[parenscript-devel] Bug: macroexpansion broken by lexical scoping change
Here's one thing that the latest PS breaks in our code. (ps (list 1 2 3)) => "[1,2,3];" (ps (let ((list nil)) (setf list (list 1 2 3)))) => "var list5180 = null; list5180 = list5180(1, 2, 3);" I think this behavior is incorrect. It is quite surprising for a lexical variable to shadow a macro definition. Daniel
Just pushed out a patch that fixes this. The solution turned out to be a lot more involved than I originally suspected - basically the PS compiler had to be refactored to give control of macroexpansion over to special forms (the way it was previously handling macroexpansion was quite naive, but would have been suitable for a single-namespace Lisp). The changes that have been going on in the recent while are starting to make Parenscript almost look like a "real Lisp compiler." Vladimir On Wed, May 6, 2009 at 2:00 PM, Daniel Gackle <danielgackle@gmail.com> wrote:
Here's one thing that the latest PS breaks in our code.
(ps (list 1 2 3)) => "[1,2,3];"
(ps (let ((list nil)) (setf list (list 1 2 3)))) => "var list5180 = null; list5180 = list5180(1, 2, 3);"
I think this behavior is incorrect. It is quite surprising for a lexical variable to shadow a macro definition.
Daniel
_______________________________________________ parenscript-devel mailing list parenscript-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel
Given the recent bug fixes, I tried to upgrade PS again. I'm now encountering an error on trying to expand a simple loop form. The error appears in commit 5a69278. (ps-macroexpand '(loop for x from 1 below 10 do (bah))) => (NIL . #<Anonymous Function #x300042319FEF>) is not of type (OR SYMBOL FUNCTION), and can't be FUNCALLed or APPLYed [Condition of type TYPE-ERROR] Daniel On Sat, May 9, 2009 at 5:29 PM, Vladimir Sedach <vsedach@gmail.com> wrote:
Just pushed out a patch that fixes this. The solution turned out to be a lot more involved than I originally suspected - basically the PS compiler had to be refactored to give control of macroexpansion over to special forms (the way it was previously handling macroexpansion was quite naive, but would have been suitable for a single-namespace Lisp).
The changes that have been going on in the recent while are starting to make Parenscript almost look like a "real Lisp compiler."
Vladimir
On Wed, May 6, 2009 at 2:00 PM, Daniel Gackle <danielgackle@gmail.com> wrote:
Here's one thing that the latest PS breaks in our code.
(ps (list 1 2 3)) => "[1,2,3];"
(ps (let ((list nil)) (setf list (list 1 2 3)))) => "var list5180 = null; list5180 = list5180(1, 2, 3);"
I think this behavior is incorrect. It is quite surprising for a lexical variable to shadow a macro definition.
Daniel
_______________________________________________ parenscript-devel mailing list parenscript-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel
_______________________________________________ parenscript-devel mailing list parenscript-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel
Start with a clean Lisp image or set *ps-macro-toplevel* to an empty hash table - that error is due to the fact that the PS internal macro representations have changed. Vladimir On Sat, May 9, 2009 at 8:40 PM, Daniel Gackle <danielgackle@gmail.com> wrote:
Given the recent bug fixes, I tried to upgrade PS again. I'm now encountering an error on trying to expand a simple loop form. The error appears in commit 5a69278.
(ps-macroexpand '(loop for x from 1 below 10 do (bah)))
=> (NIL . #<Anonymous Function #x300042319FEF>) is not of type (OR SYMBOL FUNCTION), and can't be FUNCALLed or APPLYed [Condition of type TYPE-ERROR]
Daniel
On Sat, May 9, 2009 at 5:29 PM, Vladimir Sedach <vsedach@gmail.com> wrote:
Just pushed out a patch that fixes this. The solution turned out to be a lot more involved than I originally suspected - basically the PS compiler had to be refactored to give control of macroexpansion over to special forms (the way it was previously handling macroexpansion was quite naive, but would have been suitable for a single-namespace Lisp).
The changes that have been going on in the recent while are starting to make Parenscript almost look like a "real Lisp compiler."
Vladimir
On Wed, May 6, 2009 at 2:00 PM, Daniel Gackle <danielgackle@gmail.com> wrote:
Here's one thing that the latest PS breaks in our code.
(ps (list 1 2 3)) => "[1,2,3];"
(ps (let ((list nil)) (setf list (list 1 2 3)))) => "var list5180 = null; list5180 = list5180(1, 2, 3);"
I think this behavior is incorrect. It is quite surprising for a lexical variable to shadow a macro definition.
Daniel
_______________________________________________ parenscript-devel mailing list parenscript-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel
_______________________________________________ parenscript-devel mailing list parenscript-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel
_______________________________________________ parenscript-devel mailing list parenscript-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel
Oops sorry, I must have neglected to clean some fasls. On Sat, May 9, 2009 at 9:04 PM, Vladimir Sedach <vsedach@gmail.com> wrote:
Start with a clean Lisp image or set *ps-macro-toplevel* to an empty hash table - that error is due to the fact that the PS internal macro representations have changed.
Vladimir
On Sat, May 9, 2009 at 8:40 PM, Daniel Gackle <danielgackle@gmail.com> wrote:
Given the recent bug fixes, I tried to upgrade PS again. I'm now encountering an error on trying to expand a simple loop form. The error appears in commit 5a69278.
(ps-macroexpand '(loop for x from 1 below 10 do (bah)))
=> (NIL . #<Anonymous Function #x300042319FEF>) is not of type (OR SYMBOL FUNCTION), and can't be FUNCALLed or APPLYed [Condition of type TYPE-ERROR]
Daniel
On Sat, May 9, 2009 at 5:29 PM, Vladimir Sedach <vsedach@gmail.com> wrote:
Just pushed out a patch that fixes this. The solution turned out to be a lot more involved than I originally suspected - basically the PS compiler had to be refactored to give control of macroexpansion over to special forms (the way it was previously handling macroexpansion was quite naive, but would have been suitable for a single-namespace Lisp).
The changes that have been going on in the recent while are starting to make Parenscript almost look like a "real Lisp compiler."
Vladimir
On Wed, May 6, 2009 at 2:00 PM, Daniel Gackle <danielgackle@gmail.com> wrote:
Here's one thing that the latest PS breaks in our code.
(ps (list 1 2 3)) => "[1,2,3];"
(ps (let ((list nil)) (setf list (list 1 2 3)))) => "var list5180 = null; list5180 = list5180(1, 2, 3);"
I think this behavior is incorrect. It is quite surprising for a
lexical
variable to shadow a macro definition.
Daniel
_______________________________________________ parenscript-devel mailing list parenscript-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel
_______________________________________________ parenscript-devel mailing list parenscript-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel
_______________________________________________ parenscript-devel mailing list parenscript-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel
_______________________________________________ parenscript-devel mailing list parenscript-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel
participants (2)
-
Daniel Gackle -
Vladimir Sedach