summary refs log tree commit diff stats
path: root/compiler/ccgexprs.nim
Commit message (Collapse)AuthorAgeFilesLines
* Fixes #1005Dominik Picheta2014-04-151-0/+3
|
* resolved conflictAndreas Rumpf2014-04-101-143/+158
|\
| * implemented region pointersAraq2014-04-091-4/+4
| |
| * Merge branch 'devel' of github.com:Araq/Nimrod into develZahary Karadjov2014-03-061-1/+1
| |\
| | * implemented 'union' and 'unchecked' pragmasAraq2014-03-051-1/+1
| | |
| * | split the inline and closure iterators into different symbol kinds for ↵Zahary Karadjov2014-03-061-1/+1
| |/ | | | | | | easier discrimination between them
| * vm2: new representation of registersAraq2014-02-211-1/+0
| |
| * fixes #914Araq2014-02-141-1/+1
| |
| * fixes #887Simon Hafner2014-02-091-1/+1
| |
| * bugfix: codegen issue that caused getGMTime() to leak memoryAraq2014-02-071-1/+1
| |
| * case consistency for evalffiAraq2014-02-011-1/+1
| |
| * compiler warns when you use GC'ed memory and '--gc:none'Araq2014-01-191-0/+16
| |
| * 'nil' as a statement is deprecated, use an empty 'discard' insteadAraq2014-01-191-4/+4
| |
| * case consistency: cs:partial bootstraps on windowsAraq2013-12-291-18/+18
| |
| * case consistency: next stepsAraq2013-12-291-8/+8
| |
| * case consistency part 4Araq2013-12-271-102/+102
| |
| * case consistency part 1Araq2013-12-271-2/+2
| |
* | Merge pull request #691 from rbehrends/fix-shallow-pragmaAndreas Rumpf2014-04-091-7/+39
|\ \ | |/ |/| Fixes shallow pragma for objects and tuples.
| * Fixes various issues with shallow copying.Reimer Behrends2013-12-011-7/+39
| | | | | | | | | | | | | | | | | | | | | | | | | | * The {.shallow.} pragma is now properly respected for objects and tuples that are not handled by genGenericAsgn(). Mirroring the decision in generic assignments, whether an assignment is shallow depends on the destination of the assignment, but not the source. * String literals and complex inline constants are now properly assigned a location kind of locData. * Assignments from a locData source will now force deep copying to occur, even if shallowCopy or {.shallow.} is being used. This avoids potential crashes from `shallowCopy s, "xxx"` or from `shallowCopy t, ("a", "b")`.
* | fixes a regression where memset was used without including <string.h>Araq2013-12-031-0/+6
|/
* fixes #575Araq2013-09-101-2/+1
|
* fixes #569; C++ codegen works againAraq2013-08-311-12/+7
|
* float64 is now an alias to 'float'; fixes #545Araq2013-08-301-6/+8
|
* hacky fix for generic constraints matchingZahary Karadjov2013-08-191-2/+2
|
* fixes #411Araq2013-05-261-8/+12
|
* attempt to fix a codegen bugAraq2013-05-201-1/+7
|
* fixes #432, fixes #427Araq2013-05-191-3/+28
|
* fixes #439Araq2013-05-191-1/+1
|
* 'result/return' triggers a void contextAraq2013-05-061-3/+13
|
* completed expr/stmt unificationAraq2013-05-031-109/+113
|
* first steps to the expr/stmt unificationAraq2013-04-301-3/+2
|
* fixes #287; bugfix: subrange checking is performed againAraq2013-04-121-2/+8
|
* fixes #366Araq2013-03-171-1/+4
|
* Removes executable bit for text files.Grzegorz Adam Hankiewicz2013-03-161-0/+0
|
* documented object constrs; endb works againAraq2013-03-091-30/+5
|
* object constructors work nowAraq2013-03-081-82/+116
|
* next steps for object construction expressionsAraq2013-03-071-1/+32
|
* next steps for object construction expressionsAraq2013-03-071-22/+26
|
* fixes for the new overloading resolutionAraq2013-03-031-1/+2
|
* make some tests greenAraq2013-03-031-1/+1
|
* codegen produces better code for complex assignmentsAraq2013-02-241-2/+61
|
* cleaner GC switchingAraq2013-01-311-5/+5
|
* merged upstream masterZahary Karadjov2013-01-271-5/+14
|\
| * fixed some closure related bugsAraq2013-01-221-1/+1
| |
| * fixes #270Araq2013-01-081-1/+1
| |
| * first version of ropes.nim with unsafeNew (broken)Araq2012-12-021-3/+12
| |
* | cleaned up some debugging codeZahary Karadjov2013-01-201-10/+0
| |
* | track the "owner" heap object in the ref write barrierZahary Karadjov2012-12-011-4/+20
| | | | | | | | See the papers for reference counting with heap sliding views for details:
* | disables the compile-time rope formatting during bootstrappingZahary Karadjov2012-11-281-90/+90
| |
* | first batch of rope code converted to `rfmt`Zahary Karadjov2012-11-281-37/+37
|/ | | | | | The results so far are inconclusive: The rope cache tries have been reduced by more than 125K (nearly 20%), but the bootstrap time seems to be ever slightly increasing (0.01s).