register required classes:
+ add lazy references to classes referred to in the current unit
+ define reference symbols for classes defined in the current unit
+ add lazy references to parent classes of classes defined in the current
unit
git-svn-id: branches/objc@13679 -
based on patch by Dmitry Boyarintsev (mantis #14508)
o Todo: while parsing the class declaration, check whether the
field types are valid for use in an obj-c class
* use a common pool for selector names generated by objcselector()
and by the rtti info for implemented classes
git-svn-id: branches/objc@13663 -
confuse the assembler (patch by Dmitry Boyarintsev, from mantis
#14508)
* use objextname and not objrealname in the method name (extname is
the class/protocol/category name as it appears/will appear to the
ObjC runtime)
git-svn-id: branches/objc@13655 -
libtool from Xcode 3.2 (Mac OS X 10.6) does not support all new linker
options (in particular -no_order_inits and -no_order_data, both of
which must be specified when compiling dynamic libraries with FPC and
Xcode 3.2)
git-svn-id: trunk@13645 -
* don't allow "type x = type y" for objcclass/objcprotocol for now
(would require generating new rtti for the new name)
git-svn-id: branches/objc@13641 -
after merge of r13618 from trunk
* also allow "varargs" for objclass/objcprotocol methods
* set the cdecl calling convention for objcclass/objcprotocol methods in
handle_calling_convention() rather than in a separate routine in
pdecobj (and do the same for cppclass methods)
* check that array of const parameters and varargs are only used with
external objc-methods (can only be done after the class is parsed,
because earlier on it's not known yet whether or not it is external)
git-svn-id: branches/objc@13624 -
from the def)
+ added (disabled) code to correctly calculate the parameter offsets for
ppc, see comments why it's not activated
git-svn-id: branches/objc@13619 -
routines (just like "varargs"), because if implemented in Pascal then
on the callee side this array of const parameter is treated as a Pascal-
style array of const
* don't give the "cdecl'ared functions have no high parameter" warning for
array of const parameters for cdecl external routines and procvars
git-svn-id: trunk@13618 -
Darwin platforms, so that no data can be placed below that address.
This fixes the strange Windows-compatible resource API, which
assumes that addresses <64kb do not exist.
git-svn-id: trunk@13600 -
* left old cpu-specific set helper code under ifdef FPC_OLD_BIGENDIAN_SETS
in case someone wants to write new assembler set helpers (although most
of them should be optimally generated by the compiler already if
http://wiki.freepascal.org/FPC_HowToDo#Bit.28field.29_getting.2Fsetting_primitives
are optimally implemented)
git-svn-id: trunk@13582 -
change, resulting in slightly different DWARF debug information
(mantis #13508)
The cause was saving a node pointer in a local variable, processing
it further, and later on checking whether it changed by comparing
the stored and the current instance pointer. The problem was that
the node could have been freed and reallocated at the same address
(but with different contents), so this check sometimes resulted
in (hard to reproduce) false negatives.
git-svn-id: trunk@13580 -
byte is used for zero-termination (so we now always read blocks of
exactly 32kb from disk)
* set the first byte of the buffer to #0 when opening a file initially
git-svn-id: trunk@13576 -
(and use the resulting compiler as final compiler to compile the rest of
the source tree and to install)
Note: this means that if you manually use "make cycle RELEASE=1", the RTL
afterwards will not be usable, because it will have been compiled with
WPO for the compiler; if you use "make all" in the root FPC dir, there
is no problem since that makefile recompiles the RTL anyway.
git-svn-id: trunk@13560 -
type definitions (was at the position of the invalid x type in
"class of x", rather than at the position of the "class of x" definition
itself)
git-svn-id: trunk@13555 -
explicitly using -Ooasmcse for now. It will probably be completely
removed before the next major release after 2.4.0 (fixes mantis
#14363)
git-svn-id: trunk@13545 -
changes from that commit, because not hardcoding the use of
current_procinfo in all cases when obtaining the GOT register is
cleaner
git-svn-id: trunk@13470 -
associated with an Objective-C protocol. Code generator part will
always generate an internalerror currently, because it cannot yet
be implemented (needs support for generating RTTI for Objective-C
classes)
git-svn-id: branches/objc@13461 -
only the last source file specified will be compiled, not the first one
* also print out both the old and the new source file name to be compiled
when showing that message (to more easily find where the error is)
* give an error if an unknown parameter is used, instead of silently
ignoring it
git-svn-id: trunk@13429 -
Darwin anymore), because otherwise if unit A is compiled without debug
info and unit B uses one of its types, the debug info will contain no
DW_TAG_typedef for that type and hence gdb will not recognise it as
a type definition that's part of the program.
git-svn-id: trunk@13424 -
depend on that (and correct a number of cases where this was wrong)
* set the correct subreg type for xmm x86_64 parameter registers
(resolved mantis #14067)
git-svn-id: trunk@13410 -
starting with a previous 2.3.1 or compiler built from the objc branch
+ added basic objcprotocol support (only for external protocols
currently)
o use in type declaration: "type xp = objcprotocol ... end;"
o when defining a root class that implements it:
"type yc = objcclass(xp) ... end" (note: no support yet
for something like "objcclass(id,xp)" or so)
o when defining a non-root class that implements a protocol:
"type zc = objcclass(nsobject,xp) ... end"
o includes support for "required" and "optional" sections
o no support yet for the objcprotocol(<protocol>) expression
that enables getting a class instance representing the
protocol (e.g., for use with "conformsToProtocol:")
o message names have to specified in protocol declarations,
but if an objcclass implements a protocol, the message names do
not have to be repeated (but if they are, they have to match;
the same goes when overriding inherited methods)
+ allow specifying the external name of Objective-C classes and
protocols, since classes and protocols can have the same name
(and you cannot use the same Pascal identifier in such caseq)
+ added NSObject protocol, and make the NSObject class use it
+ added missing NSObject class methods that have the same name
as instance methods (added "class" name prefix to avoid clashes)
* fixed several cases where the compiler did not treat Objective-C
classes/protocols the same as Object Pascal classes/interfaces
(a.o., forward declarations, alignment, regvars, several type
conversions, ...)
* allow "override" directive in objcclass declarations, and print
a hint if it's forgotten in an external declaration (because it
doesn't really matter there, and may make automated header
conversion harder than necessary) and an error if will be used in
a non-external declaration (because it is not possible to start
a new vmt entry-tree in Objective-C, you can only override parent
methods)
* reject objcclasses/protocols as parameters to typeof()
* don't try to test VMT validity of objcclasses/protocols
git-svn-id: branches/objc@13375 -
svn+ssh://jonas@svn.freepascal.org/FPC/svn/fpc/trunk
........
........
r13157 | marco | 2009-05-17 14:22:41 +0200 (Sun, 17 May 2009) | 2 lines
* mingw textmode IDE debugging support. First working version (for me), atexit handler for mingw and cygwin stuff under ifndef USE_MINGW_GDB
........
r13158 | marco | 2009-05-17 14:43:57 +0200 (Sun, 17 May 2009) | 1 line
* Fix for gdb66/mingw (general?) program reset issue. Instructs GDB to turn confirmation off. Mantis 0011968
........
r13159 | yury | 2009-05-17 14:43:58 +0200 (Sun, 17 May 2009) | 1 line
* Replaced some ifdef x86_64 by ifdef cpu64bitaddr, since it is more correct.
........
r13160 | yury | 2009-05-17 14:47:26 +0200 (Sun, 17 May 2009) | 1 line
* Set default image base for win64 dlls to $110000000.
........
........
........
r13163 | jonas | 2009-05-17 16:33:20 +0200 (Sun, 17 May 2009) | 3 lines
* support {$i *.ext} to include source files with the same base name, but
with a different extension than the current source file (mantis #13696)
........
r13164 | jonas | 2009-05-17 16:42:17 +0200 (Sun, 17 May 2009) | 2 lines
* copyright in comments updated (from mantis #13048)
........
r13165 | jonas | 2009-05-17 16:50:44 +0200 (Sun, 17 May 2009) | 2 lines
* small fixes from Karl-Michael Schindler (mantis #13048)
........
r13166 | marco | 2009-05-17 17:37:27 +0200 (Sun, 17 May 2009) | 3 lines
* gtk_cell_renderer_combo_new returns gtkcellrenderer not -combo (mantis 13689),
probably to support some kind of procedural polymorphism.
........
r13167 | jonas | 2009-05-17 18:01:13 +0200 (Sun, 17 May 2009) | 4 lines
* when loading the functions for a particular OpenGL function, also
automatically load those for lower versions (patch by "satan", mantis
#13687)
........
r13168 | marco | 2009-05-17 19:09:26 +0200 (Sun, 17 May 2009) | 3 lines
* reversed destruction order of TFPObjectList as per Mantis 13715.
Note that it is _NOT_ good coding to rely on such things
........
........
........
........
r13172 | sergei | 2009-05-18 00:56:51 +0200 (Mon, 18 May 2009) | 9 lines
DOM test suite enhancements:
* Do not convert tests which request implementation attribute 'signed'='true'.
Such tests aren't applicable to our unsigned DOM, they only cause compiler warnings
and noise in the test report.
+ Support for default properties (obj.item(x) -> obj[x]).
+ Support black-listing of testcases. Some of them (in HTML testsuite) are easier to
rewrite by hand than to convert.
+ Support adding certain units to 'uses' clause (e.g. HTML suite must use dom_html).
........
........
........
r13175 | sergei | 2009-05-18 22:33:43 +0200 (Mon, 18 May 2009) | 1 line
dom.pp: Ignore node read-only property when its owning document is being destroyed (enables destroying nodes in arbitrary order).
........
r13176 | sergei | 2009-05-19 00:03:37 +0200 (Tue, 19 May 2009) | 2 lines
* Moved TDOMAttrDef class from xmlread.pp to dom.pp (in order to implement default attribute handling at DOM level).
* Using regular CloneNode() method to create copies of TDOMAttrDef.
........
........
........
........
r13180 | sergei | 2009-05-21 01:08:24 +0200 (Thu, 21 May 2009) | 13 lines
Final strike for #13605:
src/dom.pp:
* GetElementsByTagName[NS] results now get cached in a hashtable. Repeated calls to
GetElementsByTagName with same arguments return the same instance of NodeList. All NodeLists
created during document lifetime are destroyed with the document.
src/xmlutils.pp:
* THashTable.Lookup(), changed SetString to SetLength+Move because SetString truncates on #0
+ added THashTable.RemoveData() method
tests/api.xml:
- No longer need to 'garbage collect' the NodeLists.
........
r13181 | jonas | 2009-05-21 11:57:45 +0200 (Thu, 21 May 2009) | 4 lines
- unconditionally enabled {$ifdef PARAOUTFILE}-code (that define was
decomissioned quite a while ago, but this file was forgotten in
that process)
........
r13182 | marco | 2009-05-21 13:26:54 +0200 (Thu, 21 May 2009) | 2 lines
* win64 compatibility fix from Vincent. TLibHandle is now equal to system.thandle
........
r13183 | marco | 2009-05-21 16:11:16 +0200 (Thu, 21 May 2009) | 1 line
* some minor vista/w7 related updates to message numbers
........
r13184 | sergei | 2009-05-21 22:45:16 +0200 (Thu, 21 May 2009) | 1 line
dom.pp: Added node memory management code (pure addition, no functionality changes this time).
........
r13185 | sergei | 2009-05-22 01:07:32 +0200 (Fri, 22 May 2009) | 10 lines
dom.pp: Customized memory management, fixes 17 test cases and all memory leaks in testsuite:
* Every node created by Document.CreateXXX method is now guaranteed to be destroyed with the
document, whether it is part of the tree or not. Therefore, DOM methods which remove nodes
from the tree (namely, TDOMNode.RemoveChild, TDOMNode.ReplaceChild,
TDOMElement.SetAttributeNode and TDOMElement.SetAttributeNodeNS) no longer need to destroy
their return value and are now conformant to the specs.
* Nodes are allocated in arrays of instances (emulates 'placement new operator' in C++ terms).
Allocation and freeing are as fast as possible (just assigns a couple of pointers).
* Behaviour of nodes that are created by direct call to constructor is unchanged.
........
r13186 | jonas | 2009-05-23 12:10:56 +0200 (Sat, 23 May 2009) | 2 lines
+ Darwin-specific fcntl constants
........
........
r13188 | marco | 2009-05-23 17:40:13 +0200 (Sat, 23 May 2009) | 1 line
* thread creation failure for win<x>. Mantis 13798
........
r13189 | joost | 2009-05-24 13:17:59 +0200 (Sun, 24 May 2009) | 1 line
* Modified patch from taka_jp to set the connection-dataset for mysql, bug #13588.
........
r13190 | joost | 2009-05-24 13:20:45 +0200 (Sun, 24 May 2009) | 1 line
* Fixed resource-leak on exceptions during creating/dropping databases, patch from taka_jp, bug #13587
........
r13191 | jonas | 2009-05-24 13:48:49 +0200 (Sun, 24 May 2009) | 5 lines
* changed the parameters of gotoxy() and the return values of wherex/wherey
from byte into tcrtcoord=1..255, since all crt unit coordinates are
1-based (not for Window() procedure, see comments in crth.inc;
mantis #13788)
........
r13192 | jonas | 2009-05-24 14:25:33 +0200 (Sun, 24 May 2009) | 2 lines
* again made less complex in case of PIC, otherwise it fails
........
r13193 | jonas | 2009-05-24 14:26:38 +0200 (Sun, 24 May 2009) | 2 lines
* make sure example also works without an application bundle
........
r13194 | jonas | 2009-05-25 20:34:04 +0200 (Mon, 25 May 2009) | 2 lines
* fixed evaluating "nil +/- int_const" expressions (mantis #13815)
........
r13195 | jonas | 2009-05-25 21:41:14 +0200 (Mon, 25 May 2009) | 4 lines
* fixed missing reference counting in case a refcounted type was implicitly
type casted to a non-refcounted type, e.g. dynarray to pointer in Delphi
mode (mantis #13820)
........
r13196 | sergei | 2009-05-25 22:28:40 +0200 (Mon, 25 May 2009) | 8 lines
dom.pp: No more failures on level 1 testsuite... at last.
+ Implemented handling of default attributes:
* creating an element also creates and attaches the default attributes;
* removing an attribute restores it with default value, if there's one.
+ Attribute nodes remove themselves from the owner element upon destruction, making it possible
to Free attributes manually.
* TDOMNamedNodeMap.SetNamedItem does not reset attribute OwnerElement if the argument node is already
contained in the map (and whole operation is therefore is a no-op).
........
r13197 | marco | 2009-05-26 14:24:58 +0200 (Tue, 26 May 2009) | 5 lines
* uxtheme moved from Lazarus.
* Disabled (ifdeffed) lock around init/finalization because of circular
reference between fcl-base and winunits-base. If needed it can be
recoded directly using system specific (windows.pas) routines.
........
r13198 | jonas | 2009-05-26 18:22:41 +0200 (Tue, 26 May 2009) | 2 lines
* make references PIC-safe before pushing them as parameter (mantis #13829)
........
r13199 | marco | 2009-05-26 18:54:49 +0200 (Tue, 26 May 2009) | 2 lines
* tmschema as include file added at Paul's request
........
r13200 | sergei | 2009-05-27 00:14:21 +0200 (Wed, 27 May 2009) | 6 lines
+ Implementation of TDOMDocument.CreateAttributeNS and CreateElementsNS
+ Initial part of TDOMElement.SetAttributeNS
+ TDOMImplementation.CreateDocumentType checks validity of QualifiedName
* Use 'managed' memory allocation in TDOMEntity.CloneNode and TDOMNotation.CloneNode to avoid leaks
* TDOMDocument.RemoveID now using THashTable.RemoveData(), simplifies things
........
r13201 | marco | 2009-05-27 09:29:04 +0200 (Wed, 27 May 2009) | 2 lines
* Marquee patch from Paul + some Vista PBM_ constants.
........
r13202 | marco | 2009-05-27 10:46:27 +0200 (Wed, 27 May 2009) | 2 lines
* first half of vista updates.
........
r13203 | marco | 2009-05-27 18:26:32 +0200 (Wed, 27 May 2009) | 2 lines
* vista items till WC_STATICA, roughly at 80% now.
........
r13204 | jonas | 2009-05-27 19:26:50 +0200 (Wed, 27 May 2009) | 2 lines
* various optimizations by Dave Richards (mantis #12440)
........
r13205 | jonas | 2009-05-27 19:50:55 +0200 (Wed, 27 May 2009) | 3 lines
* forbid "nil +/- ordinal" expressions (Delphi-compatible, and the
expression is meaningless since nil is not a valid address)
........
r13206 | marco | 2009-05-27 22:08:56 +0200 (Wed, 27 May 2009) | 2 lines
* more Vista fixes, and implemented some macro's that somehow were not implemented.
........
r13207 | hajny | 2009-05-27 23:18:35 +0200 (Wed, 27 May 2009) | 1 line
* extension of r13191 to OS/2, EMX and potential other platforms using crt.inc based implementation of unit crt
........
r13208 | florian | 2009-05-29 00:28:58 +0200 (Fri, 29 May 2009) | 1 line
* patch to fix fillpoly, by borsa77
........
r13209 | michael | 2009-05-29 09:24:02 +0200 (Fri, 29 May 2009) | 1 line
* Change so that when FindFirst initially returns nonzero, FindClose is no longer needed (Windows compatible)
........
r13210 | michael | 2009-05-29 09:27:41 +0200 (Fri, 29 May 2009) | 1 line
* Change so that when FindFirst initially returns nonzero, FindClose is no longer needed (Windows compatible)
........
r13211 | michael | 2009-05-29 09:30:46 +0200 (Fri, 29 May 2009) | 1 line
* Fix from Ken Wright for broken FPC changes
........
r13212 | jonas | 2009-05-30 13:29:54 +0200 (Sat, 30 May 2009) | 2 lines
* fixed/cleaned up tests for "nil + 1"
........
r13213 | michael | 2009-05-30 14:07:16 +0200 (Sat, 30 May 2009) | 1 line
* Some TCustomVariant methods implemented
........
r13214 | sergei | 2009-05-31 10:18:06 +0200 (Sun, 31 May 2009) | 3 lines
+ Processing of prefixed attributes and prefix bindings. This completes namespace support at
the XML reader side.
........
r13215 | sergei | 2009-05-31 11:14:37 +0200 (Sun, 31 May 2009) | 1 line
* Added .txt extension to README_DOM (this one was missed while fixing Mantis #12358)
........
r13216 | sergei | 2009-05-31 12:38:28 +0200 (Sun, 31 May 2009) | 1 line
Bringing readme's up to date
........
r13217 | Legolas | 2009-05-31 14:15:24 +0200 (Sun, 31 May 2009) | 4 lines
* updated nds/gba linker scripts and reverted some changes for 2.2.4a release
- Removed unused/outdated stuff from libndsfpc
+ Added new examples for libndsfpc
+ Added working (I hope so...) makefile.fpc for all libndsfpc/libgbafpc examples
........
git-svn-id: branches/objc@13348 -
can create ambiguities for the parser in case the field names also exist
as modifiers (TP- and Delphi-compatible, mantis #13971) + tests
* fixed tests that broke because of this change
git-svn-id: trunk@13334 -
* reordered fields of tsetting to avoid unaligned fields
* fixed compilation of arm after disableddircache was added
* increased ppu version
git-svn-id: trunk@13327 -
parameters (mantis #13956)
* when reporting an error about too few specified parameters to a call,
return the column of the call itself rather than after the last parameter
(because this ends up after the end of an indexed property in case of
properties, which is confusing)
git-svn-id: trunk@13326 -
case the getter is a function, the result is that first the getter is
called, and subsequently the temp holding the function result is
overwritten (thus not changing anything). This is Delphi-compatible,
and fixes tests/tbf/tb0214*
git-svn-id: trunk@13320 -
pointer is of a reference-counted type (reverts r2165, fixes mantis #14019)
* modified test for mantis #4086 (which was "fixed" by r2165) so it still
works and no longer depends on whether or not the compiler uses a
refcounted temp to evaluate certain expressions
git-svn-id: trunk@13310 -
TDebugInfoDwarf, so DWARFv2 and DWARFv3 generators can share protected
methods.
* fixed the debug info for sets whose language-level lower bound is not
equal to the actual set base used when storing the data (e.g., a "set
of 1..5" is actually stored as a "set of 0..7" or "set of 0..31")
(mantis #13984)
+ interactive test for the above
git-svn-id: trunk@13302 -
* Use DW_AT_byte_stride when possible, because gdb does not support DW_AT_stride_size
* Add stride to dynamic-array definitions
git-svn-id: trunk@13280 -
another, it's not enough that the temp sizes are the same, because
the assignment may only apply to part of them. In such cases,
perform a regular copy (mantis #13948)
git-svn-id: trunk@13255 -
vmt entries of virtual methods that can never be called with references
to FPC_ABSTRACTERROR. Some virtual methods are always considered to be
reachable: published methods, and methods used as getter/setter for a
published property.
git-svn-id: trunk@13238 -
instantiated (for wpo)
* also replace vmt-entries for classes for which we don't have any
information at all with FPC_ABSTRACTERROR (since that means they
certainly are not instantiated), except for their published and
virtual class methods
* fixed check for published methods in wpo
git-svn-id: trunk@13219 -
- Removed unused/outdated stuff from libndsfpc
+ Added new examples for libndsfpc
+ Added working (I hope so...) makefile.fpc for all libndsfpc/libgbafpc examples
git-svn-id: trunk@13217 -
designated parent to NSObject, because you can have multiple root classes
(e.g., NSProxy is also an Objective-C root class in Cocoa). Instead,
give a hint when an Objective-C class is declared without any parent
(because it can very well be a mistake, given that Pascal classes all
do automatically inherit form TObject).
git-svn-id: branches/objc@13177 -
o support for declaring external Objective-C classes (see
rtl/inc/objcbase.pas), including derived classes
o support for converting methods of objcclasses into selectors
(see tests/test/tobjc1.pp)
o support for loading from/storing to fields of objcclasses
o support for calling Objective-C methods using regular
Object Pascal syntax (see tests/test/tobjc1.pp)
o some things that are known to be not yet working:
o automatic conversion from ID to objcclasses and back
o declaring and implementing new objcclasses/methods in Pascal code
o debug information (objcclasses are currently plain pointers
as far as the debugger knows)
git-svn-id: branches/objc@13162 -
* changed type of minstacksize, maxstacksize and imagebase from aword
(= size of integer registers of the target platform) into puint
(= size of address space of target platform)
git-svn-id: trunk@13155 -
sized integer (succeed for Delphi, fail for objfpc)
* fixed compiler so it only allows typecasting a class/interface to a
differently sized integer in Delphi mode
git-svn-id: trunk@13152 -
dynamic library initialization (this automatically enables dead code
stripping/smart linking in all cases though, independent of -XX/CX;
on the other hand, there is no real reason not to use it anymore on
Mac OS X, since support was introduced in Xcode 1.5 for
Mac OS X 10.3.9 and that's also the earliest version we support
for running FPC on)
git-svn-id: trunk@13151 -
linker does not like two global labels both pointing at the end of a
section (causes "atom sorting error" warnings), and this happened for
units without any code after r13098
git-svn-id: trunk@13111 -
* Warning when 64-bit object file contains 32-bit absolute relocations. In such case an executable image can be loaded into lower 4Gb of address space only.
git-svn-id: trunk@13110 -
* Proper handling of 64-bit COFF relocations.
* Use jumps with 32-bit relative addressing in imports on win64.
* This fixes are related to bug #13657. The bug is not fully fixed yet...
git-svn-id: trunk@13104 -
* Write zero to relocs in debug sections, which point to unused sections, eliminated by linker.
* It seems internal linker creates correct dwarf debug info now.
git-svn-id: trunk@13093 -
can remove regallocs for the allocated register and reinstert them
before already processed instructions, thereby preventing the
peephole optimizer from ever seeing the allocation for this register
(mantis #13612)
git-svn-id: trunk@13085 -
use the alignment of these elements rather than their size (fixes
alignment of arrays of int64/qword/double on darwin/i386 and darwin/arm)
git-svn-id: trunk@13068 -
mode (mantis #11859)
* cleaned up some superfluous "eq=te_incompatible" checks (probably from
copy/pasting conditions from elsewhere)
git-svn-id: trunk@13050 -
initialization code of units/programs, because they are already
zero and initializing them explicitly is Delphi-incompatible
(mantis #13345)
git-svn-id: trunk@13042 -
* test now tests both ansistring and widestring constants, instead of only
ansistring
* fixed writing of widestring constants to ppu file
* fixed setting length of read widestring constants (don't only set maxlen)
* handle cross-endian reading of widestring constants
git-svn-id: trunk@13034 -
of a bitpacked record in assembler code
* convert the offsets of byte-aligned fields of bitpacked records from bits
to bytes (mantis #13563)
git-svn-id: trunk@13027 -
physical registers
* free the physical return registers at the caller side for 64 bit
systems
* make sure that we do not double-free registers in case a return
value is not used (mantis #13536)
git-svn-id: trunk@13023 -
attribute is available via the new append_attribute() method
(to enable easily adding extra attributes)
* write the definitions of methods as child entries of the
objectdef, instead of in the global scope
* only write DW_AT_calling_convention and DW_AT_external attributes
if their value is different from the DWARF default one
+ write DW_AT_virtuality and DW_AT_vtable_elem_location attributes
for virtual methods
* write the debug info for the hidden "self" parameter before all
other parameters and mark it as "artificial", because that is
how GDB distinguishes regular methods from static methods
git-svn-id: trunk@13003 -
Borland's i386 calling convention (its value may still change)
* specify the calling convention of each procedure in the DWARF info
* don't add the function result twice or even thrice to the debug
info, and don't add it at all to the parameter lists (even if it's
a hidden parameter, gdb will automatically add it according to the
ABI rules when calling the function)
-> calling "register" functions/procedures works with my experimental
gdb patch (methods not yet tested, and probably won't work yet)
git-svn-id: trunk@12995 -
when they are not used in the unit where they are defined (since they
may be used elsewhere, discovered while analyzing mantis #13345)
git-svn-id: trunk@12954 -
(and hence those directories were interpreted as source files themselves)
+ added -gostabsabsincludes option to make the compiler store absolute
paths to include files in Stabs debugging information (since no
search paths are used). This is not the default because it breaks
when only the base path of an entire source tree changes (relative
paths, which are the previous behaviour and the current default,
work fine, but the absolute paths don't anymore; this can happen
with shared network drives, and possibly with distros shipping
external .dbg files)
git-svn-id: trunk@12933 -
work correctly with DWARF)
* only force the internal linker for STABS when generating relocatable
output (rather than for all debugging formats)
git-svn-id: trunk@12917 -
inlining, because that results in alignment settings of "0"
(voidtype.savesize = 0 -> size2align(0) = 0), which is interpreted
by the code generators as unaligned (and is actually even invalid
nowadays). This caused errors on NDS when writing to memory mapped
I/O devices (mantis #13343)
git-svn-id: trunk@12916 -
* renamed al_dwarf enum to al_dwarf_frame since this enum only enables frame generation
* renamed create_dwarf procedure for the same reason
* skipped check for dwarf support in assember, since we generate all dwarf info ourselves
git-svn-id: trunk@12906 -
of in the DWARF, as it's more efficient and avoids a gdb bug which
occurs on some platforms (thanks to Jan Kratochvil for the tip,
http://sourceware.org/ml/gdb/2009-03/msg00094.html )
git-svn-id: trunk@12905 -
they always use the same parameter pushing order
* define the pushleftright_pocalls constant only for i386 so it cannot
be accidentally used anymore for other platforms
git-svn-id: trunk@12891 -
