gmane.editors.j.devel

Re: EXTERNAL-FORMAT support r11409 -branches/open-external-format/

Erik Huelsmann — 2008-12-01T13:35:28

Resending to the right list: On Mon, Dec 1, 2008 at 9:59 AM, Erik Huelsmann gmail.com> wrote: ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: :EXTERNAL-FORMAT support for OPEN

Hideo at Yokohama — 2008-11-30T15:29:07

Hi, Thanks for your effort incorporating my patch. I tried out the branch, and with LANG set to "ja_JP.UTF-8", abcl successfuly printed Japanese utf-8 text to the tty. Great! So, this worked: (with-open-file (s "test-ja.txt" :encoding "utf-8") (format t "~a~%" (read-line s))) However, when I didn't pass any argument for :external-format, it went into an infinite loop. This happens only when the file has non-ascii chars, and I don't specify external-format. So, this didn't work: (with-open-file (s "test-ja.txt") (format t "~a~%" (read-line s))) When I specify "ISO-8859-1" as :external-format, the program will produce output. The characters get gobbled but that is as expected. I don't have time to investigate it right now. Just a quick report. Regards, Hideo. On Sun, 30 Nov 2008 22:52:41 +0900, Erik Huelsmann gmail.com> wrote:

Re: :EXTERNAL-FORMAT support for OPEN

Erik Huelsmann — 2008-11-30T13:52:41

Hi Hideo, Thanks for the patch! I started to commit it bit-by-bit. See my comments below. On Sat, Nov 29, 2008 at 5:34 PM, Hideo at Yokohama gmail.com> wrote: Yes. I didn't want to enable the encodings while the basic behaviour wasn't stable. Your patch fixed that, so I also committed this bit. That's completely correct. However, I intended the external-format to follow the format as used by the argument list to MAKE-EXTERNAL-FORMAT for FLEXI-STREAMS (see http://weitz.de/flexi-streams/). I guess there's no real harm in doing it this way. I'm in the process of adding that, but committed your contribution with the FIXMEs in the mean time. I haven't committed the sockets bit: I'd like to have a single implementation to do the eol-style recoding and a maximum of 2 implementations for character conversion (one for random-access files and one for streams). I just committed a fix to your changes which at least should use the system default locale for reading off the input stream, so,

Re: :EXTERNAL-FORMAT support for OPEN

Hideo at Yokohama — 2008-11-30T05:02:46

Erik, I ran the ansi tests with my modifications that I sent as a gzipped svn diff. FILE-POSITION.5 passed, but OPEN.66 fails. open.66 requires that the lisp function open can take an open stream object as the first argument for the open function. The first argument is typically a path name, rather than a stream. open.lisp in abcl doesn't look like it is prepared to handle such situation. For me, enough is working for now. Regards, Hideo. On Sun, 23 Nov 2008 20:33:09 +0900, Erik Huelsmann gmail.com> wrote:

Re: IKVM Patch progress

2008-11-29T23:49:08

Oops I forgot to mention for running on win32/64 I had to modify IkvmSite.java slightly Adding: if (res == null) { if (Utilities.isPlatformWindows) { String revSlash = filename.replace("\\", "/"); //someone stripped the "/" ? if (revSlash.charAt(0) != '/') revSlash = "/" + revSlash; if (!revSlash.equals(filename)) { res = Lisp.class.getResource(revSlash); } } to: public static File ikvmFile(String filename) throws IOException I attaced the new j/src/org/armedbear/lisp/IkvmSite.java ----- Original Message ----- From: users.sourceforge.net> To: "Ville Voutilainen" gmail.com> Cc: "Mark Evenson" panix.com>; "kino coursey" gmail.com>; lists.sourceforge.net> Sent: Saturday, November 29, 2008 3:41 PM Subject: Re: IKVM Patch progress ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's

Re: IKVM Patch progress

2008-11-29T23:41:00

The only test that i *have* to comment out is MISC.512 That single test FIND-SYMBOL.ERROR.2 on my machine might need a half hour to run.!:(. I have been writing it off as a mono gc headroom bug however, here is the latest results I am getting on .NET 2.0 on WinXP: notice the time is just as bad as on on mono. 57 out of 21701 total tests failed: (DEFINE-SETF-EXPANDER.1 DEFINE-SETF-EXPANDER.6 DEFINE-SETF-EXPANDER.7 MULTIPLE-VALUE-SETQ.5 MULTIPLE-VALUE-SETQ.8 REINITIALIZE-INSTANCE.ERROR.1 SHARED-INITIALIZE.ERROR.4 CHANGE-CLASS.1.11 CHANGE-CLASS.ERROR.4 MAKE-INSTANCE.ERROR.3 MAKE-INSTANCE.ERROR.4 DEFGENERIC.ERROR.20 DEFGENERIC.ERROR.21 DEFGENERIC.30 CALL-NEXT-METHOD.ERROR.1 CALL-NEXT-METHOD.ERROR.2 DEFMETHOD.ERROR.14 DEFMETHOD.ERROR.15 RESTART-CASE.29 RESTART-CASE.30 RESTART-CASE.31 MAKE-CONDITION.3 MAKE-CONDITION.4 DELETE-PACKAGE.5 DELETE-PACKAGE.6 DO-ALL-SYMBOLS.6 DO-ALL-SYMBOLS.9 DO-ALL-SYMBOLS.12 ABS.14 ABS.15 ABS.16 ABS.17 ABS.18 ABS.19 ABS.20 ABS.21 MAP.48 MAP.SPECIALIZED-STRING.3 TYPE-OF.1 ENSURE-DIR

Re: IKVM Patch progress

Ville Voutilainen — 2008-11-29T22:25:19

I applied the patch, rebuilt and run the ikvmc. The tests don't seem to run for me, the tests hang at FIND-SYMBOL.ERROR.2 I noticed that my environment differs from dmiles' env so that I have jdk 1.5, dmiles seems to have jdk 1.7. Could someone verify whether the similar hang occurs for anyone else? We currently support 1.5, so it would be important to get these modifications to work there, if at all possible. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: [armedbear-cvs] r11401 -branches/open-external-format/src/org/armedbear/lisp

Erik Huelsmann — 2008-11-29T20:20:02

ah. right. sorry. I'll have another look at the code again. Bye, Erik. On Sat, Nov 29, 2008 at 9:05 PM, Ville Voutilainen gmail.com> wrote: ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

[armedbear-cvs] r11401 -branches/open-external-format/src/org/armedbear/lisp

Ville Voutilainen — 2008-11-29T20:05:24

On Sat, Nov 29, 2008 at 8:42 PM, Erik Huelsmann common-lisp.net> wrote: This code seems dubious, we're in an else-part of if (eolStyle == EolStyle.CRLF) { Thus the test that you added is always false, is it not? There's no need to write CRs in this case, because the if-part takes care of that and this else part should not. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Documentation on embedding ABCL

Brad Garton — 2008-11-29T18:18:54

I've seen questions like this come up before on this list, and to be honest it was difficult for me to figure at first, too (I'm not a highly-skilled java or lisp coder!). I haven't taken a good look at the example code now distributed with the new abcl, nor have I been able to find much on the wiki site, so this may be redundant. And it also may not be the best way to do this (see the comment a few sentences above about my java/lisp skills) -- any improvement-advice would be appreciated! I found my answer to questions I had about calling abcl from java by scanning the mailing list archives + some good help from list- contributors (esp. Alex Mizrahi), so I thought I'd post a few quick lines here. I *think* these are correct; they're digested from a denser patch of code. To set up a basic call to the interpreter from java: Interpreter interpreter; String lispbuf, evalstring; interpreter = Interpreter.createInstance(); lispbuf = new String(); // fill lispbuf with the LISP to be int

abcl testimonial

Brad Garton — 2008-11-29T18:16:13

Hey abcl-ers -- I've been meaning to write this for quite awhile now, but one thing kept getting in front of another, blah blah blah... This is one of the 'testimonial' missives (with links to some code even!). I used abcl (older, v. 0.0.10) as the interpreter underlying an object for a music/media graphical programming environment called Max/MSP (http://www.cycling74.com/products/max5). abcl allowed me to imbed an entire LISP interpreter inside the Max/MSP package, and it has worked extremely well. Max/MSP is designed primarily for real-time/interactive applications, used a lot by computer music academics, DJs/VJs, bizarro jazz performers, audio plugin developers, etc. With the imbedded abcl object, I am able to schedule LISP functions and work interactively with complex processes -- the abcl interpreter has been rock-solid! The source for the object is here (scroll down to the bottom of the page): http://music.columbia.edu/~brad/maxlispj/ And I also used it extensively in a large application built

Re: :EXTERNAL-FORMAT support for OPEN

Hideo at Yokohama — 2008-11-29T16:40:48

Excuse me. I was still editing when I sent it out by accident. I was trying to say, didn't work. I tried to compile with "ant abcl.wrapper" but it reproduced the compile error with the character duplication issue. Compiling with sbcl using build-abcl.lisp did the job for me. Bye. On Sun, 30 Nov 2008 01:34:46 +0900, Hideo at Yokohama gmail.com> wrote:

Re: :EXTERNAL-FORMAT support for OPEN

Hideo at Yokohama — 2008-11-29T16:34:46

Hi again. I got things starting to work. Attached is the gzipped diff. What I did : (1) Went through the make-file-stream on the lisp side, since it seemed to be half way done. I made it pass the external-format parameter. (2) Changed the way the java side code handles the external-format parameter. I'm not quite sure how abcl represent objects of various types, I kind of guessed that if a string is passed, the LispObject would be of type AbstractString, and you could get the value by getStringValue. It works now so you can say :external-format "utf-8" . I still don't know handle symbols or keyword symbols. (3) Changed socket_stream so that it will accept an external-format argument. (4) compiled it with sbcl. ant What I got working: (1) Invoke abcl from slime, read a file encoded in utf-8, write it to the slime terminal. The following code worked: (with-open-file (s "test-ja.txt" :external-format "utf-8") (format t "~a~%" (read-line s))) What I haven't go

Re: IKVM Patch progress

2008-11-28T03:37:33

Patch URL: https://sourceforge.net/mailarchive/attachment.php?list_name=armedbear-j-devel&message_id=87F8482FD00C43A684490C12668794C9%40lotus&counter=1 Here are some results of the IKVM [root< at >titan ansi-tests]# /usr/lib/jvm/jdk1.7.0/bin/java -cp ../build/classes/ org.armedbear.lisp.Main --eval '(load "doit.lsp")' Armed Bear Common Lisp 0.0.11.1 (built Thu Nov 27 2008 14:47:50 PST) Java 1.7.0-ea Sun Microsystems Inc. Java HotSpot(TM) 64-Bit Server VM Low-level initialization completed in 0.298 seconds. Startup completed in 1.223 seconds. .......... 48 out of 21701 total tests failed: (DEFINE-SETF-EXPANDER.1 DEFINE-SETF-EXPANDER.6 DEFINE-SETF-EXPANDER.7 MULTIPLE-VALUE-SETQ.5 MULTIPLE-VALUE-SETQ.8 REINITIALIZE-INSTANCE.ERROR.1 SHARED-INITIALIZE.ERROR.4 CHANGE-CLASS.1.11 CHANGE-CLASS.ERROR.4 MAKE-INSTANCE.ERROR.3 MAKE-INSTANCE.ERROR.4 DEFGENERIC.ERROR.20 DEFGENERIC.ERROR.21 DEFGENERIC.30 CALL-NEXT-METHOD.ERROR.1 CALL-NEXT-METHOD.ERROR.2 DEFMETHOD.ERROR.14 DEFMETHOD.ERROR.15 RESTART-CASE.29 RESTART-CASE

Re: IKVM Patch progress

2008-11-27T22:27:50

Moved the bulk code to a new File called IkvmSite.java This patch replaces the previous ones Here is the preview of the attached .patch minus the new file Index: src/org/armedbear/lisp/Keyword.java =================================================================== --- src/org/armedbear/lisp/Keyword.java (revision 11399) +++ src/org/armedbear/lisp/Keyword.java (working copy) < at >< at > -35,6 +35,9 < at >< at > public final class Keyword extends Lisp { + public static final Package PACKAGE_KW = + Packages.createPackage("KEYWORD", 1024); + public static final Symbol ABCL = internKeyword("ABCL"), ABORT = internKeyword("ABORT"), Index: src/org/armedbear/lisp/Lisp.java =================================================================== --- src/org/armedbear/lisp/Lisp.java (revision 11399) +++ src/org/armedbear/lisp/Lisp.java (working copy) < at >< at > -79,7 +79,7 < at >< at > // ### nil // Constructing NIL forces the Symbol class to be loaded (since Nil extends // Symbol). - pu

Re: IKVM Patch progress

2008-11-27T20:39:14

Oh to try it out in IKVM ant abcl.dist ikvmc -compressresources dist/abcl.jar mono abcl.exe [12:28] It doesn't look bad, but I want the ikvmfile functions gone from Load.java, and put to a separate file. [12:29] no, I didn't peek. [12:29] V-ille, yeah.. i can also put the stuff to Site.java as well [12:29] i mean remove the Site.java additions to a differnt file [12:30] dmiles_afk: That would be good. Wherever possible, maybe we should consolidate the ikvm-specific stuff into, well, [12:32] ok, yeah goign to do that and fix the CL_PACKAGE thing in Nil.. since its actualy null there [12:34] i'm still trying to figure out the extent of the needs for ikvmFileSafe(...) basically it allows searching of the DLL and then unpacks it to a temp file.. the right thing is to return the filestream ----- Original Message ----- From: users.sourceforge.net> To: "Mark Evenson" panix.com>; "Ville Voutilainen"

Re: Bug

Ville Voutilainen — 2008-11-26T20:48:45

On Wed, Nov 26, 2008 at 10:24 PM, Reinhard Oldenburg math.uni-frankfurt.de> wrote: You can try running the code inside a debugger (jdb, eclipse, netbeans). That should give a more precise idea, but debugging compiled lisp functions with any of the above is a bit painful, because you'd have to manually follow cars and cdrs without seeing the actual lisp code. I would like to have the file, so if you can put it on a web page, please do. I suppose you can mail it to me directly, just remember not to cc the list in that case. :) Can I get the donp2 from somewhere? I had trouble googling it. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Bug

Reinhard Oldenburg — 2008-11-26T20:24:00

I successfully compiled ABCL 0.0.11.1 checked out on the 24th of November. It compiled Maxima 5.16.3 fast als the Maxima works. But then I gave the donp2 package, translated by Ray Toy from Fortran to CL a try and got stuck with the following error: ; Caught STYLE-WARNING: ; The variable IUMIN is defined but never used. java.lang.VerifyError: (class: org/armedbear/lisp/o8opti_2, method: execute signature: ([Lorg/armedbear/lisp/LispObject;)Lorg/armedbear/lisp/LispObject;) Illegal target of jump or branch at java.lang.Class.getDeclaredConstructors0(Native Method) at java.lang.Class.privateGetDeclaredConstructors(Class.java:2389) at java.lang.Class.getConstructor0(Class.java:2699) at java.lang.Class.getConstructor(Class.java:1657) at org.armedbear.lisp.Lisp.loadCompiledFunction(Lisp.java:1147) at org.armedbear.lisp.Lisp.loadCompiledFunction(Lisp.java:1074) at org.armedbear.lisp.o8opti_1.(o8opti.lisp:1) Is there a way to find out the line

Re: ABCL 0.0.11 tagged in SVN

Erik Huelsmann — 2008-11-26T10:09:32

The effect at least is what I intended. Sorry to have been vague about it. After your mails I went to Trac and found that r11356 is indeed the magic number. Bye, Erik. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: ABCL 0.0.11 tagged in SVN

Mark Evenson — 2008-11-26T08:53:02

I wasn't entirely sure of which revision to tag, so I specified r11360 as the version of the repository that certainly contained a good version of ABCL 0.0.11, as it was the first revision post Oct 19 (the day that Erik announced the release), and r11360 made a copy of 'trunk' which I inferred as Erik's implicit sanction as the natural "0.0.11" tag revision. I made a mistake in the comment for the commit which unfortunately I don't think can be edited without a fair amount of pain from what I remember (I think one has to dump the repository, replay all the commits up until the "bad" comment, etc.) r11356 corresponds to the last change in the repository that was underneath the "trunk" which is what I intended by specifying r11360 as the version to tag as 'abcl-0.0.11'. This is what I presumed from looking at the svn logs/Trac timeline but ensured from using r11360 as the peg revision to copy. My logic might be a bit convoluted here, but as far as I know the 'abcl-0-0-11' tag faithfully corr

Re: ABCL 0.0.11 tagged in SVN

Ville Voutilainen — 2008-11-26T07:46:51

That raised my eyebrow.. r11360 is the import of the scripting branch. The commit message says r11360 (well, actually it says r113960 which doesn't exist :), but the tag itself says it's a copy of trunk version r11356. Most puzzling. :) ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/