- 04 Nov, 2018 1 commit
-
-
Vladimir Sedach authored
-
- 03 Nov, 2018 3 commits
-
-
Vladimir Sedach authored
-
Vladimir Sedach authored
-
Vladimir Sedach authored
-
- 02 Nov, 2018 1 commit
-
-
Vladimir Sedach authored
-
- 25 Oct, 2018 6 commits
-
-
Vladimir Sedach authored
Thanks to Olaf Ruppert for the suggestion.
-
Vladimir Sedach authored
Deprecated forms are no longer generated by LOOP or other macros
-
Vladimir Sedach authored
-
Vladimir Sedach authored
-
Vladimir Sedach authored
-
Vladimir Sedach authored
-
- 24 Oct, 2018 2 commits
-
-
Vladimir Sedach authored
-
Vladimir Sedach authored
-
- 16 Oct, 2018 2 commits
-
-
Vladimir Sedach authored
-
Vladimir Sedach authored
-
- 15 Oct, 2018 2 commits
-
-
Vladimir Sedach authored
-
Vladimir Sedach authored
-
- 13 Oct, 2018 1 commit
-
-
Vladimir Sedach authored
-
- 11 Oct, 2018 1 commit
-
-
Vladimir Sedach authored
-
- 04 Oct, 2018 1 commit
-
-
Vladimir Sedach authored
-
- 07 Jul, 2018 1 commit
-
-
Vladimir Sedach authored
-
- 18 Feb, 2016 1 commit
-
-
Jason Miller authored
Previously multiple-values were very fragile. Examples of things that didn't work, but now do: (multiple-value-bind (x y) (values 1 2) y) (defun foo () (values 1 2)) (defun bar () foo) (multiple-value-bind (x y) (bar) y) Biggest downside of this change is that every return statement must be prefixed by a clearing of the MV register.
-
- 09 Jun, 2015 1 commit
-
-
Boris Smilga authored
Additions to 116f9ca9, to not replace Φ() with Φ.call(this) unless Φ refers to this. Also, fixed / added tests.
-
- 05 Mar, 2015 1 commit
-
-
Jason Miller authored
-
- 09 Feb, 2015 1 commit
-
-
Boris Smilga authored
Fixed: an occurence of the macro WITH-DECLARATION-EFFECTS preceded its definition, leading to ps-compile errors (see https://github.com/vsedach/Parenscript/issues/15).
-
- 25 Dec, 2014 1 commit
-
-
Boris Smilga authored
1. Named function bodies in Common Lisp may contain multiple DECLARE forms before or after the docstring (or even before *and* after). This feature is important for meta-programming by macros, and so PS should also support it. 2. Declarations should be allowed in PS forms whose Lisp namesakes allow them, to wit: in the macros WITH-SLOTS, MULTIPLE-VALUE-BIND, DO*, DO, DOTIMES, DOLIST, DESTRUCTURING-BIND, LET*, DEFUN, DEFSETF, and in the special forms LET, FLET, LABELS (letting out DEFMACRO, MACROLET, SYMBOL-MACROLET who are not translated to JavaScript). Before the present commit, most of them didn't handle declarations correctly. 3. Individual DECLARE forms should be allowed to contain multiple declaration specifiers, in particular, multiple SPECIAL specifiers. WITH-DECLARATION-EFFECTS used to ignore SPECIALs after the first one. 4. This commit also introduces LOCALLY as PS form because it simplifies the implementation of stuff from (2).
-
- 18 Dec, 2014 1 commit
-
-
Boris Smilga authored
-
- 30 Sep, 2014 1 commit
-
-
Boris Smilga authored
See https://groups.google.com/forum/#!topic/parenscript/eQ4NS5lslHE for a more detailed description of the bug.
-
- 10 Jul, 2014 1 commit
-
-
Fred LeMaster authored
Added a unary-plus to the IR and removed the automatic conversion of (setf x (logior x 0)) to "x |= 0;".
-
- 14 Mar, 2013 4 commits
-
-
Daniel Gackle authored
-
Daniel Gackle authored
-
Daniel Gackle authored
the poor-man's codewalking of try-expressionizing-if?. Example: (try-expressionizing-if? '(for-in (label (foo)))) The codewalker macroexpands (label (foo)) even though it is not a macro invocation. Since LABEL is a deprecated macro, a warning was leaking through.
-
Daniel Gackle authored
inside RETURN-EXP so it presents one interface to the outside world. This makes the implementation of source mappings a little easier.
-
- 31 Jan, 2013 1 commit
-
-
Daniel Gackle authored
whether or not to expressionize an IF.
-
- 09 Oct, 2012 1 commit
-
-
Vladimir Sedach authored
-
- 08 Oct, 2012 1 commit
-
-
Vladimir Sedach authored
First try at a maybe-once-only macro that doesn't introduce bindings for constant and variable expressions in PS code
-
- 15 Sep, 2012 1 commit
-
-
Vladimir Sedach authored
Fixed problem with loop-return-arg and loop bodies getting double compiled by getting rid of loop-return-arg. Thanks to Daniel Gackle for the bug report.
-
- 13 Sep, 2012 1 commit
-
-
Vladimir Sedach authored
Removed the pass-through capability from the new global variable-based multiple value return system and prettied up associated code generation.
-
- 12 Sep, 2012 1 commit
-
-
Vladimir Sedach authored
Full multiple values implementation. Works correctly but produces ugly code. This is here for later use, but will be rolled back in the next patch.
-
- 11 Aug, 2012 1 commit
-
-
Vladimir Sedach authored
-