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

[armedbear-devel] Compiled type declaration and type checks (http://abcl.org/trac/ticket/323)
by Theam Yong Chew 08 Apr '15

08 Apr '15

Just an FYI (is trac open to edits for the public? I guess not), I managed to narrow down the problem here to this: http://abcl.org/trac/ticket/323 (defun test (stream) (declare ;; (type stream stream) (stream stream) ;; (optimize speed (safety 1) (debug 0)) ) (write-byte 1 stream)) (flex:with-output-to-sequence (s) (test s)) If you evaluate test (C-x C-e), the second form works. If test is compiled, (C-c C-c), then the second form fails. Some sort of type check error? The flexi-streams *is* a stream: (block a (flex:with-output-to-sequence (s) (return-from a (list (type-of s) (typep s 'stream))))) => (FLEXI-STREAMS::VECTOR-OUTPUT-STREAM T) Yong

2 1

[armedbear-devel] Email version of list?
by Steve Nunez 08 Apr '15

08 Apr '15

Hope this makes it through. It's not easy trying to find out where to sign up for the ABCL list. Does anyone know where this is done? This silly gmane web interface is complaining about lines longer than 80 characters.

3 3

[armedbear-devel] Looking for someone(s) for a project
by Alan Ruttenberg 08 Apr '15

08 Apr '15

STOP Reading here if you absolutely no interest in or bandwidth for doing work for hire. --- The project will probably *not* be implemented in lisp, but many of you can, I know, program in any language. We do intend that the server side be based on the JVM so that there is flexibility in choice of language used. Whether or not lisp can be used, and to what extent, will depend on how persuasive I am in convincing the group that owns the existing system that allowing development in a number of different languages is a Good Thing. One project is to help start moving a legacy windows(PowerBuilder, .net, SQL server)-based app, typically accessed through Citrix, towards a standards-based web platform (e.g. firefox on the client). The strategy is to start adding new functionality using web standards and have them interoperate with the existing system. As time permits they will move old functionality over. The system is an electronic dental record and so issues that will have to be dealt with are having secure connections, coordinated timeouts of screens (after any activity in either the app window or web app the timer is reset, after the last activity windows need be locked). This means there has to be a system by which windows on both sides of the application are coordinated, with failsafes in the case of losing network connectivity. Other issues to address are: - Coordination with Citrix - ideally the VPN enabling the web app is carried by the single Citrix connection. - Storage backend. There is interest in having new storage being triple-store-based. - Possibility of using AWS Perhaps by the same person, or by another, we will want to start building functionality on the web side. In this case I would want someone with solid experience using a modern javascript toolkit and either direct design experience or who works with a designer. I would expect to see a portfolio of previous work. Telecommuting will be an option, though there may be need for a visit to Buffalo at some point in the project. Schedule: It would be beneficial to aim for a quick proof of concept, then move into development. The next milestone will be deployment in a new clinic in a matter of months (legacy plus whatever new functionality we can add during that time). How much effort is spent in the legacy system versus in the new standards-based system depends on how quickly we can show that the strategy works and can be a reasonable basis for development. A likely project is to extend the current system with the ability to record medical history, with this interaction taking place on the web side. The project, all told, could exercise quite a spectrum of skills. Depending on the people working on the project, and the needs of the future clinic denizens we will obviously need to be strategic in setting prioritizes and crafting a plan. The ideal candidate would have experience and previous success in planning implementation so they can be a useful participant in our project strategizing. If you are perhaps interested in this work, please contact me offline at alanruttenberg(a)gmail.com with the understanding that while I am tasked getting some leads on skilled programmers I am not the person responsible for hiring. But I'll have influence and will have some level of ongoing relation to the project. Regards, Alan ps. I think the project will be fun. YMMV.

1 0

[armedbear-devel] Building from trunk broken?
by Alan Ruttenberg 08 Apr '15

08 Apr '15

I'm getting boatloads of messages of the sort: [java] Failed to require CLOS because 'The function (SETF VALUES) is undefined.' [java] ; SYSTEM:%MAKE-CONDITION triggers autoloading of org.armedbear.lisp.make_condition ... [java] ; Autoloaded org.armedbear.lisp.make_condition (0.0030 seconds) [java] Failed to require PRINT-OBJECT because 'Don't know how to REQUIRE CLOS.' [java] ; COMMON-LISP:INVOKE-DEBUGGER triggers autoloading of debug ... [java] ; Autoloaded debug (0.013 seconds) [java] ; COMMON-LISP:FRESH-LINE triggers autoloading of pprint ... [java] ; Loading /Users/alanr/repos/abcl/src/org/armedbear/lisp/print.lisp ... [java] ; Loaded /Users/alanr/repos/abcl/src/org/armedbear/lisp/print.lisp (0.053 seconds) [java] ; Loading /Users/alanr/repos/abcl/src/org/armedbear/lisp/print-object.lisp ... [java] ; Loading /Users/alanr/repos/abcl/src/org/armedbear/lisp/clos.lisp ... [java] Failed to require CLOS because 'The function (SETF VALUES) is undefined.' [java] Failed to require PRINT-OBJECT because 'Don't know how to REQUIRE CLOS.'

1 1

[armedbear-devel] Why field LispStatckFrame.UNAVAILABLE_ARG is not static ?
by Dmitry Nadezhin 08 Apr '15

08 Apr '15

Memory profiling of ABCL shows that the classes with largest allocation count are LispStackFrame and LispStackFrame.UnavailableArgument . It seems to me that LispStatckFrame.UNAVAILABLE_ARG can be a singleton object, can't it ? =================================================================== --- LispStackFrame.java (revision 14571) +++ LispStackFrame.java (working copy) @@ -44,7 +44,7 @@ private final LispObject third; private final LispObject[] args; - private final class UnavailableArgument extends LispObject + private static final class UnavailableArgument extends LispObject { public UnavailableArgument () { } @Override @@ -53,7 +53,7 @@ } } - private final LispObject UNAVAILABLE_ARG = new UnavailableArgument(); + private static final LispObject UNAVAILABLE_ARG = new UnavailableArgument(); public LispStackFrame(LispObject operator) { ====================================================================

2 1

[armedbear-devel] Fwd: One-jar
by Alan Ruttenberg 08 Apr '15

08 Apr '15

Howdy, I was wondering whether anyone had tried to use this with ABCL. Looks like a nice packaging solution but I'm not sure how it would interact with ABCL's classloader (yes I can try :) ) -Alan ---------- Forwarded message ---------- From: Blaisure, Jonathan <jcblaisu(a)buffalo.edu> Date: Wed, Nov 13, 2013 at 9:12 AM Subject: One-jar To: Alan Ruttenberg <alanruttenberg(a)gmail.com> http://one-jar.sourceforge.net/ Might be useful. -- Regards, Jonathan Blaisure, MS ---------------------- *Database Administrator* *Institute for Health care Informatics* 716.878.2996

3 3

[armedbear-devel] Lazy allocation of LispStackFrames
by Dmitry Nadezhin 08 Apr '15

08 Apr '15

LispStackFrame object are allocated on every LispThread.execute(...) . However, they are seldom (to inspect stack trace). This patch delays allocation of LispStackFrame objects until they are requested. Raw information about stack frames is stored in stack. Stack is an Object[] array ( more precisely a list of 8 MB Object[] arrays ).

3 3

[armedbear-devel] Pdfdoclet - Javadoc for ABCL in PDF format
by Sean Champ 08 Apr '15

08 Apr '15

Hello! In brief: Attached is a shell script for running pdfdoclet[6], a Javadoc doclet component that generates PDF format documentation, instead of HTML, when used with the javadoc tool. For sake of convenience, I was going to send the PDF format ABCL javadoc generated by pdfdoclet, but the resulting PDF file is 13.3 megs in size. Maybe the little utility shell script, as attached, could be enough. The PDF javadoc could be useful, i think, in the interest of studying ABCL's Java API - perhaps more intermediately, for modeling ABCL's Java API in UML - for instance, using Modelio[1], Argo[2], or Papyrus[3], on the laptop, and Cubetto Mobile[4] on the tablet. As an additional plug: As far as the PDF format itself, and PDF's annotation support, typically on mobile, i tend to prefer Goodreader[5]. It offers good support for PDF annotations and so on. PDFDoclet can also be used from Ant. I'm not very well familiar with Ant, offhand. Of course, I notice that Ant is used in ABCL - will take a look... Cheers! [1] http://www.modeliosoft.com/ [2] http://argouml.tigris.org/ [3] http://www.eclipse.org/papyrus/ [4] http://www.semture.de/en/cubetto-mobile-en/overview [5] http://www.goodiware.com/goodreader.html [6] http://pdfdoclet.sourceforge.net/ <http://pdfdoclet.sourceforge.net/><http://www.semture.de/en/cubetto-mobile-en/overview><http://www.eclipse.org/papyrus/>

2 1

[armedbear-devel] mvn snapshots
by Cyrus Harmon 08 Apr '15

08 Apr '15

The attached patch adds support for maven snapshot versions with the following defsystem component syntax: (:mvn "org.openscience.cdk/cdk-bundle" :version "1.5.6" :snapshot t) I’m open to other suggestions to solving this, but this works for my immediate needs. thanks, Cyrus

3 8

[armedbear-devel] Re: What are you working on?
by Rolando Pereira 08 Apr '15

08 Apr '15

ubikation _ <ubikation-Re5JQEeQqe8AvxtiuMwx3w(a)public.gmane.org> writes: > What projects are people working on using ABCL? If they are > open-source, would you mind sharing a link to the repository? Since the lack of answers is slightly disappointing I guess I can post something regarding an experiment I did an afternoon a couple months ago which tried to link Jython to ABCL. Since it was an experiment I didn't put the source code online. Also currently the source is basically just a "*slime-scratch*" buffer saved to a file which means that the code is a hairy mess. But here's an example to show how to use it: > (add-to-classpath "/home/rolando/src/jython/jython-standalone-2.5.3.jar") > (defvar *interpreter* (jss:new 'org.python.util.PythonInterpreter)) > > (#"exec" *interpreter* "import feedparser") > > (defvar *rss-parse* (gethash "parse" > (py-unwrap (#"getDict" > (gethash "feedparser" > (py-unwrap (#"getLocals" *interpreter*))))))) > > ;; Get the ABCL mailing list RSS > (defvar *rss* (pycall *rss-parse* "http://rss.gmane.org/gmane.lisp.armedbear.devel")) Having defined the "*rss*" variable you can then do this: > CL-USER> (gethash "title" (nth 1 (gethash "entries" (py-unwrap *rss*)))) > "Web frameworks with ABCL (was Re: SBCL->ABCL migration questions)" > T Something I wanted to make was importing the Python modules as Lisp functions so instead of defining the "*rss-parse*" as > (defvar *rss-parse* (gethash "parse" > (py-unwrap (#"getDict" > (gethash "feedparser" > (py-unwrap (#"getLocals" *interpreter*))))))) you could do: > (feedparser:parse "http://rss.gmane.org/gmane.lisp.armedbear.devel") and that would also allow you slime's autocomplete to complete the python code. I've thought about doing the above using a macro around the "inspect" Python module, but never got around to it. Best Regards, Rolando Pereira

1 0