the frame of the caller rather than of the current routine, which could
cause e.g. stackframes to be missed in RTL helpers that are known to be
only called from run time error routines)
git-svn-id: trunk@44064 -
out-of-range constants, because of the comments mentioned in #35753
(except for enums, as apparently Delphi does the same)
* added range check warnings about explicit type casts that throw away
bits (e.g. byte($fff)), without giving warnings for most common cases
(like cardinal(-1))
* fixed masking/sign exting constant array indices (must be based on index
range type size/signedness rather than on array size/"signedness")
git-svn-id: trunk@42275 -
Allows compilation of compiler using -dDEBUG_NODE_XML
which will generate a NAME-node-dump.xml file for each
unit, program or library compiled,
containing a XML description of the nodes handled during
compilation of the unit, program or library.
git-svn-id: trunk@42271 -
sense as a function (the original input needed to be replaced by its
result to make any sense), and someone already mistakenly used it like that
in htypechk causing a bug/memory leak
git-svn-id: trunk@40729 -
transforming inc/dec to an addn/subn in case of range checking) from
triggering the transformation logic for accessing nested variables
on LLVM/JVM multiple times
* also prevent it from triggering errors during inlining, in case a
parameter to a nested routine gets replaced by a local variable
from a parent routine (in that case, it does not need to be
accessed via the parentfpstruct, as we are being inlined in the
parent) -> don't duplicate the check from pass_typecheck in pass_1,
but check whether it needs to be accessed via the parentfpstruct
by verifying that tloadnode.left is assigned (which pass_typecheck
will ensure if needed)
git-svn-id: trunk@40630 -
* disable matching volatile references in the assembler optimisers, so they
can't be removed (more conservative than needed, but better than removing
too many)
o the CSE optimiser will ignore them by default, because they're an unknown
inline node for it
* also removed no longer used fpc_in_move_x and fpc_in_fillchar_x inline node
identifiers from rtl/inc/innr.inc, and placed fpc_in_unaligned_x at the
right place
git-svn-id: trunk@40465 -
An uninitialized function Result of a managed type needs special handling.
When passing it as a var parameter a warning need to be emitted, since a user
may expect Result to be empty (nil) by default as it happens with local vars
of a managed type. But this is not true for Result and may lead to serious issues.
The only exception is SetLength(Result, ?) for a string Result. A user always
expects undefined contents of the string after calling SetLength(). In such
case a hint need to be emitted.
+ Tests for this.
git-svn-id: trunk@40216 -
+ mask underflow and precision on startup
+ check for floating point exceptions after inlined float routine helpers
- do not check for floating point exceptions after floating point moves
git-svn-id: branches/laksen/riscv_new@39645 -
checking is on, when handling the inc/dec inline nodes, instead of using
current_settings.localswitches
* when creating inline nodes in the optloadmodifystore optimization pass, copy
localswitches from the node, that is being replaced, because otherwise,
localswitches is copied from current_settings.localswitches at the time the
new node is created, and that can already be in a different state, since
optloadmodifystore is performed in a separate pass, after the current
procedure has already been parsed and in this moment, it reflects the state
of localswitches after the end of the procedure.
* these two fixes fix a bug, where an internalerror 2017032701 can happen, when
compiling with -O3 code that turns on and off range/overflow checking in the
middle of a procedure.
git-svn-id: trunk@36195 -
* 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 -
optimization, which prevented the optimization from ever being performed. This
should also fix the test failure of tbs/tb0627b.pp on all 64-bit platforms.
git-svn-id: trunk@36071 -
less-than-64-bit CPUs that have a 64-bit OP_SHR/OP_SHL/OP_SAR implementation
in their cg64 backend code generator. This is enabled only for i386 for now.
git-svn-id: trunk@36022 -
be used (TBD in a future commit) for optimizing x:=-x and x:=not x on CPUs
that support performing these operations directly in memory (such as x86)
git-svn-id: trunk@35749 -
op=and/or/xor). They generate more optimal code on certain architectures
(including x86). The new inline nodes aren't generated by the compiler yet,
but will be used in the future, at certain optimization levels, whenever the
pattern x:=x op y is detected by the compiler.
git-svn-id: trunk@35666 -
Since overloading compilerprocs does not work each procedure got its own unique name, but they are using the new compilerproc extension to map them to the Insert and Delete symbol so that error messages can be shown with the respective name for the procedure declarations instead of fpc_shortstr_delete for example.
git-svn-id: trunk@33895 -
ninl.pas, tinlinenode:
* pass_typecheck: let code pass on to simplify() for undefineddefs in Low() and High()
* simplify: create a constant 0 as fallback
+ added test
git-svn-id: trunk@33884 -
with a ctemprefnode.create_offset() hack into a proper array, because
the old way does not work at all with LLVM (and is ugly)
git-svn-id: trunk@33724 -
