Python 2 and 3 by fixing a few text/binary ambiguities and using "as" in an except clause. https://bugzilla.gnome.org/show_bug.cgi?id=756763
* transformer: Strip out GLib _autoptr typedefsColin Walters2015-10-101-0/+3
| | | | | | | They're just internal bits for users of C, shouldn't be part of the API. https://bugzilla.gnome.org/show_bug.cgi?id=755882
* maintransformer: Fix regression in callback closure assignmentColin Walters2015-10-101-5/+6
| | | | | | | The nullable code needs to search via index lookup of closure_name, which drops the need to (incorrectly) assign closure_name again. https://bugzilla.gnome.org/show_bug.cgi?id=756352
* giscanner.message: Don't print node as the default contextMikhail Zabaluev2015-10-051-3/+1
| | | | | | | Node is always present in the output of .log_node(), so printing it another time as context just adds noise. https://bugzilla.gnome.org/show_bug.cgi?id=756088
* giscanner: Mark gpointer nodes as nullable by defaultPhilip Withnall2015-10-045-13/+62
| | | | | | | | | | | | | gpointer parameters and return types should be marked as nullable by default, unless: • also annotated with (type) and not with (nullable); or • explicitly annotated with (not nullable). This introduces the (not nullable) annotation as a direct opposite to (nullable). In future, (not) could be extended to invert other annotations. https://bugzilla.gnome.org/show_bug.cgi?id=729660
* giscanner: Mark (closure) parameters as (nullable) by conventionPhilip Withnall2015-10-041-0/+7
| | | | | | | | All user_data pointers should be nullable, and they should all be annotated as closures too. I have not found any counter-examples where a closure is non-nullable. https://bugzilla.gnome.org/show_bug.cgi?id=729660
* scanner: Warn and ignore return annotations when there is no return valueGarrett Regier2015-10-042-5/+11
| | | | | | | | Otherwise the .gir can contain invalid data which would likely cause a crash at runtime. https://bugzilla.gnome.org/show_bug.cgi?id=752044 Signed-off-by: Garrett Regier <garrett.regier@riftio.com>
* scanner: Warn and ignore on incorrect optional/nullable/allow-none annotationsGarrett Regier2015-10-041-3/+18
| | | | | | | These can easily be misunderstood, especially optional. https://bugzilla.gnome.org/show_bug.cgi?id=752065 Signed-off-by: Garrett Regier <garrett.regier@riftio.com>
* giscanner.ast: Map size_t and ssize_t to gsize and gssize, respectivelyMikhail Zabaluev2015-10-041-2/+1
| | | | | | | The mapping to long int types seems bogus; the code above it did the right thing, but only for size_t. https://bugzilla.gnome.org/show_bug.cgi?id=756016
* giscanner.message: Fix module-level logging functionsMikhail Zabaluev2015-09-301-3/+7
| | | | | | | | MessageLogger.get() calls the class constructor without arguments. The __init__ signature, however, did not default the namespace parameter, so any usage of the logging functions caused a bogus exception. https://bugzilla.gnome.org/show_bug.cgi?id=755890
* giscanner: Add a Parameter.name propertyPhilip Withnall2015-09-291-0/+4
| | | | | | | This is an alias of Parameter.argname, which makes it easier to duck-type debugging of AST nodes by printing out their name property. https://bugzilla.gnome.org/show_bug.cgi?id=747979
* giscanner: Store direction in TypeContainer instancePhilip Withnall2015-09-291-5/+5
| | | | | | | | Instead of storing a direction property on both Parameter and Return separately, hoist it up to TypeContainer so it’s inherited. This neatens things up a bit, but doesn’t really change anything in practice. https://bugzilla.gnome.org/show_bug.cgi?id=747979
* giscanner: Store parent link in AST for TypeContainer instancesPhilip Withnall2015-09-292-0/+35
| | | | | | | | | | It is useful for navigating the AST to have a link from each TypeContainer instance (e.g. each Parameter and Return) to the parent AST node, such as a Function. This will be used in the g-ir-diff tool. https://bugzilla.gnome.org/show_bug.cgi?id=747979
* scanner: Fix allow-none on return valuesGarrett Regier2015-09-291-1/+2
| | | | | | | | The optional annotation was being applied which is invalid for return values. https://bugzilla.gnome.org/show_bug.cgi?id=752029 Signed-off-by: Garrett Regier <garrett.regier@riftio.com>
* scanner: Pacify PEP8Colin Walters2015-09-294-5/+7
|
View Lorry Controller Status