armedbear-devel

armedbear-devel@common-lisp.net

1 participants
1200 discussions

[armedbear-devel] compile doesn't return correct values for warnings-p and failure-p
by Stas Boukarev 11 Aug '12

11 Aug '12

(compile nil '(lambda () (char=))) ; Caught WARNING: ; Wrong number of arguments for CHAR=. ; Compilation unit finished ; Caught 1 WARNING condition #<FUNCTION (LAMBDA ()) {456E3EB3}> NIL NIL while it should be T, T. jvm-compile, which handles this doesn't succeed because the underlying with-compilation-unit eats all the warnings and errors so that jvm-compile's handler-bind doesn't get a piece of them. Binding *resignal-compiler-warnings* to T inside jvm-compile resolves this. -- With best regards, Stas.

1 0

[armedbear-devel] char= compilation warning
by Stas Boukarev 11 Aug '12

11 Aug '12

(compile (defun test (a b c) (if (char= a b c) t))) ; Caught WARNING: ; Wrong number of arguments for CHAR= (expected 2, but received 3). The error is caused by p2-test-char= in compiler-pass2.lisp:2667 calling (check-arg-count form 2). -- With best regards, Stas.

1 1

[armedbear-devel] [Patch]: Which allows drakma to run.
by Stas Boukarev 11 Aug '12

11 Aug '12

Drakma uses its own type OCTET as element-type for streams it's defined as (deftype octet () '(unsigned-byte 8)) But read-byte and friends don't recognize type because it's not expanded before-hand. So, what I did is to apply EXPAND-DEFTYPE beforehand. Perhaps UPGRADED-ELEMENT-TYPE would've been better, or something entirely different in an another place or more places. But it allows me to use drakma successfully. So, one can do (push :drakma-no-ssl *features*) ;; I can't load CFFI needed for cl+ssl (asdf:load-system :drakma) (drakma:http-request "http://bing.com") => fancy results On another note, I couldn't get usocket to work on Debian until I discovered that I need to either do echo 0 > /proc/sys/net/ipv6/bindv6only or launch java with -Djava.net.preferIPv4Stack=true Otherwise usocket returns a very cryptic error SOCKET-TYPE-NOT-SUPPORTED-ERROR which is in reality "java.net.SocketException: Network is unreachable". -- With best regards, Stas.

2 2

[armedbear-devel] MACROLET environment issue in DO?
by Theam Yong Chew 11 Aug '12

11 Aug '12

Hi devs, I've just stumbled upon something that looks buggy. In the following form, I am particularly focusing on the end-test-form in the do macro. (macrolet ((foo () '(print "macrolet foo"))) (foo) (do ((p 0 (1+ p))) ((foo)) (print "LOOP"))) Prints "macrolet foo" "macrolet foo" then returns ==> NIL Looks fine, but if I wrap that in a function, (defun use-foo () (macrolet ((foo () '(print "macrolet foo"))) (foo) (do ((p 0 (1+ p))) ((foo)) (print "LOOP")))) (use-foo) Prints "macrolet test" then raises an error: The function FOO is undefined. [Condition of type UNDEFINED-FUNCTION] Restarts: 0: [CONTINUE] Try again. 1: [USE-VALUE] Specify a function to call instead. 2: [RETURN-VALUE] Return one or more values from the call to FOO. 3: [RETRY] Retry SLIME interactive evaluation request. 4: [*ABORT] Return to SLIME's top level. 5: [ABORT] Abort thread. Backtrace: 0: (#<FUNCTION {1879842}> #<UNDEFINED-FUNCTION {1426780}> #<FUNCTION {1879842}>) 1: (APPLY #<FUNCTION {1879842}> (#<UNDEFINED-FUNCTION {1426780}> #<FUNCTION {1879842}>)) 2: (SYSTEM::RUN-HOOK SYSTEM::*INVOKE-DEBUGGER-HOOK* #<UNDEFINED-FUNCTION {1426780}> #<FUNCTION {1879842}>) 3: (INVOKE-DEBUGGER #<UNDEFINED-FUNCTION {1426780}>) 4: (ERROR #<UNDEFINED-FUNCTION {1426780}>) 5: (SYSTEM:UNDEFINED-FUNCTION-CALLED FOO NIL) 6: (FOO) Locals: "??" = "??" 7: (USE-FOO) Locals: "??" = "??" 8: (SYSTEM::%EVAL (USE-FOO)) Locals: "??" = "??" 9: (EVAL (USE-FOO)) Locals: "??" = "??" .... If I define a function with same name as the local macro, (defun test () (print "function test")) (defun use-test () (macrolet ((test () '(print "macrolet test"))) (test) (do ((p 0 (1+ p))) ((test)) (print "LOOP")))) (use-test) Prints "macrolet test" "function test" then returns ==> NIL If I manually macroexpand the do form & substitute the expansion back into the function (with some gensym tidy up), (defun macro-expanded-do () (macrolet ((test () '(print "macrolet test"))) (test) (BLOCK NIL (LET ((P 0)) (TAGBODY (GO G4759) G4758 (PRINT "LOOP") (PSETQ P (1+ P)) G4759 (UNLESS (TEST) (GO G4758)) (RETURN-FROM NIL (PROGN))))))) (macro-expanded-do) Prints "macrolet test" "macrolet test" ; No value and returns ==> NIL Is this a bug? I'm running on Windows, using (lisp-implementation-version) ==> "1.0.1-svn-13750-13751" Thanks, Yong.

2 1

[armedbear-devel] breakage in r14066
by James M. Lawrence 09 Aug '12

09 Aug '12

(defstruct foo (data nil)) (defun test () (let ((foo (make-foo))) (symbol-macrolet ((x (foo-data foo))) (setf x 99)) (foo-data foo))) (test) ;=> NIL

2 1

[armedbear-devel] IGNORE declaration in inline function not heeded
by James M. Lawrence 09 Aug '12

09 Aug '12

(declaim (inline foo)) (defun foo (x) (declare (ignore x)) 99) (defun bar (a) (foo a)) compile-file emits ; Caught STYLE-WARNING: ; The variable X is defined but never used. 1.1.0-dev-svn-14066

2 1

[armedbear-devel] Bogus error signaled when compiling bogus RETURN call
by James M. Lawrence 09 Aug '12

09 Aug '12

CL-USER> (lisp-implementation-version) "1.1.0-dev-svn-14066" CL-USER> (defun foo () (return)) FOO CL-USER> (compile 'foo) ; Caught ERROR: ; RETURN-FROM NIL: no block named NIL is currently visible. The value #<EQUAL HASH-TABLE 0 entries, 11 buckets {AE9DCF}> is not of type STRUCTURE-OBJECT. [Condition of type TYPE-ERROR] 0: (#<FUNCTION {1FA8BA4}> #<TYPE-ERROR {1E5E70E}> #<FUNCTION {1FA8BA4}>) 1: (APPLY #<FUNCTION {1FA8BA4}> (#<TYPE-ERROR {1E5E70E}> #<FUNCTION {1FA8BA4}>)) 2: (SYSTEM::RUN-HOOK SYSTEM::*INVOKE-DEBUGGER-HOOK* #<TYPE-ERROR {1E5E70E}> #<FUNCTION {1FA8BA4}>) 3: (INVOKE-DEBUGGER #<TYPE-ERROR {1E5E70E}>) 4: org.armedbear.lisp.Lisp.error(Lisp.java:382) 5: org.armedbear.lisp.Lisp.type_error(Lisp.java:423) 6: org.armedbear.lisp.LispObject.getSlotValue(LispObject.java:1241) 7: org.armedbear.lisp.compiler_pass2_389.execute(compiler-pass2.lisp:7453) 8: org.armedbear.lisp.compiler_pass2_375.execute(compiler-pass2.lisp:7408) 9: org.armedbear.lisp.LispThread.execute(LispThread.java:628)

1 0

[armedbear-devel] destroy-thread and unwind-protect
by James Lawrence 08 Aug '12

08 Aug '12

(defun test-cleanup () (let* ((cleanedp nil) (thread (threads:make-thread (lambda () (unwind-protect (sleep 999) (setf cleanedp t)))))) (sleep 0.5) (threads:destroy-thread thread) (sleep 0.5) cleanedp)) (test-cleanup) ABCL-1.0.0 and ABCL-0.27.0 return NIL. Other implementations I tested all return T -- Allegro, Clozure, LispWorks, SBCL (changing "threads" to "bordeaux-threads"). Of course this isn't a bug since ABCL is free to do what it wishes. Does ABCL wish that?

5 5

[armedbear-devel] abcl-1.1.0 engineering scrum @ 2100 CET (1900 GMT) on #abcl
by Mark Evenson 06 Aug '12

06 Aug '12

We will be holding the weekly abcl-1.1.0 engineering scrum @ 2100 CET on #abcl. Anyone with questions, priorities, patches, welcome to join. The meeting will probably be sparsely attended, so I will just work on recording the [current status as I see it on the Armed Bear Trac Wiki][1]. We can move to audio/video if there is a need. [1]: http://trac.common-lisp.net/armedbear/wiki/releng/abcl-1.1.0 -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."

1 0

[armedbear-devel] Unqualified FIND-PACKAGE call in ABCL compiled files
by Vladimir Sedach 06 Aug '12

06 Aug '12

Hi, Currently, ABCL doesn't qualify the package of FIND-PACKAGE when it serializes a package in a compiled file. This causes a bug when loading a compiled file that has code (for example, the EOS:TEST macro from the EOS unit-testing library) that serializes a package, and occurs after an IN-PACKAGE into a package that either shadows or does not import FIND-PACKAGE from COMMON-LISP. I didn't work out a minimum test case, but the problem can be observed when trying to load the Parenscript test suite: (ql:quickload :parenscript.test) Here's the fix: Index: src/org/armedbear/lisp/compiler-pass2.lisp =================================================================== --- src/org/armedbear/lisp/compiler-pass2.lisp (revision 14055) +++ src/org/armedbear/lisp/compiler-pass2.lisp (working copy) @@ -1191,7 +1191,7 @@ (defun serialize-package (pkg) "Generate code to restore a serialized package." - (emit 'ldc (pool-string (concatenate 'string "#.(FIND-PACKAGE \"" + (emit 'ldc (pool-string (concatenate 'string "#.(CL:FIND-PACKAGE \"" (package-name pkg) "\")"))) (emit-invokestatic +lisp+ "readObjectFromString" (list +java-string+) +lisp-object+)) Index: src/org/armedbear/lisp/Package.java =================================================================== --- src/org/armedbear/lisp/Package.java (revision 14055) +++ src/org/armedbear/lisp/Package.java (working copy) @@ -865,7 +865,7 @@ public String printObject() { if (_PRINT_FASL_.symbolValue() != NIL && name != null) { - StringBuilder sb = new StringBuilder("#.(FIND-PACKAGE \""); + StringBuilder sb = new StringBuilder("#.(CL:FIND-PACKAGE \""); sb.append(name); sb.append("\")"); return sb.toString(); Happy hacking, Vladimir

2 1

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

armedbear-devel