summaryrefslogtreecommitdiff
path: root/compiler/ncgbas.pas
Commit message (Expand)AuthorAgeFilesLines
* * 16-bit and 8-bit ALU fixes in tcgtempdeletenode.pass_generate_codenickysn2013-12-231-2/+44
* * fixes several issues which cause warnings by the dfa code when using it to ...florian2013-12-011-0/+2
* * keep managed types in registers if possible. Under certain circumstances (i...florian2013-06-231-4/+2
* * report resolveref errors on the correct line numberjonas2012-12-151-7/+7
* * give a proper error when an mmx/fpu/sse register is used inside a memoryjonas2012-12-151-6/+7
* * support referencing "result" in nostackframe pure assembler routines ifjonas2012-08-211-0/+14
* * only don't pre-finalize funcret temps that are passed by reference to thejonas2012-07-231-1/+1
* * don't pre-finalize temp nodes that are passed by reference as hiddenjonas2012-07-231-2/+4
* + ttgobj.gethltemptyped() routine for allocating managed types with high leveljonas2011-08-201-1/+1
* * properly deal with reference temps in tcgtempdeletenode.pass_generate_codejonas2011-08-201-0/+15
* + "reference" temps that sort of implement pointer-style functionality forjonas2011-08-201-16/+42
* + gethltemp() to allocate a temp specifying full type information evenjonas2011-08-201-1/+1
* * replaced all calls to cg.g_incr/decrrefcount/g_initialize/g_finalize withjonas2011-08-201-2/+2
* * Fixed spell error revealed by lintian.mazen2010-10-061-2/+2
* * renamed pi_has_goto into pi_has_label because the problematic thing is if a...florian2010-08-051-3/+3
* * changed tprocdef.funcretloc[] from a tlocation into a tcgpara so it canjonas2010-05-301-36/+2
* * removed "and not is_class()" condition from is_managed_type(), becausejonas2010-05-221-2/+1
* + tempcreate nodes can now take an init. value, this allows to initilialize t...florian2010-01-101-0/+8
* * handle mm register properly when used as top_local, resolves #14566florian2010-01-031-0/+15
* * don't allocate persistent temps for classes (they are not refcounted)jonas2009-12-081-1/+2
* o added ARM VPFv2/VFPv3 support:jonas2009-12-031-3/+1
* * the compiler now explicitly keeps track of the minimally guaranteedjonas2009-02-081-5/+10
* * always specify an explicit alignment for tgobj.gettemp (so e.g.jonas2008-11-081-1/+1
* * split cpu64bit compiler define intojonas2008-02-131-4/+4
* * no longer remove AsmBlockStart and AsmBlockEnd markers fromjonas2007-09-221-12/+1
* * changed boolean fields in ttempinfo to a set for easierjonas2007-07-291-11/+11
* * fixed calling inline functions (with exit statements) from insidejonas2007-07-201-0/+7
* * only check validity of assembler instructions in pass_2, after alljonas2007-05-291-1/+11
* * remove tdictionary and tindexarraypeter2006-11-031-1/+1
* * settings refactoredflorian2006-10-301-3/+3
* * remove ttypepeter2006-10-291-27/+27
* * fixed memory leak with temp. variables, fixes #7282florian2006-10-221-0/+3
* * fixed compilation on non x86 systemsflorian2006-04-201-0/+2
* * Support in tempnodes for SSE floats in XMM registerspeter2006-04-201-7/+35
* Merged revisions 2921-2922,2925 via svnmerge from peter2006-03-161-29/+29
* * rewrite of optimizer optionspeter2006-03-131-2/+2
* Merged revisions 2669,2673,2677,2683,2696,2699-2702,2704,2708,2712-2715,2718,...peter2006-03-051-5/+3
* * tai_const refactoredflorian2006-01-221-6/+1
* * optimized register allocation for tempnodesjonas2006-01-081-18/+40
* * support multiple asmlabel types, renamed getlabel topeter2005-09-181-1/+1
* * ie on powerpc fixed: synching of fpu register variables wasn't implemented,...florian2005-09-131-0/+9
* + internal error addedflorian2005-09-101-0/+2
* * inline and compilerproc are now procoptions instead of proccall typesjonas2005-07-181-1/+1
* * log and id tags removedfpc2005-05-211-10/+0
* initial importfpc2005-05-161-0/+511