It's now possible to mark struct fields and function params as "references",
which causes them to become pointers in C (in bindings, they become whatever
is necessary). They're not a part of the type and are much more restricted
than pointers, allowing bindings to be easier. This system will be gradually
utilized and expanded as required.
@feature
Now when dealing with pointer types, we will not get pointer to
pointer semantics in callbacks and eina_promise_owner_value_set
for Eina_Promise.
It will work as expected:
Eina_Promise_Owner* promise = eina_promise_add();
void* p = malloc(sizeof(T));
eina_promise_owner_value_set(promise, p, &free);
Complex types (i.e. list, array, hash, accessor etc.) now do not require
pointers with them anymore (the pointer is implied) and the same goes for
class handles. Eolian now explicitly disallows creating pointers to these
as well. This is the first part of the work to remove pointers from Eolian
completely, with the goal of simplifying the DSL (higher level) and therefore
making it easier for bindings (as well as easier API usage).
@feature
Modify the way hooks are defined and used by promise generation in
Eolian in the Eo API.
Instead of passing macro names as parameters to EO_FUNC_BODY macros,
just re-define the actual hooks when it is needed.
Add a promise object to allows Eolian interface to include promises
as a way to have asynchronous value return and composibility.
The usage is like this in a .eo file:
class Foo {
methods {
bar {
params {
@inout promise: Promise<int>;
}
}
}
}
Which will create the following API interface:
void foo_bar(Eo* obj, Eina_Promise** promise);
and a Eina_Promise_Owner for the implementation, like this:
void _foo_bar(Eo* obj, Private_Data* pdata, Eina_Promise_Owner* promise);
Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
Add two parameters for macros that generate API functions in Eo so
that the generation can be customized with macros used by Eolian.
Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
Reverting this at Felipe's request following my email. There are many
things I strongly object to in this commit. I've touched the surface of
those on the ML (which doesn't work at the moment), though we need to
better discuss it.
The gist:
1. dlsym is a really bad hack that is not even needed.
2. I don't see why eo should even be aware of promises. It's not aware
of list, hash and etc.
3. The eolian changes were done wrong.
This should have been discussed and consulted before done, even if only
because of the amount of hacks it includes and the cross-domain (ecore,
eo and eolian) nature of it.
This reverts commit f9ba80ab33.
Add a promise object that allows Eolian interface to include promises
as a way to have asynchronous value return and composibility.
The usage is like this in a .eo file:
class Foo {
methods {
bar {
params {
promise: Promise<int>;
}
}
}
}
Which will create the following API interface:
void foo_bar(Ecore_Promise** promise);
and the equivalent declaration for implementation.
However, the API function will instantiate the Promise for the
user and the implementer of the class.
This allows generators to silence type errors in validation in order
to reduce duplicate error messages when generating multiple files.
Also adjusted the C generator to only emit type errors when generating
Eo header files.
@feature
Until now it wasn't allowed/possible to init (eo_init) eo after it has
been shut down (eo_shutdown). This commit fixes that, so now that is
fully legal to have as many init/shutdown cycles as you want.
There was a previous workaround for this issue:
e47edc250d.
This should allow more flexibility when using the EFL in loadable
modules and in various other scenarios.
The problem is that the class_get() functions cache the previously
created class for efficiency, but the class is freed if eo is shut down,
so the cached pointer is actually invalid.
The solution to the problem was to maintain a generation count
(incremented every time we shut down eo), and compare that to a locally
saved version in class_get(). If they don't match, recreate the class,
as it has already been freed.
@feature
As described by Carsten in his email to edev ML titled:
"[E-devel] eo stability - i think we need to postpone that"
with the switch to Eo2 we significantly increased our usage of RW memory
pages, and thus significantly increased our memory usage when running
multiple applications.
The problem was that during the migration to Eo2 the op id cache and the
op description arrays were merged, causing the op description arrays to
no longer be RO. This patch enables users of Eo (mainly Eolian) to
declare those arrays as const (RO) again, saving that memory.
There might be performance implications with this patch. I had to remove
the op desc array sorting, and I used a hash table for the lookup. I
think the op desc sorting doesn't really affect performance because that
array is seldom accessed and is usually pretty short. The hash table
is not a problem either, because it's behind the scenes, so it can be
changed to a more efficient data structure if the hash table is not good
enough. The hash table itself is also rarely accessed, so it's mostly
about memory.
Please keep an eye for any bugs, performance or excessive memory usage.
I believe this should be better on all fronts.
This commit *BREAKS ABI*.
@fix
This hasn't been used for a while. Since we are going to break Eo a bit anyway
it's a good opportunity to drop this.
This may cause a slight performance issues with legacy events, such as
smart callbacks. This shouldn't really be a problem as we've migrated away from
them. If it does, we need to migrate the remaining parts. Only relevant
for callbacks that are added before the classes are created, which
shouldn't be possible except for smart, only for old evas callbacks.
It's now generalized to work properly with legacy functions
as well as correctly append suffixes. Thus, it can be easily
used with the documentation generator and in other places.
@feature
Otherwise there would be conflicts in certain circumstances.
This also requires adding const on many existing functions,
and similar work is necessary in Elementary.
@fix
This resulted in a field using the next field's docs rather
than its own, because it incremented the member before
even trying to retrieve the docs.
@fix
While it is true that 'strrchr' may return NULL, in this case it
never will (there is always a newline). Either way, silence Coverity.
Fixes CID 1304558.
This commit adds the necessary generator logic to emit doc
comments from the new doc syntax. Old doc comments are kept
in for the time being as they're used within the EFL but they
will be removed eventually. This new generator focuses all the
important code in one place, making usage easy.
@feature
Typedef word was not generated for enums. Camel case was not respected
for type definition.
Tests have been updated to check enums generation.
@fix T2400
Local and base class functions are supported.
When @empty is provided, dummy functions (initializing the parameters with default
values if needed) are generated.
When @auto is provided on properties, access to internal data variables is done. On
set, it will assign parameters values to private data members. On get,
parameters are set with private data members values.
See the supplied tests as examples.
@feature
As the implements list will soon contain all methods and properties,
do some preparations. The Eolian library now fills in class field in
implements early on when the implement is local. The Eolian C generator
now checks for local implements and skips them (so that things don't break).
This is needed when get properties or methods have to return a
value in case of failure or to initialize parameters.
The way used is to generate an intermediate function that will
initialize the parameters and then invoke the "user" function.
Because of Eina_Value being less than optimal in our usage, we're dropping it in Eolian.
This simplifies the code and makes it easier to bind to other languages, which
will aid new generators. Also, we're dropping long double support from eo files
and expressions as it causes an ABI breakage in gcc 4.4.
This change adds a new API to retrieve the legacy prefix for enum members
and also adds support for generating enums into the C generator. Besides that,
it prevents generation of entire struct contents when the .eo file declares
a named struct inside of a typedef - as the struct itself gets generated later on.
This commit also does several side (related) changes. Particularly, it updates
the Eolian C generator to use the new API, it adds missing expr types (null, char)
and masks, updates the API dealing with default return values to use expressions
instead of strings and does several fixes (mainly in lexer) around the place.
It also disallows single quoted strings as those are reserved for characters.
so the test suites were useless because they didn't ecompile what was
generated. it was missing a ) at the end of calling eo_do_super to
start with. it was missing a call to do_uper fo destructors too. also
it didnt deal with the #include "blah.eo.c" at the file end either so
it produced uncompilable src at the start. again - also fix. this also
fixes test case comparisons to match the now correct generation that
compiles.
This new API supports function pointer types, multiple type subtypes, const attribute without parsing
the name string, own attribute for any partial type and more. This commit also updates the C and C++
generators so that they compile and generate correct code.
@feature
.eot files are similar to .eo, but they can't contain classes; all .eot files found in include paths
are parsed and information is added into the database. They're there for typedefs and eventually also
enums, structs and constants.
@feature
This patch:
- removes the @def from Doxygen, as it is not correct for API
functions.
- fixes the generation of class comments. When no class description is
supplied, no comment should be added.
place.
Instead of having the calculation (class name + function name + set/get)
in many places in the code, it is now in one place and accessible via a
function environment structure.
In this case, the section 'implements' contains bad information about the
function to override. If the first (at least) function is correct, it
will never fail but use the last correct information retrieved from the
database.
The patch fixes it by checking the result of the database function
eolian_implement_information_get.
@fix
The function eolian_implement_information_get was returning strings for
the class and the function. It was written in this way at the beginning
because it was not needed to verify the correctness of the class and
the function.
Now that we have the namespace feature, this function must check it,
meaning that the class and the function are now known.
So we can return them instead of returning the strings.
The generators had to find the class from the classname. It is no more
needed.
The C++ generator has been adapted to this new API.
To reduce the invocations to strings conversions, we store the
classname, the Eo prefix in upper and lower cases in global variables.
The problem comes when we have to handle overriding functions. A lot of
conflicts between class base and class inheriting can happen.
The chosen solution is to create independent environments storing the
converted strings.
When using -O2 or -O3, the Eina_Bool legacy_support (unsigned char) was
overriding the int eo_needed.
The result was a failure during options check:
Eo flag is not specified (use --eo). Aborting eo generation.
@fix
Until now, the functions giving access to class information were taking
the class name as parameter.
Except the fact that we needed to search into a hash table for the internal
class structure, no flexibility is possible.
This change consists in modifying most of the APIs using the class name
with a new Eolian_Class type and adapt the code of the C and C++
generators accordingly.
By using -gi option, the generator appends the functions that are
present into the given eo file and missing into the developer file
(given via -o option as an in/out file).
@feature