summaryrefslogtreecommitdiffstats
path: root/generic
Commit message (Collapse)AuthorAgeFilesLines
* split off TclInitStubTable() as separate function - which does the actual ↵jan.nijtmans2013-06-163-24/+80
| | | | stub table initialization - previously part of Tcl_InitStubs().
* Fix env.test when running mingw-w64 (AMD64 build) test-suite on wine64. jan.nijtmans2013-06-131-9/+9
|\ | | | | Fix some gcc warnings which become visible with -Wextra
| * Fix some gcc warnings which become visible with -Wextrajan.nijtmans2013-06-132-28/+28
| |
* | Revise INST_EXPAND_STKTOP so that it no longer makes use of its operand.dgp2013-06-111-14/+23
| | | | | | | | | | | | | | | | | | | | All the information required to do a proper expansion of the exec stack to support expanded command invocation is already present. The operand doesn't provide any essential information. By ignoring it, we eliminate the risk that the compiler might fill in the operand with a bad stack depth estimate value. INST_EXPAND_STKTOP doesn't need an operand, but in order to support loading of existing bytecodes we cannot change it now. There's also no need to change what the compiler tries to place in the operand, though changing it to always be zeros would be acceptable now.
* | Select improvements in stack depth estimates brought over from mig-review.dgp2013-06-115-71/+17
| | | | | | | | | | Mostly these are just simplifications, removing code that wasn't needed. Some changes make the stack depth estimate more accurate instruction by instruction.
* | Repairs to compile/exec debugging output.dgp2013-06-112-8/+16
| |
* | Stack housekeeping repair for last checkin.dgp2013-06-111-0/+2
| |
* | [33b7abb8a2] [7174354ecb] Rewrite TclCompileThrowCmd().dgp2013-06-111-59/+46
| |
* | Improve compatibility detection for <time.h> and <sys/stat.h>: jan.nijtmans2013-06-111-3/+4
|\ \ | |/ | | | | - Move <time.h> before other includes on Windows, so we are sure the time_t definition being checked doesn't come from <sys/types.h>. - Padding at the end of Tcl_StatBuf doesn't influcence binary compatibility, so relax panic check accordingly.
| * Improve compatibility detection for <time.h> and <sys/stat.h>: jan.nijtmans2013-06-111-3/+4
| | | | | | | | - Move <time.h> before other includes on Windows, so we are sure the time_t definition being checked doesn't come from <sys/types.h>. - Padding at the end of Tcl_StatBuf doesn't influcence binary compatibility, so relax panic check accordingly.
* | Fix the problems with code generation; behavior now appears correct.dkf_better_try_compilationdkf2013-06-091-129/+204
| |
* | Improving tests, fixed one case.dkf2013-06-091-27/+44
| |
* | Working on a better compiler for [try]; found some bugs in previous compilationdkf2013-06-081-23/+123
| | | | | | code which aren't resolved yet.
* | More informative comment describing INST_SYNTAX.dkf2013-06-081-1/+2
| |
* | Factor out stereotypical ways of getting variable indices.dkf2013-06-083-160/+63
| |
* | Simplify stack depth management.dkf2013-06-071-31/+33
| |
* | More efficient instruction sequence for [dict for] with correct exception ↵dkf2013-06-061-23/+10
| | | | | | | | depth handling.
* | 3614382 Fix stack management of compiled [dict for] by shifting limitsdgp2013-06-061-3/+4
| | | | | | of the catch range.
* | 3614360 Repair stack demands of optimized compiled [return LITERAL].dgp2013-06-061-0/+1
| |
* | Working on the optimizer.dkf2013-06-062-82/+223
|\ \
| * | Minor grammar fix.dkf2013-06-061-3/+3
| | |
| * | Split the optimizer up. Remove the dreaded 'goto' from which doesn't need it.dkf2013-06-061-81/+127
| | |
| * | Added optimizing of jump-to-nop and jump-to-jump cases. Ta to AK for suggesting.dkf2013-06-051-57/+152
| | |
* | | Working on the optimizer, and finding a few bugs in bytecode generation.dkf2013-06-054-176/+290
|\ \ \ | |/ /
| * | Added the optimizer...dkf2013-06-051-0/+287
| | |
| * | Corrected wrong information about instruction width that was causing an ↵dkf2013-06-051-1/+1
| | | | | | | | | | | | optimizer crash.
| * | merge fixesdkf2013-06-052-17/+9
| |\ \ | |/ / |/| |
| * | More cleaning up; factor out optimizer to new file. Some weird problems still.dkf2013-06-053-187/+18
| | |
* | | Repair TCL_COMPILE_DEBUG guardsbug_3614342dgp2013-06-051-5/+7
| | |
* | | Stack Depth fixups.dgp2013-06-052-12/+4
|/ /
* | merge trunkdkf_compile_improvementsdkf2013-06-051-7/+7
|\ \
| * \ fix for perf bug detected by Kieran ↵mig2013-06-031-7/+7
| |\ \ | | |/ | | | | | | (https://groups.google.com/forum/?fromgroups#!topic/comp.lang.tcl/vfpI3bc-DkQ)
| | * fix for perf bug detected by Kieran ↵mig2013-06-031-5/+6
| | | | | | | | | | | | (https://groups.google.com/forum/?fromgroups#!topic/comp.lang.tcl/vfpI3bc-DkQ)
| | * Fix C99 comment-ism breaking the native AIX compiler. HPUX likely as well.andreask2013-05-291-2/+2
| | |
* | | Stack cleanup works now even in the most evil expansion cases.dkf2013-06-055-60/+130
| | |
* | | Next stage of fixing the break/continue generation.dkf2013-06-034-33/+61
|/ /
* | Generate [continue] optimally in [for] next clauses. Add tests for Bug 3614226.dkf2013-06-033-8/+51
| |
* | Fix a stack depth calculation.dkf2013-06-021-1/+1
| |
* | Remove useless macro, use existing macro where it makes sense.dkf2013-06-022-5/+2
| |
* | Many improvements to code generation of efficient break and continue.dkf2013-06-024-81/+235
| |
* | Getting better at doing more efficient break/continue instruction handling.dkf2013-06-013-50/+111
| |
* | derpdkf2013-05-301-2/+2
| |
* | Working towards the next batch of optimizations.dkf2013-05-304-19/+168
| |
* | Corrected code generation when doing the second run with an 'infinite' loop.dkf2013-05-301-1/+1
| |
* | Simplifications and tidying up of stack management issues. Work in progress.dgp2013-05-291-14/+13
| |
* | 3614102 - Reset stack housekeeping when compileProc fails.dgp2013-05-291-2/+6
| |
* | Stop emitting the instructions INST_*_SCALAR_STK*. They are identical todgp2013-05-296-94/+67
| | | | | | | | | | | | | | their INST_*_STK* counterparts. Having done that, it is clear the "simpleVarName" return from TclPushVarName provides nothing of use to any of its callers. Eliminate that. Also make TPVN return void, instead of int. Bring the TPVN header comments up to date; they were quite rotten.
* | Use the routines that provide "basic compile" instead of reinventing them.dgp2013-05-281-34/+16
| |
* | Plug the memory leak. Greater ambitions to improve this routine are provingdgp2013-05-281-0/+1
| | | | | | more difficult than expected.
* | Missed bits of dup code elimination.dgp2013-05-281-23/+0
| |
generated by cgit v0.12 at 2025-03-27 11:42:31 (GMT)