mesa/src/mapi/glapi/gen/gl_apitemp.py

335 lines
10 KiB
Python
Raw Normal View History

Mammoth update to the Python code generator scripts that live in src/mesa/glapi. Basically, the scripts that did simple things (like gl_offsets.py) were simple, and the scripts that did more complicated things (like glX_proto_send.py) were getting progressively more and more out of control. So, I re-write the foundation classes on which everything is based. One problem with the existing code is that the division between the GL API database representation and the way the output code is generated was either blury or nonexistant. The new code somewhat follows the Model-View-Controller pattern, minus the Controller. There is a distinct set of classes that model the API data, and there is a distinct set of classes that generate code from that data. One big change is in the class that represents GL functions (was glFunction, is now gl_function). There used to be an instance of this calls for each function and for each alias to that function. For example, there was an instance for PointParameterivSGIS, PointParameterivEXT, PointParameterivARB, and PointParameteriv. In the new code, there is one instance. Each instance has a list of entrypoint names for the function. In the next revision, this will allow a couple useful things. The script will be able to verify that the parameters, return type, and GLX protocol for a function and all it's aliases match. It will also allow aliases to be represented in the XML more compactly. Instead of repeating all the information, an alias can be listed as: <function name="PointParameterivARB" alias="PointParameterivEXT"/> Because the data representation was changed, the order that the alias functions are processed by the scripts also changed. This accounts for at least 2,700 of the ~3,600 lines of diffs in the generated code. Most of the remaining ~900 lines of diffs are the result of bugs *fixed* by the new scripts. The old scripts also generated code with some bugs in it. These bugs were discovered while the new code was being written. These changes were discussed on the mesa3d-dev mailing list back at the end of May: http://marc.theaimsgroup.com/?t=111714569000004&r=1&w=2 Xorg bug: 3197, 3208
2005-06-22 00:42:43 +01:00
# (C) Copyright IBM Corporation 2004, 2005
# All Rights Reserved.
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# on the rights to use, copy, modify, merge, publish, distribute, sub
# license, and/or sell copies of the Software, and to permit persons to whom
# the Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
# IBM AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
# IN THE SOFTWARE.
#
# Authors:
# Ian Romanick <idr@us.ibm.com>
import argparse
import gl_XML, glX_XML
import license
Mammoth update to the Python code generator scripts that live in src/mesa/glapi. Basically, the scripts that did simple things (like gl_offsets.py) were simple, and the scripts that did more complicated things (like glX_proto_send.py) were getting progressively more and more out of control. So, I re-write the foundation classes on which everything is based. One problem with the existing code is that the division between the GL API database representation and the way the output code is generated was either blury or nonexistant. The new code somewhat follows the Model-View-Controller pattern, minus the Controller. There is a distinct set of classes that model the API data, and there is a distinct set of classes that generate code from that data. One big change is in the class that represents GL functions (was glFunction, is now gl_function). There used to be an instance of this calls for each function and for each alias to that function. For example, there was an instance for PointParameterivSGIS, PointParameterivEXT, PointParameterivARB, and PointParameteriv. In the new code, there is one instance. Each instance has a list of entrypoint names for the function. In the next revision, this will allow a couple useful things. The script will be able to verify that the parameters, return type, and GLX protocol for a function and all it's aliases match. It will also allow aliases to be represented in the XML more compactly. Instead of repeating all the information, an alias can be listed as: <function name="PointParameterivARB" alias="PointParameterivEXT"/> Because the data representation was changed, the order that the alias functions are processed by the scripts also changed. This accounts for at least 2,700 of the ~3,600 lines of diffs in the generated code. Most of the remaining ~900 lines of diffs are the result of bugs *fixed* by the new scripts. The old scripts also generated code with some bugs in it. These bugs were discovered while the new code was being written. These changes were discussed on the mesa3d-dev mailing list back at the end of May: http://marc.theaimsgroup.com/?t=111714569000004&r=1&w=2 Xorg bug: 3197, 3208
2005-06-22 00:42:43 +01:00
class PrintGlOffsets(gl_XML.gl_print_base):
def __init__(self, es=False):
gl_XML.gl_print_base.__init__(self)
Mammoth update to the Python code generator scripts that live in src/mesa/glapi. Basically, the scripts that did simple things (like gl_offsets.py) were simple, and the scripts that did more complicated things (like glX_proto_send.py) were getting progressively more and more out of control. So, I re-write the foundation classes on which everything is based. One problem with the existing code is that the division between the GL API database representation and the way the output code is generated was either blury or nonexistant. The new code somewhat follows the Model-View-Controller pattern, minus the Controller. There is a distinct set of classes that model the API data, and there is a distinct set of classes that generate code from that data. One big change is in the class that represents GL functions (was glFunction, is now gl_function). There used to be an instance of this calls for each function and for each alias to that function. For example, there was an instance for PointParameterivSGIS, PointParameterivEXT, PointParameterivARB, and PointParameteriv. In the new code, there is one instance. Each instance has a list of entrypoint names for the function. In the next revision, this will allow a couple useful things. The script will be able to verify that the parameters, return type, and GLX protocol for a function and all it's aliases match. It will also allow aliases to be represented in the XML more compactly. Instead of repeating all the information, an alias can be listed as: <function name="PointParameterivARB" alias="PointParameterivEXT"/> Because the data representation was changed, the order that the alias functions are processed by the scripts also changed. This accounts for at least 2,700 of the ~3,600 lines of diffs in the generated code. Most of the remaining ~900 lines of diffs are the result of bugs *fixed* by the new scripts. The old scripts also generated code with some bugs in it. These bugs were discovered while the new code was being written. These changes were discussed on the mesa3d-dev mailing list back at the end of May: http://marc.theaimsgroup.com/?t=111714569000004&r=1&w=2 Xorg bug: 3197, 3208
2005-06-22 00:42:43 +01:00
self.name = "gl_apitemp.py (from Mesa)"
self.license = license.bsd_license_template % ( \
"""Copyright (C) 1999-2001 Brian Paul All Rights Reserved.
(C) Copyright IBM Corporation 2004""", "BRIAN PAUL, IBM")
self.es = es
self.undef_list.append( "KEYWORD1" )
self.undef_list.append( "KEYWORD1_ALT" )
self.undef_list.append( "KEYWORD2" )
self.undef_list.append( "NAME" )
self.undef_list.append( "DISPATCH" )
self.undef_list.append( "RETURN_DISPATCH" )
self.undef_list.append( "DISPATCH_TABLE_NAME" )
self.undef_list.append( "UNUSED_TABLE_NAME" )
self.undef_list.append( "TABLE_ENTRY" )
def printFunction(self, f, name):
p_string = ""
o_string = ""
t_string = ""
comma = ""
if f.is_static_entry_point(name):
keyword = "KEYWORD1"
else:
keyword = "KEYWORD1_ALT"
n = f.static_name(name)
silence = ''
space = ''
for p in f.parameterIterator(name):
if p.is_padding:
continue
if p.is_pointer():
cast = "(const void *) "
else:
cast = ""
t_string = t_string + comma + p.format_string()
p_string = p_string + comma + p.name
o_string = o_string + comma + cast + p.name
comma = ", "
silence += "%s(void) %s;" % (space, p.name);
space = ' '
if f.return_type != 'void':
dispatch = "RETURN_DISPATCH"
else:
dispatch = "DISPATCH"
need_proto = False
if not f.is_static_entry_point(name):
need_proto = True
elif self.es:
cat, num = api.get_category_for_name(name)
if (cat.startswith("es") or cat.startswith("GL_OES")):
need_proto = True
if need_proto:
print('%s %s KEYWORD2 NAME(%s)(%s);' % (keyword, f.return_type, n, f.get_parameter_string(name)))
print('')
print('%s %s KEYWORD2 NAME(%s)(%s)' % (keyword, f.return_type, n, f.get_parameter_string(name)))
print('{')
if silence:
print(' %s' % (silence))
if p_string == "":
print(' %s(%s, (), (F, "gl%s();\\n"));' \
% (dispatch, f.name, name))
else:
print(' %s(%s, (%s), (F, "gl%s(%s);\\n", %s));' \
% (dispatch, f.name, p_string, name, t_string, o_string))
print('}')
print('')
return
def printRealHeader(self):
print('')
self.printVisibility( "HIDDEN", "hidden" )
print("""
/*
* This file is a template which generates the OpenGL API entry point
* functions. It should be included by a .c file which first defines
* the following macros:
* KEYWORD1 - usually nothing, but might be __declspec(dllexport) on Win32
* KEYWORD2 - usually nothing, but might be __stdcall on Win32
* NAME(n) - builds the final function name (usually add "gl" prefix)
* DISPATCH(func, args, msg) - code to do dispatch of named function.
* msg is a printf-style debug message.
* RETURN_DISPATCH(func, args, msg) - code to do dispatch with a return value
*
* Here is an example which generates the usual OpenGL functions:
* #define KEYWORD1
* #define KEYWORD2
* #define NAME(func) gl##func
* #define DISPATCH(func, args, msg) \\
* struct _glapi_table *dispatch = CurrentClientDispatch; \\
* (*dispatch->func) args
* #define RETURN DISPATCH(func, args, msg) \\
* struct _glapi_table *dispatch = CurrentClientDispatch; \\
* return (*dispatch->func) args
*
*/
#if defined( NAME )
#ifndef KEYWORD1
#define KEYWORD1
#endif
#ifndef KEYWORD1_ALT
#define KEYWORD1_ALT HIDDEN
#endif
#ifndef KEYWORD2
#define KEYWORD2
#endif
#ifndef DISPATCH
#error DISPATCH must be defined
#endif
#ifndef RETURN_DISPATCH
#error RETURN_DISPATCH must be defined
#endif
#ifdef MemoryBarrier
#undef MemoryBarrier
#endif
""")
return
def printInitDispatch(self, api):
print("""
#endif /* defined( NAME ) */
/*
* This is how a dispatch table can be initialized with all the functions
* we generated above.
*/
#ifdef DISPATCH_TABLE_NAME
#ifndef TABLE_ENTRY
#error TABLE_ENTRY must be defined
#endif
#ifdef _GLAPI_SKIP_NORMAL_ENTRY_POINTS
#error _GLAPI_SKIP_NORMAL_ENTRY_POINTS must not be defined
#endif
_glapi_proc DISPATCH_TABLE_NAME[] = {""")
for f in api.functionIterateByOffset():
print(' TABLE_ENTRY(%s),' % (f.dispatch_name()))
print(' /* A whole bunch of no-op functions. These might be called')
print(' * when someone tries to call a dynamically-registered')
print(' * extension function without a current rendering context.')
print(' */')
for i in range(1, 100):
print(' TABLE_ENTRY(Unused),')
print('};')
print('#endif /* DISPATCH_TABLE_NAME */')
print('')
return
Mammoth update to the Python code generator scripts that live in src/mesa/glapi. Basically, the scripts that did simple things (like gl_offsets.py) were simple, and the scripts that did more complicated things (like glX_proto_send.py) were getting progressively more and more out of control. So, I re-write the foundation classes on which everything is based. One problem with the existing code is that the division between the GL API database representation and the way the output code is generated was either blury or nonexistant. The new code somewhat follows the Model-View-Controller pattern, minus the Controller. There is a distinct set of classes that model the API data, and there is a distinct set of classes that generate code from that data. One big change is in the class that represents GL functions (was glFunction, is now gl_function). There used to be an instance of this calls for each function and for each alias to that function. For example, there was an instance for PointParameterivSGIS, PointParameterivEXT, PointParameterivARB, and PointParameteriv. In the new code, there is one instance. Each instance has a list of entrypoint names for the function. In the next revision, this will allow a couple useful things. The script will be able to verify that the parameters, return type, and GLX protocol for a function and all it's aliases match. It will also allow aliases to be represented in the XML more compactly. Instead of repeating all the information, an alias can be listed as: <function name="PointParameterivARB" alias="PointParameterivEXT"/> Because the data representation was changed, the order that the alias functions are processed by the scripts also changed. This accounts for at least 2,700 of the ~3,600 lines of diffs in the generated code. Most of the remaining ~900 lines of diffs are the result of bugs *fixed* by the new scripts. The old scripts also generated code with some bugs in it. These bugs were discovered while the new code was being written. These changes were discussed on the mesa3d-dev mailing list back at the end of May: http://marc.theaimsgroup.com/?t=111714569000004&r=1&w=2 Xorg bug: 3197, 3208
2005-06-22 00:42:43 +01:00
def printAliasedTable(self, api):
print("""
/*
* This is just used to silence compiler warnings.
* We list the functions which are not otherwise used.
*/
#ifdef UNUSED_TABLE_NAME
_glapi_proc UNUSED_TABLE_NAME[] = {""")
normal_entries = []
proto_entries = []
for f in api.functionIterateByOffset():
normal_ents, proto_ents = self.classifyEntryPoints(f)
# exclude f.name
if f.name in normal_ents:
normal_ents.remove(f.name)
elif f.name in proto_ents:
proto_ents.remove(f.name)
normal_ents = [f.static_name(ent) for ent in normal_ents]
proto_ents = [f.static_name(ent) for ent in proto_ents]
normal_entries.extend(normal_ents)
proto_entries.extend(proto_ents)
print('#ifndef _GLAPI_SKIP_NORMAL_ENTRY_POINTS')
for ent in normal_entries:
print(' TABLE_ENTRY(%s),' % (ent))
print('#endif /* _GLAPI_SKIP_NORMAL_ENTRY_POINTS */')
print('#ifndef _GLAPI_SKIP_PROTO_ENTRY_POINTS')
for ent in proto_entries:
print(' TABLE_ENTRY(%s),' % (ent))
print('#endif /* _GLAPI_SKIP_PROTO_ENTRY_POINTS */')
print('};')
print('#endif /*UNUSED_TABLE_NAME*/')
print('')
return
def classifyEntryPoints(self, func):
normal_names = []
normal_stubs = []
proto_names = []
proto_stubs = []
# classify the entry points
for name in func.entry_points:
if func.has_different_protocol(name):
if func.is_static_entry_point(name):
proto_names.append(name)
else:
proto_stubs.append(name)
else:
if func.is_static_entry_point(name):
normal_names.append(name)
else:
normal_stubs.append(name)
# there can be at most one stub for a function
if normal_stubs:
normal_names.append(normal_stubs[0])
elif proto_stubs:
proto_names.append(proto_stubs[0])
return (normal_names, proto_names)
def printBody(self, api):
normal_entry_points = []
proto_entry_points = []
for func in api.functionIterateByOffset():
normal_ents, proto_ents = self.classifyEntryPoints(func)
normal_entry_points.append((func, normal_ents))
proto_entry_points.append((func, proto_ents))
print('#ifndef _GLAPI_SKIP_NORMAL_ENTRY_POINTS')
print('')
for func, ents in normal_entry_points:
for ent in ents:
self.printFunction(func, ent)
print('')
print('#endif /* _GLAPI_SKIP_NORMAL_ENTRY_POINTS */')
print('')
print('/* these entry points might require different protocols */')
print('#ifndef _GLAPI_SKIP_PROTO_ENTRY_POINTS')
print('')
for func, ents in proto_entry_points:
for ent in ents:
self.printFunction(func, ent)
print('')
print('#endif /* _GLAPI_SKIP_PROTO_ENTRY_POINTS */')
print('')
self.printInitDispatch(api)
self.printAliasedTable(api)
return
Mammoth update to the Python code generator scripts that live in src/mesa/glapi. Basically, the scripts that did simple things (like gl_offsets.py) were simple, and the scripts that did more complicated things (like glX_proto_send.py) were getting progressively more and more out of control. So, I re-write the foundation classes on which everything is based. One problem with the existing code is that the division between the GL API database representation and the way the output code is generated was either blury or nonexistant. The new code somewhat follows the Model-View-Controller pattern, minus the Controller. There is a distinct set of classes that model the API data, and there is a distinct set of classes that generate code from that data. One big change is in the class that represents GL functions (was glFunction, is now gl_function). There used to be an instance of this calls for each function and for each alias to that function. For example, there was an instance for PointParameterivSGIS, PointParameterivEXT, PointParameterivARB, and PointParameteriv. In the new code, there is one instance. Each instance has a list of entrypoint names for the function. In the next revision, this will allow a couple useful things. The script will be able to verify that the parameters, return type, and GLX protocol for a function and all it's aliases match. It will also allow aliases to be represented in the XML more compactly. Instead of repeating all the information, an alias can be listed as: <function name="PointParameterivARB" alias="PointParameterivEXT"/> Because the data representation was changed, the order that the alias functions are processed by the scripts also changed. This accounts for at least 2,700 of the ~3,600 lines of diffs in the generated code. Most of the remaining ~900 lines of diffs are the result of bugs *fixed* by the new scripts. The old scripts also generated code with some bugs in it. These bugs were discovered while the new code was being written. These changes were discussed on the mesa3d-dev mailing list back at the end of May: http://marc.theaimsgroup.com/?t=111714569000004&r=1&w=2 Xorg bug: 3197, 3208
2005-06-22 00:42:43 +01:00
def _parser():
"""Parser arguments and return a namespace."""
parser = argparse.ArgumentParser()
parser.add_argument('-f',
metavar='<input file name>',
dest='filename',
default="gl_API.xml",
help="An XML file describing the API.")
parser.add_argument('-c',
action='store_true',
dest='es',
help="Enable OpenGL ES compatibility")
return parser.parse_args()
def main():
"""Main function."""
args = _parser()
api = gl_XML.parse_GL_API(args.filename, glX_XML.glx_item_factory())
printer = PrintGlOffsets(args.es)
printer.Print(api)
if __name__ == '__main__':
main()