vishapoberon/compiler
1
0
Fork 0
mirror of https://github.com/vishapoberon/compiler.git synced 2026-04-06 08:42:24 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Tidy (#41)

Browse source 
* Deduplicate common constants into OPM and do some source format tidying.

* Fix postpush buildall script to force checkout of updated buildall.

* Show enlistment branch in makefiles

* Support non-printables in string literals and tidy case alignment and constant literals.

* Common code for MIN and MAX of integer types.

* Common code for SInt/Int/LInt in ConstOp parameter preparation.

* Common code for SInt/Int/LInt in Op parameter preparation.

* Refactor SetIntType to work with byte size directly. Prepare to revert my incorrect VAL changes.

* Original meaning of VAL restored. Many library files disabled until use of VAL in 64 bits fixed.

* Make Reals.Mod independent of INTEGER size and add reals tests.

* Implement fraction, IsInfinity and IsNaN in oocLowReal.Mod.

* OPB little simplifications and ShorterSize/LongerSize functions.

* Add test for alignment computability

* Replace alignment constants with calculated alignment.

* typ.size aware OPV.Convert

* Add SYSTEM_INT64 and make tests name independent.

* Remove SYSTEM.H includes (string.h and stdint.h).

* Replace uses of uintptr_t and size_t with SYSTEM_ADDRESS.

* Sad hack to make FreeBSD and OpenBSD happy with memcpy declaration.

* Detect 64 bit on FreeBSD, and size_t defined on OpenBSD.

* %zd not supportd by mingw, cast strnlen return to int.

* Add debug for intermittent failure only on OpenBSD.

* Add textTexts as a confidence test and tidy up a couple of other tests.

* Update binary test process.
This commit is contained in:
David C W Brown 2016-08-25 14:41:00 +01:00 • committed by GitHub
parent 1f41d80b1e
commit da88496c5f
224 changed files with 7494 additions and 8065 deletions
Download patch file Download diff file Expand all files Collapse all files

2
src/system/Heap.Mod
View file

@ -186,7 +186,7 @@ MODULE Heap;
correctly regardless of the size of an address. Specifically on 32 bit address
architectures with 64 bit LONGINT, it loads 32 bits and extends it to LONGINT
rather than loading 64 bits. *)
PROCEDURE -FetchAddress(pointer: LONGINT): LONGINT "(LONGINT)(uintptr_t)(*((void**)((uintptr_t)pointer)))";
PROCEDURE -FetchAddress(pointer: LONGINT): LONGINT "(LONGINT)(SYSTEM_ADDRESS)(*((void**)((SYSTEM_ADDRESS)pointer)))";
PROCEDURE ExtendHeap(blksz: LONGINT);
VAR size, chnk, j, next: LONGINT;