rawbytestring so the compiler does not try to convert it to the declared
string type of the function result, as this may not be available
(mantis #32510)
git-svn-id: trunk@37889 -
* compinnr.inc include file converted to a unit
* inline number field size stored in ppu increased from byte to longint
* inlines in the parse tree (when written with the -vp option) now printed with
their enum name, instead of number
git-svn-id: trunk@36174 -
code from one unit to another, but in splitting some methods
I made a couple of mistakes :/ (causing global variables to be
initialised three times)
git-svn-id: trunk@35461 -
for the InitFinalTable, because they are declared as ALUUInt in the RTL and
sinttype doesn't always correspond to the ALU size (it differs on platforms
with the combination of a 64-bit ALU and 32-bit address space)
git-svn-id: trunk@34629 -
- register a symbol in the unit's public or extern asm symbol list
- add an indirect symbol when defining the symbol
- use an indirect symbol instead of the direct one
This will be needed to correctly handle resourcestrings for packages
git-svn-id: trunk@34182 -
symbol is defined
o removed all places where AB_INDIRECT symbols were explicitly generated
o only generate AB_INDIRECT symbols for AT_DATA on systems_indirect_var_imports
o for some symbols an indirect symbol is always required (because they are
dereferenced by code in RTL units) -> use new AT_DATA_FORCEINDIRECT type
git-svn-id: trunk@34165 -
tasmdata.DefineAsmSymbol() and all routines that call it
o will be used to automatically generate AB_INDIRECT sybols when
necessary
git-svn-id: trunk@34164 -
compiler/ngenutil.pas, tnodeutils:
* InsertThreadvarTablesTable: reference a unit's (and the program's) threadvar table using a indirect symbol
* InsertThreadvars: generate an indirect symbol for the threadvar table
rtl/inc/threadvr.inc:
* TltvInitTablesTable: add an additional indirection for the tables field
git-svn-id: trunk@34043 -
Adjust non-Windows resources to work with indirect main information as well.
rtl/inc/systemh.inc, TEntryInformation:
+ new cross platform field ResLocation which stores the pointer to the resources
rtl/inc/intres.inc:
* change the type of ResHeader from PResHdr to PPResHdr (and adjust code that uses it accordingly; Note: the first dereferencing is assumed to be always valid!)
* adjust declaration of ResHeader depending on whether indirect main information is used or not
rtl/darwin/sysinit.pas & rtl/linux/si_impl.inc:
* pass the location of the resources through SysInitEntryInformation
rtl/bsd/system.pp & rtl/linux/system.pp:
+ new public variable FPCResLocation which is setup by SysEntry
rtl/win32/sysinit.inc:
* initialize ResLocation of SysInitEntryInformation as Nil
rtl/win32/system.pp:
* initialize ResLocation of EntryInformation as Nil
........
For systems that support both internal and external resources (in this case only Darwin ones) we need to generate the FPC_RESLOCATION symbol always, because the SysInit unit is resource type agnostic.
ngenutil.pas, tnodeutils:
* InsertResourceInfo: generate FPC_RESLOCATION symbol pointing to 0 in case the program is compiled on Darwin with resources set to external ones (-We)
git-svn-id: trunk@33952 -
ngenutil.pas, tnodeutils:
* InsertInitFinalTable: the init and finalize functions are in the *local* symboltable not the global one
* InsertInitFinalTable.add_initfinal_import: reenable the internal error
git-svn-id: trunk@33508 -
ngenutil.pas, tnodeutils.InsertInitFinalTable:
* add_initfinal_import: disable the internal error until I've found out why the INIT$ function isn't there
git-svn-id: trunk@33504 -
Also add the initialization/finalization sections and the class constructors/destructors to the imported symbols of the current program/library.
ngenutil.pas, tnodeutils.InsertInitFinalTable:
+ new function add_initfinal_import to find the procsym for the initialization/finalization code and to add it to the imported unit symbols
* write_struct_inits: import the found class constructor/destructor symbols if they are not in the program/library unit
* use add_initfinal_import for the initialization and finalization sections of all used units beside the program/library unit
........
git-svn-id: trunk@33501 -
Note: It needs to be checked what adjustments will need to be done for JVM and LLVM
Note 2: The indirect symbols will always be generated even if the target won't use them so that assembler code can be consistent among the targets (e.g. i386-win32 needs them while i386-linux does not)
ptconst.pas, read_typed_const:
* generate the indirect symbol in a read only data section
ngenutil.pas, tnodeutils:
* insertbsssym: generate the indirect symbol in a read only data section
git-svn-id: trunk@33279 -
accessed by a nested routine, assign it back to the original funcretsym
at the end because the code generator is hardcoded to use funcretsym.
We can't replace funcretsym with an absolutevarsym that redirects to
the parentfpstruct field, because the code generator can only handle
simple variables
git-svn-id: trunk@32687 -
const builder
o also put the resourcestring tables table in rodata instead of in data,
since it doesn't change at run time (unlike the individual resourcestring
tables)
git-svn-id: trunk@32397 -
section to the high level typed const builder, and use this functionality
for the resource strings (generating items: todo)
* fixed dead stripping unused resource strings (all string data was put in the
same section since r30342)
git-svn-id: trunk@32394 -
* extend tcallnode with the ability to pass a tspecializationcontext so that tcallcandidates can do a final specialization
* the final procdef is registered at the end of tcallnode.pass_typecheck
git-svn-id: trunk@31763 -
* for now all procdefs are created as registered
Note: an additional parameter instead of an overload is used for tprocdef.create as otherwise both constructors would need to be overridden in potential descendant CPU-specific classes...
git-svn-id: trunk@31593 -
o made all (non-abstract) tdef and tsym constructors virtual
o added c*def/c*sym classref types for every (non-abstract) t*def/t*sym
class
o added cpusym unit for every architecture that derives a tcpu*def/tcpu*sym
class from the base classes, and initialises the c*def/c*sym classes with
them. This is done so that the llvm target will be able to derive from
the tcpu*def/sym classes without umpteen ifdefs, and it also means that
the WPO can devirtualise everything because the c* variables are only
initialised with one class type
o replaced all t*def/t*sym constructor calls with c*def/c*sym constructor
calls
git-svn-id: trunk@27361 -
overridden by the llvm target
* override a bunch of other ngenutil methods with empty stubs for things
that are not yet supported when targeting llvm
git-svn-id: branches/hlcgllvm@26051 -
that may trigger JVM bytecode verification errors if they are used
before they are initialised (this includes passing them as a "var"
parameter) + test
* sorted -CT parameters alphabetically and alligned them
git-svn-id: trunk@25387 -
ngenutil.pas, AddToStructsInit:
* don't add class constructors/destructors to the init table if they belong to a generic type
+ test
git-svn-id: trunk@25234 -
global, and no longer make it global for all symbols when smartlinking
is enabled on systems that use section-based smartlinking
git-svn-id: trunk@22241 -
and also use this code to trash local variables in inlined routines
o fixes mantis #22088
o makes it possible to also implement it for the jvm target in the future
git-svn-id: trunk@21393 -
o support for the new codepage-aware ansistrings in the jvm branch
o empty ansistrings are now always represented by a nil pointer rather than
by an empty string, because an empty string also has a code page which
can confuse code (although this will make ansistrings harder to use
in Java code)
o more string helpers code shared between the general and jvm rtl
o support for indexbyte/word in the jvm rtl (warning: first parameter
is an open array rather than an untyped parameter there, so
indexchar(pcharvar^,10,0) will be equivalent to
indexchar[pcharvar^],10,0) there, which is different from what is
intended; changing it to an untyped parameter wouldn't help though)
o default() support is not yet complete
o calling fpcres is currently broken due to limitations in
sysutils.executeprocess() regarding handling unix quoting and
the compiler using the same command lines for scripts and directly
calling external programs
o compiling the Java compiler currently requires adding ALLOW_WARNINGS=1
to the make command line
git-svn-id: branches/jvmbackend@20887 -
o these classes get an "enum" flag in the class files
o these classes get a class field (whose type is that same enum
class) per enum in the type, which also gets the "enum" flag
o those class fields are initialised in the class constructor
with the name of the enum and their order in the declaration
o if the enum has jumps in FPC (lowest value is not 0, or not
all values are contiguous), then we add an extra field
to hold the FPC ordinal value of the enum
o these classes get a class field valled $VALUES that contains
a reference to the aforementioned class fields in order of
declaration (= ordinal->instance mapping, JDK-mandated)
o apart from the JDK-mandated instance methods (values, valueOf),
also add FPCOrdinal (returns FPC ordinal value; same as order
of declaration in case of no jumps) instance method and FPCValueOf
(returns enum corresponding to FPC ordinal value) static class
method
o the mapping between FPC ordinals and enum instances in case of
jumps is stored in a hashmap whose size is the next prime number
greater or equal than the number of enum elements
o moved several extra JDK types to the system unit for the enum
support, and for future boxing and Java set support
o several new synthetic method identifiers to generate the enum class
methods/constructor/class constructor
o enums with jumps are ordered by FPC ordinal value in the JVM
$VALUES array so that the java.lang.Enum.doCompare() method
will properly compare them
git-svn-id: branches/jvmbackend@18616 -
o moved several routines from pmodules to ngenutil and overrode them
in njvmutil (for unit initialisation tables, resource strings, ...)
o force the evaluation stack size to at least 1 for the main program,
because the unit initialisation triggers are inserted there afterwards
and they require one stack slot
git-svn-id: branches/jvmbackend@18507 -
assignment-nodes. For global typed constants and typed constants/
local variable initialisers in regular functions/procedurs, the
assignments are performed in the unit initialisation code. For
those in object/record definitions and their methods, it's done
in the class constructor. Since we may not yet have parsed all
method implementations when the class constructor is parsed, part
of these may be initialised in a helper routine called from the
class constructor. The ones known when the class constructor is
parsed are inited there, because the ones marked as "final" and
declared as static class fields must be initialised in the class
constructor for Java
o new set systems_typed_constants_node_init in systems unit that
indicates that a target uses node trees to initialise typed consts
instead of an initialised data section
o mark typed constants in {$j-} mode as "final" for JVM
o mangle the name of staticvarsyms inside localtables a bit to avoid
name clashes (only with procedure names for now, no parameters yet
so can still cause problems with overloaded routines)
o after a routine has been parsed, it is now processed by
cnodeutils.wrap_proc_body(), which can add extra nodes before code
generation (used for injected the typed constant node trees)
git-svn-id: branches/jvmbackend@18475 -
and initialise global variables that are wrapped (records, arrays)
in those sections
o check whether pd.localst is assigned in dbgjasm, because it's
not for the unit initialisation routine
o moved insertbssdata() from ncgutil to ngenutil and override it
njvmutil (it does nothing in the latter, since global variables
are added as fields to the class representing the unit; the
initialisation is done in gen_initialize_code() in thlcgjvm)
o added force_init() and force_final() methods to ngenutil, so
that targets can force init/final routines separate from the
regular managed types infrastructure (used by JVM for forcing
an init section in case of records/arrays)
git-svn-id: branches/jvmbackend@18460 -
nutils.pas into virtual class methods of a new tnodeutils class defined
in ngenutil (global factory: cnodeutils), so they can be overridden by
architecture-specific implementations (required by the JVM backend)
git-svn-id: branches/jvmbackend@18364 -
