For both GLib and GIO. (GObject, GIRepository and GModule don’t have any
platform specific APIs.)
This is needed for two reasons:
* If the same GIR file is shipped on multiple platforms, it has no way
to conditionally define/not-define an API based on the platform (like
an `#ifdef` in a C header). So we either end up shipping differing
GIR APIs on different platforms, or shipping a GIR file which
declares APIs which aren’t resolvable by `dlopen()` on certain
platforms, and will cause a language runtime error.
* The API reference documentation is now generated from the GIR, and
similar problems are present there: if the GIR contains different
symbols depending on the platform, there is no way to generate API
documentation for the union of all of them.
The fix is to ensure that there are no conditional symbols in a GIR, by
splitting out the platform specific symbols into platform specific GIR
files. Platform specific documentation can then be generated from these,
in addition to the main, platform agnostic, documentation.
The documentation changes will following in a subsequent commit.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3037
There was no obvious logical need to list the `GAppInfo` subclass
sources separately in the build. It makes more sense to add them to the
platform-specific source lists, since they are platform specific.
This will be used in an upcoming commit which generates
platform-specific GIR files, so needs the full platform-specific lists
of sources.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3037
It was causing a build failure on some platforms:
```
In file included from ../girepository/gthash.c:29:
In file included from ../girepository/gitypelib-internal.h:30:
../girepository/girepository-private.h:26:10: fatal error: 'ffi.h' file not found
#include <ffi.h>
^~~~~~~
```
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Instead, add a method on `GIRegisteredTypeInfo` which indicates whether
the registered type is a boxed type. This will return true for
`GIStructInfo` and `GIUnionInfo` instances which are boxed (not all
structs and unions are).
This makes `GIBoxedInfo` redundant, and it’ll be dropped in a following
commit.
---
There are several different things which typelibs need to be able to
represent:
1. Plain old datatype (POD) structs
2. POD unions
3. Structs with a copy func and/or free func
4. Unions with a copy func and/or free func
5. Structs which are the ‘GType struct’ for an object or interface (i.e.
the class or instance or interface struct)
6. Structs with a copy func and free func *and* boxed GType
7. Unions with a copy func and free func *and* boxed GType
8. Boxed GTypes which represent something other than a struct or union
So there’s a lot going on here. In commit
e28078c70cbf4a57c7dbd39626f43f9bd2674145, a lot of this was reworked,
and support was added for boxed unions and boxed ‘others’ (the last item
on the list above).
Since then, support for boxed types other than structs seems to have
atrophied a bit, and the whole lot has got a bit confusing.
It was perhaps less confusing when all the `GIBaseInfo` subclasses were
actually aliases of each other, but now they have subtype relationships,
the position of `GIBoxedInfo` in that type hierarchy has become unclear.
How is it related to `GIStructInfo`, `GIUnionInfo` and
`GIRegisteredTypeInfo`?
Since a boxed type is necessarily a `GIRegisteredTypeInfo`, and the
methods of `GIRegisteredTypeInfo` are all written to allow a `GType` to
be optional, so that `GIStructInfo` and `GIUnionInfo` can safely derive
from it and still be used to represent plain old datatypes without
`GType`s, it makes sense to add a method to `GIRegisteredTypeInfo` to
indicate that the registered type is derived from `G_TYPE_BOXED`.
Accordingly, the things above are now represented in libgirepository’s
type system as:
1. `GIStructInfo` instance, `GIRegisteredTypeInfo` methods return no
`GType` info
2. `GIUnionInfo` instance similarly
3. `GIStructInfo` instance, `GIRegisteredTypeInfo` methods return no
`GType` info, `gi_struct_info_get_{copy,free}_function_name()` return
non-`NULL` values
4. `GIUnionInfo` instance similarly
5. `GIStructInfo` instance, `GIRegisteredTypeInfo` methods return no
`GType` info, `gi_struct_info_is_gtype_struct()` returns true
6. `GIStructInfo` instance, `GIRegisteredTypeInfo` methods return valid
`GType` information, `gi_registered_type_info_is_boxed()` returns
true, `gi_struct_info_get_{copy,free}_function_name()` return
`NULL` values (because the copy/free functions are hidden inside the
boxed type registration at runtime)
7. `GIUnionInfo` instance similarly
8. Not representable, but could be represented in future by re-adding a
`GIBoxedInfo` type which derives from `GIRegisteredTypeInfo` and is
used solely for boxed ‘other’ types, *not* boxed structs or unions
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Fixes: #3245
When creating `GIBaseInfo` instances from binary blobs, so far the code
has used an undocumented implicit conversion from one enum type to the
other. That’s a bit nasty, as it means the two enum types need to be
kept in sync (without that need being documented).
Introduce an explicit conversion function to make things more explicit.
Currently it does nothing, but in an upcoming commit it will be used to
deprecate a blob type.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3245
While it’s an internal API which `girwriter.c` has access to, it’s not
available inside `libgirepository-internals.so`. This wasn’t spotted
before commit 343027d5d landed because none of the existing users of
`libgirepository-internals.so` use the relevant code in `girwriter.c`,
so it got compiled out (`libgirepository-internals.so` is statically
linked and can be optimised like this).
Now that `gi-decompile-repository` uses the relevant code from
`girwriter.c`, the problem is obvious and `gi-decompile-repository`
fails to link.
Fix that by no longer using `gi_base_info_get_info_type()` in
`girwriter.c`. These are changes which would have eventually been made
anyway in issue #3253.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3253
`--version` is conventionally used to find the version of the program
itself. While that’s not currently implemented in
`gi-inspect-repository`, let’s not box ourselves into a corner.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Update the code to the latest standards and to use girepository-2.0
after it’s been imported to this repository.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3155
Update the code to the latest standards and to use girepository-2.0
after it’s been imported to this repository.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3155
Rename them to `gi-decompile-repository` and `gi-inspect-repository` to
match the existing `gi-compile-repository`. The names have to differ
from those used in girepository-1.0 to avoid collisions.
‘generate’ has been changed to ‘decompile’ because it does the inverse
of what `gi-compile-repository` does: it converts a typelib to a GIR
file. ‘generate’ never really made much sense for this anyway — it’s
almost a synonym of ‘compile’.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3155
GLib is used a subproject with GObject Introspection and overriding
this binary breaks that usage. For now we will make overriding the
binary conditional until Meson is updated to handle the binary changes.
In the project from which this copy of cmph was forked, this was fixed
in version 2.0.2 by commit
bbf77c63c9/
Signed-off-by: Simon McVittie <smcv@collabora.com>
These can then be used by a parent project if GLib is used as a Meson
subproject.
In particular, GJS wants to use `GLib-2.0.typelib`
from GLib-as-a-subproject, but doesn’t want to hardcode the path to it
inside the GLib builddir.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Fixes: #3244
Various distributions (mainly RPM based so far) make use of automatic
dependencies extracted from typelib files (they can require other typelibs
and also shared libraries)
Current features
* Print used shared libraries
* Print used typelib dependencies
Based-on-patch-by: Dominique Leuenberger <dimstar@opensuse.org>
Reference: https://bugzilla.gnome.org/show_bug.cgi?id=665672
Reviewed-by: Colin Walters <walters@verbum.org>
Signed-off-by: Dominique Leuenberger <dimstar@opensuse.org>
Signed-off-by: Igor Gnatenko <ignatenko@redhat.com>
Take a GError * for typelib loading code, validate the header. This
fixes bizarre errors from gjs where g_irepository_require would happily
load old typelibs.
People have wanted support for marking (out) on functions of the
form:
/**
* clutter_color_from_pixel:
* @pixel: A pixel
* @color: (out): Color to initialize with value of @pixel
*/
void
clutter_color_from_pixel (guint32 pixel, ClutterColor *color);
Where the caller is supposed to have allocated the argument; the
C function just initializes it. This patch adds support for this
argument passing style to introspection. In this case, we see the
(out), and notice that there's only a single indirection (*) on
the argument, and assume that this means (out caller-allocates).
https://bugzilla.gnome.org/show_bug.cgi?id=604749
Foreign structs are special in the sense that there might
be native bindings (for instance PyCairo for PyGI) that provides
the same functionallity as the introspected variant.
https://bugzilla.gnome.org/show_bug.cgi?id=610357
Parse the c:prefix from the .gir, include it in the header. Armed with this
information, we can now optimize lookups of GTypes because we
have the requirement that GTypes must start with the c:prefix. We do
fall back though if a lookup fails.
Broadly speaking, this change adds the concept of <vfunc> to the .gir.
The typelib already had most of the infrastructure for virtual functions,
though there is one API addition.
The scanner assumes that any class callback slot that doesn't match
a signal name is a virtual. In the .gir, we write out *both* the <method>
wrapper and a <vfunc>. If we can determine an association between
them (based on the names matching, or a new Virtual: annotation),
then we notate that in the .gir.
The typelib gains an association from the vfunc to the function, if
it exists. This will be useful for bindings since they already know
how to consume FunctionInfo.
We now support an extensible mechanism where arbitrary key-value
pairs may be associated with almost all items, including objects,
methods, and properties.
These attributes appear in both the .gir and the .typelib.
Add --all option, which is intended to show some information not
usually included in the GIR.
Currently, it shows the size of structs and unions.
Signed-off-by: Andreas Rottmann <a.rottmann@gmx.at>
Similar to GObject class structs, we pair up GInterfaces with
their C structures.
Also, move some GLib-specific things into glibast.py, and make
the naming more generic.
Inside glibtransformer, we now look at structures ending in "Class" and
see if they have an associated GlibObject (i.e. a structure of the same
name without the "Class" suffix). If found, pair them up.
The .gir file for <class> gains an attribute denoting its associated
class struct. Any <record> many now have a glib:is-class-struct-for
annotation which tells which (if any) <class> for which it defines the
layout.
In the .typelib, we record the association between the class and
its structure. Generic structures however just have a boolean
saying whether they're a class struct. (Going from a generic class
struct to its class should not be necessary).
Finally, we expose GIRepository APIs to access both bits of information
from the .typelib.
svn path=/trunk/; revision=1088
2008-01-03 Andreas Rottmann <a.rottmann@gmx.at>
Bug 556489 – callback annotations
* giscanner/transformer.py
* tools/generate.c (write_callable_info): Write out the new scope,
closure and destroy attributes.
* giscanner/transformer.py (Transformer._type_is_callback): New
method, checking if a given type is a callback.
(Transformer._augment_callback_params): New method; adds
information (closure, destroy) to callback parameters.
(Transformer._handle_closure, Transformer._handle_destroy): New methods,
auxiliary to _augment_callback_params.
(Transformer._create_function): Call _augment_callback_params().
(Transformer._create_parameter): Handle scope option.
(Transformer._create_typedef_callback): New method, creates a
callback, and registers it in the typedef namespace
(Transformer._create_typedef): Use _create_typedef_callback()
instead of the plain _create_callback().
* giscanner/ast.py (Parameter): Added callback-related fields.
* giscanner/girwriter.py: Write out new Parameter fields.
* girepository/girnode.h (GIrNodeParam): Added fields scope,
closure and destroy.
* girepository/gtypelib.h (ArgBlob): Ditto.
* girepository/girparser.c (start_parameter): Handle new fields.
* girepository/girmodule.c (g_ir_module_build_typelib): Adjust
arg_blob_size, bump major version due to this change.
* girepository/girnode.c (g_ir_node_get_full_size_internal)
(g_ir_node_build_typelib)
* girepository/gtypelib.c (g_typelib_check_sanity): ArgBlob size
adjustments.
(g_ir_node_build_typelib): Fill in new ArgBlob flags from param.
* girepository/girepository.h (GIScope): New enumeration, listing
the different possible scopes for callbacks.
* girepository/ginfo.c (g_arg_info_get_scope)
(g_arg_info_get_closure, g_arg_info_get_destroy): Accessors for
callback-related argument indices (callback scope, closure for a
callback, destroy notification for a callback).
* tests/scanner/: Added testcases for new features.
svn path=/trunk/; revision=998
Field offsets are a) architecture dependent so they shouldn't be part
of the architecture-independent gir format which is installed in
datadir. b) Are architecture-dependent so they shouldn't be in test
expected output. Remove field offsets from girs.
(Virtual function and discriminator offsets are not removed, as they
aren't fully hooked up to the field-offset computation machinery yet.)
svn path=/trunk/; revision=877
2008-10-17 Tommi Komulainen <tommi.komulainen@iki.fi>
Bug 556732 – generate gir files consistently
* tools/generate.c (write_callable_info): write 'direction'
attribute only if other than 'in' to be consistent with
girwriter.py
* tests/boxed.gir:
* tests/object.gir: remove direction="in" which is the default
svn path=/trunk/; revision=744
2008-10-17 Tommi Komulainen <tommi.komulainen@iki.fi>
Bug 556732 – generate gir files consistently
* giscanner/ast.py (Field): add readable and writable properties
* giscanner/girparser.py (_parse_field): copy 'readable' and
'writable' attributes
* giscanner/transformer.py (_create_member): create fields as
read-write
* giscanner/glibtransformer.py (_introspect_object,
_pair_class_struct): make object instance and class fields
read-only
* giscanner/girwriter.py (_write_field):
* tools/generate.c (write_field_info): write field 'readable'
and 'writable' attributes only if non-default (read-only)
* girepository/girparser.c (start_field): in the absence of
attributes assume fields are read-only
* tests/boxed.gir:
* tests/struct.gir: remove redundant readable="1" from fields
* tests/scanner/foo-1.0-expected.gir:
* tests/scanner/utility-1.0-expected.gir: add writable="1" to
all record and union fields
svn path=/trunk/; revision=743
2008-10-15 Tommi Komulainen <tommi.komulainen@iki.fi>
* tools/generate.c (xml_printf): quote printf arguments so that
we don't generate invalid XML by writing unescaped double quotes
and such in attributes
svn path=/trunk/; revision=715
2008-10-12 Tommi Komulainen <tommi.komulainen@iki.fi>
* tools/generate.c (write_callable_info): move "transfer" and
"null-ok" attributes from <function> to <return-value> element
svn path=/trunk/; revision=691
This is a big patch. You should probably remove your installation
tree to be cleaner.
* docs/typelib-format.txt: Add nsversion entry which holds
version of namespace.
* girepository/girepository.h: Add 'version' parameter to
g_irepository_require. This may be NULL. Normally
bindings should pass an explicit version though.
* girepository/girepository.c: Lots of infrastructure to
support versioning. Add some more documentation. Disallow
some usage of NULL namespaces.
* girepository/girmodule.c: Add version parameter.
* girepository/gtypelib.c: Update header size.
* giscanner/ast.py: Add version to Namespace.
* giscanner/girparser.py: Parse version attribute from
XML, pass to Namespace.
* giscanner/girwriter.py: Write out version parameter.
* giscanner/transformer.py: Clean up include registration.
* tests/*: Add version attribute.
* tests/invoke/invoke.c: Don't try looking up test before
it's loaded in repository.
* tools/generate.c: Output version parameter.
* gir/Makefile.am: Add 2.0 version to .gir files.
svn path=/trunk/; revision=677
* girepository/girnode.c: Allow gtype_name
and gtype_init in struct and union.
* girepository/girparser.c: Parse glib:
boxed bits for both structure and union.
* girepository/gtypelib.c: Don't barf
if structure is boxed.
* giscanner/girparser.py: Parse new XML
format.
* giscanner/girwriter.py: Write out new
XML format.
* giscanner/glibast.py: Define new classes
which are both Boxed and Struct/Union, as
well as an "Other" for everything else.
* giscanner/glibtransformer.py: Handle
boxed types specially; we try to merge
them with a struct/union if one exists,
otherwise fall back to generic boxed.
* tests/*: Update.
* tools/generate.c: Write out new format.
svn path=/trunk/; revision=575
* gir/Makefile.am: Dep on Makefile
* girepository/ginfo.c: Print out a nicer error
message if we failed to load something.
* girepository/girepository.c: Clean up
default typelib handling; remove global
default_typelib variable. Ensure we handle
NULL repository in more places.
Support dependency resolution.
* tests/Makefile.am: Kill off gobject.gir,
it conflicts with the real one.
* tests/Object.gir: Depend on GObject.
* tools/generate.c: Take --includedir
argument to say which directories to search
for typelibs. Print out dependencies.
svn path=/trunk/; revision=541
2008-08-27 Tor Lillqvist <tml@novell.com>
Make g-ir-scanner work on Windows. Still problems with the typelib
code. Changes okayed by jdahlin.
* configure.ac: Check for Windows, set Automake conditional
OS_WIN32. Change backslashes to forward slashes in pyexecdir to
avoid shell quoting issues
* girepository/Makefile.am: Use -no-undefined so that libtool
agrees to build a shared library on Windows.
* girepository/girparser.c (backtrace_stderr): No backtrace() on
Windows. Empty implementation on Windows so far.
* girepository/gtypelib.c (g_typelib_check_sanity): Give more
informative error message for the assertion failures. Tell also
what the expected size of the struct is. Check all sizes first and
fail afterwards if at least one size was different from expected.
* tools/Makefile.am: Reorder libraries into proper logical
dependency order.
* tools/generate.c: Don't include <dlfcn.h>, not used.
* giscanner/Makefile.am: On Windows, link with the Python library,
and install the module DLL as _giscanner.pyd. Remove the
unnecessary import library and libtool library that libtool has
installed.
* giscanner/scannerlexer.l: Recognize the gcc __attribute__ syntax
and just skip it. Recognize also two "l" suffixes for long long
constants. Recognize also __inline__.
* giscanner/grealpath.h (g_realpath): Implement on Windows, using
GetFullPathName(). As such, GetFullPathName() does more than the
UNIX realpath(). It also changes relative paths into absolute
paths. But for our purposes that shouldn't matter.
* giscanner/giscannermodule.c (pygi_source_scanner_parse_file): On
Windows the file descriptor passed to us is from Python. Python
Python2.5 uses the msvcr71.dll C library, while mingw-built code
uses msvcrt.dll. On Windows, file descriptors are specific to
which C library is used. So we must find out what underlying OS
handle corresponds to the file descriptor Python passes us, and
then make that into a file descriptor valid for the C library this
code uses.
* giscanner/sourcescanner.py (_parse): Don't need to bypass
__attribute__ as the lexer now handles it. The definition as empty
was ineffective for mingw anyway, as mingw's _mingw.h undefines
__attribute__. Close the temp file before unlinking it.
* giscanner/cgobject.py: Use correct library name for the gobject
DLL on Windows.
* gir/Makefile.am: Must pass the full basename of the DLLs on
Windows to g-ir-scanner. It's a bit ugly that we have to "know"
that the names of the GLib DLLs are like libglib-2.0-0.dll, but in
reality they won't change, until there is a GLib 3, and then also
the Unix code here needs changing.
Must pass CPPFLAGS to g-ir-scanner when building GLib.gir so that
libintl.h is found.
svn path=/trunk/; revision=503
2008-08-22 Colin Walters <walters@verbum.org>
* girepository/girparser.c: Pass through
recursive types. Avoid overwriting errors.
* giscanner/xmlwriter.py: Always write the
XML header.
* tests/*.gir: Adjust.
* tests/scanner/Makefile.am: Build typelibs,
and generate XML from those. Once we
have a good diff mechanism...
* tests/scanner/*-expected.gir: Add XML
header.
* tools/g-ir-scanner: Accept --typelib-xml
option.
* tools/generate.c: Better defaults for transfer.
svn path=/trunk/; revision=457
2008-08-22 Colin Walters <walters@verbum.org>
* gobject-introspection-1.0.pc.in: Add g-ir-generate.
* tests/Makefile.am: Support $(DEBUG)
* tools/generate.c: Do immediate close tags if no
sub-elements.
svn path=/trunk/; revision=448
