February 2012 - armedbear-devel - mailman3.common-lisp.net

[armedbear-devel] ASDF: Cannot handle the special form EXTENSIONS:TRULY-THE
by Blake McBride 24 Mar '13

24 Mar '13

Greetings, Trying to load a package with ASDF I ran into the following error: ASDF could not load because Cannot (currently) handle the special form EXTENSIONS:TRULY-THE. #<THREAD "interpreter" {4BC78066}>: Debugger invoked on condition of type SIMPLE-ERROR Cannot (currently) handle the special form TRULY-THE Restarts: ... I traced the problem to ASDF and not the package I was attempting to load. Does anyone know anything about this? (I am running the latest SVN head of ABCL) Thanks. Blake McBride

6 14

[armedbear-devel] Double printing of the prompt
by Blake McBride 12 Aug '12

12 Aug '12

Greetings, The latest version of ABCL is double printing the prompt at times as follows: Armed Bear Common Lisp 1.1.0-dev-svn-13783 Java 1.6.0_29 Apple Inc. Java HotSpot(TM) 64-Bit Server VM Low-level initialization completed in 0.465 seconds. Startup completed in 1.453 seconds. Type ":help" for a list of available commands. CL-USER(1): 3 3 CL-USER(2): 'hello HELLO CL-USER(3): '(a b c) (A B C) CL-USER(4): nil NIL CL-USER(5): CL-USER(5): 77 77 CL-USER(6): (cdr '(hello)) NIL CL-USER(7): nil NIL CL-USER(8): CL-USER(8): 3 3 CL-USER(9):

3 3

[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] Patch with improvements to stable-sort and sort
by Jorge Tavares 07 May '12

07 May '12

Hi, I send in attach a patch with further improvements to sort and stable-sort for sequences other than lists. In short, the patch includes a merge sort for vectors. To allow different types I've written the algorithm using macros and these generate the appropriate code according to the vector type. This way the algorithm is in a single place avoiding duplication of code. The macros also take care of the situation of when no key is present, avoiding the use of unnecessary funcalls. The quicksort algorithm was also refactored in the same way. I've tested the algorithms and they seem to be working correct. Stable sort is now considerably faster since the fix before converted the sequences to a list and used the sort-list function. I've made some benchmarking to verify how fast is sort and stable-sort. The tables with the results are also in a file sent in attach [1]. For stable-sort I've compare the current trunk version with the patched one while for sort I've compared 1.0.1, the trunk and with the patch. For unsorted vectors sort has a speed up of 7.5 from 1.0.1 and this considers only vectors of size 8 to 8192 (1.0.1 hits the worst-case quite fast). For stable-sort the speed up is around 90.2 from vectors of size 8 to 32768. The sort functions become even faster for the nearly sorted vectors. I think the tables clearly show the speed-ups :-) Naturally these benchmarks cannot be used to draw definite conclusions since they lack rigorous testing but I think they can provide some indications. With this patch, I think ABCL gets good performant sorting functions, especially for large vectors. As for lists, I haven't looked at them so probably they can also be improved (but I really don't know). Cheers, Jorge [1] The tables result from the generation of simple-vectors of sizes 8 to 524288 (powers of 2 from 3 to 19) with distinct integer: unsorted, nearly sorted (distances 0, 4 and 16), sorted and reversed sorted. The nearly sorted vectors were constructed by selecting pairs where they would swap with a neighbor at a certain distance. I did 100 runs and timed only the sorting operation. The tables contain the averages of the 100 runs. They were performed in an iMac (2.5GHz i5, 4GB) with Mac OS X 10.7.3. -- http://jorgetavares.com

2 2

[armedbear-devel] compiled errors: debug.assertTrue()
by scarbonell＠ono.com 06 Mar '12

06 Mar '12

Hi, When I try to compile the mcs-core.cl file (attached) with java abcl, I get this error: Failed to get InputStream for 'jar:file:D:/otro/kernel/modules/fmcs.bin!/mcs_core_4.cls'ABCL Debug.assertTrue() assertion failed!java.lang.Error: ABCL Debug.assertTrue() assertion failed! Why is this happening? How can I resolved it? Thanks

3 3

[armedbear-devel] compiling lisp files outside Java and calling it after with abcl
by scarbonell＠ono.com 02 Mar '12

02 Mar '12

Hi,I'm trying to compile a lisp application and I get a lot of errors with java abcl. I thought that if I try to compile this lisp files outside of abcl with other systems, perhaps, I could then call this compiled functions lisp files from ABCL (look like a jar file of lisp compiled files). Is this possible?How can I compiled this files outside ABCL? Some examples?Thanks

5 10

[armedbear-devel] Upcoming abcl-1.1 branch
by Mark Evenson 25 Feb '12

25 Feb '12

On [#abcl]() yesterday, ehu and easye discussed how it would be a good idea to start packing up for shipping abcl-1.1 out the door. [#abcl]: http://webchat.freenode.net/?randomnick=1&channels=abcl&prompt=1&uio=d4 Out of our brief discussion, we'd like to branch the trunk in about a week, stabilize with testing, and then release abcl-1.1.0 on 14-FEB-2012. So, unless there are objections, we'd like to start by asking everyone who was thinking of patching something, to think about getting some time in the next week to get something done. All potential contributions are welcome. If you were going to document something, now's the time! [Rudi added real BibTex-based markup to the Manual][1], so even something as trivial as fixing the references would be appreciated. [1]: http://code.google.com/p/abcl-dynamic-install/source/detail?r=5c7faeeb35408… yers in CONS, Mark -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."

2 2

[armedbear-devel] Patch for ABCL to dynamically load CFFI JNA requirement
by Mark Evenson 24 Feb '12

24 Feb '12

Please consider the referenced patch for inclusion in cffi-0.10.7. Comments towards improvments solicited. Improvements to ABCL CFFI implementation: o autoload the 'jna.jar' dependency if possible (requires abcl-1.1.0-dev) o initial implementation of callbacks Needs ABCL specific tests, but I can load CL+SSL with this patch. [1]: http://slack.net/~evenson/abcl/cffi/abcl-cffi-20120224a.diff Cheers, easyE -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."

1 0

[armedbear-devel] 김영태님이 Abcl님을 Facebook으로 초대하셨습니다
by Young Tae Kim 05 Feb '12

05 Feb '12

안녕하세요, Abcl님! 내 사진, 동영상, 이벤트를 게시할 수 있는 Facebook 프로필을 만들었습니다. 귀하가 저의 프로필을 볼 수 있도록 친구로 추가하고 싶습니다. 그러려면 먼저 귀하가 Facebook에 가입해야 합니다! 회원으로 가입하면 귀하도 프로필을 만들 수 있습니다. 고맙습니다! Young Tae Facebook에 가입하려면 아래 링크를 클릭하세요. http://www.facebook.com/p.php?i=100003305725765&k=AQCtufVGAiIaS35gg9kHoIHhT… 이미 계정을 갖고 계신가요? 계정에 이 이메일 주소를 추가하세요. http://www.facebook.com/n/?merge_accounts.php&e=armedbear-devel%40common-li… ======================================= 메시지가 armedbear-devel(a)common-lisp.xn--net()-7s4ut57e 전송되었습니다. 앞으로 이러한 이메일을 그만 받으려면 아래의 링크를 눌러서 수신을 거부해 주세요. http://www.facebook.com/o.php?k=a23045&u=100003558065807&mid=5989382G5af3e4… Facebook, Inc. Attention: Department 415 P.O Box 10005 Palo Alto CA 94303

1 0

[armedbear-devel] Patch that fixes ticket #187 (Stack Overflow for Worst-case Vector Sort)
by Jorge Tavares 05 Feb '12

05 Feb '12

Hi, I'm sending a patch that solves the issue reported in ticket #187. The quicksort function picks the pivot by selecting a midpoint and also sorts the smaller partition first. These are enough to avoid the stack overflow problem as reported. I've performed some tests and it looks it is correct. Below I include a sample session, including the test case in the ticket: idesk:abcl jast$ ./abcl Armed Bear Common Lisp 1.1.0-dev-svn-13848M Java 1.6.0_29 Apple Inc. Java HotSpot(TM) 64-Bit Server VM Low-level initialization completed in 0.307 seconds. Startup completed in 0.945 seconds. Loading /Users/jast/.abclrc completed in 4.528 seconds. Type ":help" for a list of available commands. CL-USER(1): (ql:quickload "alexandria") To load "alexandria": Load 1 ASDF system: alexandria ; Loading "alexandria" ("alexandria") CL-USER(2): (defparameter *sorted* (make-array 1000000 :initial-contents (alexandria:iota 1000000))) *SORTED* CL-USER(3): (defparameter *numbers* (alexandria:shuffle (copy-seq *sorted*))) *NUMBERS* CL-USER(4): (equalp *sorted* *numbers*) NIL CL-USER(5): (time (progn (setf *numbers* (sort *numbers* #'<)) nil)) 2.157 seconds real time 3 cons cells NIL CL-USER(6): (equalp *sorted* *numbers*) T CL-USER(7): (let ((a (map-into (make-array 10000) (let ((x 0)) (lambda () (incf x)))))) (sort a #'<) nil) NIL CL-USER(8): (time (progn (setf *numbers* (sort *numbers* #'<)) nil)) 1.074 seconds real time 3 cons cells NIL CL-USER(9): With these changes, sort seems to be a little faster (for vectors) although I was not worried with optimizations. In ABCL 1.0.1, in my machine, sorting 1000000 random integers takes around 10s on average while now it takes 2s. However, I must point out I didn't do any serious and proper benchmarking, just some runs. I will be happy to answer any questions if necessary. Cheers, Jorge Tavares -- http://jorgetavares.com

2 2