http://blog.gmane.org/gmane.lisp.ecl.general hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://comments.gmane.org/gmane.lisp.ecl.general/9107 <pre>Hi, I am new here and a Lisp beginner at his first significant development. I am having problem to use the :step option of the (trace) form as documented here: http://ecls.sourceforge.net/new-manual/re02.html I am using the ECL 12.2.1 REPL on Linux i686. CL-USER[1]&gt; (defun factorial (n) (if (plusp n) (* n (factorial (1- n))) 1)) FACTORIAL CL-USER[2]&gt; (trace (factorial :step t)) Condition of type: SIMPLE-ERROR Not a valid argument to TRACE: (FACTORIAL :STEP T) Available restarts: 1. (RESTART-TOPLEVEL) Go back to Top-Level REPL. Broken at SI:BYTECODES. [Evaluation of: (TRACE (FACTORIAL :STEP T))] In: #&lt;process TOP-LEVEL&gt;. ** BREAK [LEVEL 2]&gt; What am I doing wrong? Does the (trace) macro accept lists, doesn't it? BTW the equivalent and very similar option in another Lisp (GNU CLISP) works: (trace (factorial :step-if t)) Thanks -- Antonio Bonifati http://ninuzzo.github.com/ ------------------------------------------------------------------------------ Live Security Virtual Conference Excl</pre> Antonio Bonifati 2012-05-26T15:20:54 http://comments.gmane.org/gmane.lisp.ecl.general/9105 <pre>Hi Juanjo，I update to new version Ecl. The Chinese words work in windows console now. Thank for your help. But if I write the code into a lisp file and use load command in repl. it report error: ;;; Loading "D:/ecl/code/hello.lisp" Condition of type: STREAM-DECODING-ERROR decoding error on stream #&lt;input stream #P"D:/ecl/code/hello.lisp"&gt; (:EXTERNAL-F ORMAT (:UTF-8 :LF)): the octet sequence (255) cannot be decoded. Available restarts: 1. (CONTINUE) Read next character 2. (USE-VALUE) Replace the bogus sequence with a character 3. (RESTART-TOPLEVEL) Go back to Top-Level REPL. Broken at SI:BYTECODES. [Evaluation of: (LOAD "code/hello.lisp")] In: #&lt;process TOP-LEVEL&gt;. I export the file as utf-8 or utf-16, it's all failed. C.M. ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint secur</pre> cm 2012-05-26T02:58:42 http://comments.gmane.org/gmane.lisp.ecl.general/9101 <pre>I used visual studio 2008 It seems that both 32 and 64bit versions fails with assertion in function ecl_stream_to_handle where s-&gt;stream.mode == 0x11 (sww_io_wcon) Full stack traces: http://redplait.blogspot.com/2012/05/ecl-1251-on-xp64.html ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ </pre> red plait 2012-05-25T16:02:50 http://comments.gmane.org/gmane.lisp.ecl.general/9098 <pre>I write a hello.lisp file like this: (print "hi") (length '(a . b 1)) and I load it in to ecl: ;;; Loading "D:/ECL/code/hello.lisp" "hi" Condition of type: SIMPLE-READER-ERROR Reader error in file #&lt;input stream #P"D:/ECL/code/hello.lisp"&gt;, position 30: Too many objects after a list dot Available restarts: 1. (RESTART-TOPLEVEL) Go back to Top-Level REPL. Broken at SI:BYTECODES. [Evaluation of: (LOAD "code/hello.lisp")] In: #&lt;process TOP-LEVEL&gt;. It has not line number in error message but postion 30. I found it is the position of characters. how to get the line number of code from error message? Thank for your help! ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122</pre> cm 2012-05-23T05:10:39 http://comments.gmane.org/gmane.lisp.ecl.general/9096 <pre>I want to build a program without the need of ecl.dll, Is there anyone can help me ? ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________ Ecls-list mailing list Ecls-list-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f&lt; at &gt;public.gmane.org https://lists.sourceforge.net/lists/listinfo/ecls-list </pre> lookwong 2012-05-21T03:09:11 http://comments.gmane.org/gmane.lisp.ecl.general/9095 <pre>Hi everybody, after spending two days reading stuff about documented and undocumented functionality in Microsoft's API and CRT, I have come to the following conclusions: * In Windows, it is impossible to reasonably use POSIX functions to read/write text from the console without letting the console do some translation for you. * Moreover, POSIX's functions do not work, because read() and write() may actually read or write more bytes from the buffer thant you would want, precisely due to this "clever" translation. * This all renders the CRT unusable, and as stated in a MSDN blog from Windows, one should stick to the Console read and write operations and forget about C entirely. Well, I have coded a prototype of a Lisp stream for the console. ECL now has code that will detect whether the input/output/error channels are consoles and wrap them around with these streams. Moreover, it will also detect the current codepage and use it as external format for the stream. All this is pretty much experimental and I n</pre> Juan Jose Garcia-Ripoll 2012-05-19T08:12:10 http://comments.gmane.org/gmane.lisp.ecl.general/9085 <pre>ECL now implements asynchronous signal handlers through a secondary thread and these handlers can now be customized. The interface is given by the functions (EXT:GET-SIGNAL-HANDLER signal-code) (EXT:SET-SIGNAL-HANDLER signal-code closure-or-symbol) where the signal handler function should have the signature (LAMBDA (&amp;KEY PROCESS)). On a multithreaded ECL, instead of having a thread that relies on sigwait(), we have a thread that gets messages from other threads, which are the ones that received the signals. In principle we could optionally add a feature such that these threads get suspended until the signal is processed. This does not seem critical to me, but it might help cope with the problem of synchronous signals (SIGSEGV, SIGFPE, etc) which are not allowed to execute arbitrary code. In any case, these are refinements for the future. I would like to ask you to help debugging and perfect the changes I just introduced, and which for me are the final milestone for a release (apart from testing and integ</pre> Juan Jose Garcia-Ripoll 2012-05-17T06:58:32 http://comments.gmane.org/gmane.lisp.ecl.general/9084 <pre>I build Ecl on Window XP with Visual Studio. the code page of cmd console is cp936. D:\ECL&gt;ecl ECL (Embeddable Common-Lisp) 12.2.1 (git:UNKNOWN) Copyright (C) 1984 Taiichi Yuasa and Masami Hagiya Copyright (C) 1993 Giuseppe Attardi Copyright (C) 2000 Juan J. Garcia-Ripoll ECL is free software, and you are welcome to redistribute it under certain conditions; see file 'Copyright' for details. Type :h for Help. Top level in: #&lt;process TOP-LEVEL&gt;. (:ECL-BYTECMP :WIN32 :FORMATTER :LITTLE-ENDIAN :UINT32-T :UINT16-T :RELATIVE-PACKAGE-NAMES :UNICODE :DFFI :CLOS-STREAMS :CMU-FORMAT :WINDOWS :MSVC :ECL-PDE :DLOPEN :CLOS :THREADS :BOEHM-GC :ANSI-CL :COMMON-LISP :IEEE-FLOATING-POINT :PREFIXED-API :FFI :PENTIUM4 :COMMON :ECL) set *standard-input* :windows-cp936) language Condition of type: SIMPLE-ERROR Read or write operation to stream #&lt;input file "stdin"&gt; signaled an error. Explanation: Invalid argument. Available restarts: 1. (RESTART-TOPLEVEL) Go back to Top-Level REPL. Broken at SI:BYTECODES. [Evaluatio</pre> cm 2012-05-17T04:31:27 http://comments.gmane.org/gmane.lisp.ecl.general/9081 <pre>I had to reboot the system for another reason, but before I did, I checked the uptime of my HTTPd and it was up online during 30 days. Most of the connections were from google and vulnerability probing spiders, although interspersed with my Apache ab(8) test floods. This was still using a non-shrinking threads pool for now. Awesome :) </pre> Matthew Mondor 2012-05-16T20:27:34 http://comments.gmane.org/gmane.lisp.ecl.general/9080 <pre>Hello again, Although it's a minor issue, I noticed lately that when entering Lisp the following form using latest ECL (head) and latest Slime, I'm getting the following error in the command buffer: (with-open-file error in process filter: Wrong number of arguments: nil, 137 I'm not sure if the problem is in Slime or in ECL. Also, is it expected that this help only is available for implementation functions and bytecode compiled custom fonctions? I built my code with the debug level 3, but I cannot get that automatic interactive help for the functions of those files. Example: (defun foo (a b c) (list a b c)) (foo&lt;space&gt;[I get the lambda-list help here] (compile 'foo) (foo&lt;space&gt;[I no longer get it] Thanks, </pre> Matthew Mondor 2012-05-16T20:20:07 http://comments.gmane.org/gmane.lisp.ecl.general/9073 <pre>Uf, I finally managed to convert all my scripts and now the tests run thanks to "make check". This simplifies my life a lot and allows me to move ECL's checking to any computer. http://ecls.sourceforge.net/logs.html Windows is still pending. That is the most complicated virtual machine to run. On a side note, I reckon that some of you might be bored of my periodic emails. To avoid overflowing the mailing list with messages about irrelevant achievements, I will try to focus on the other channel, the News RSS feed in ECL. https://sourceforge.net/news/?group_id=30035 The News items do not allow comments and are not good for interactivity, but seem to reach a wider audience. In any case, if you find that some news item is worth discussing, feel free to post an answer here. I will gladly continue the conversation in this forum. Cheers, Juanjo </pre> Juan Jose Garcia-Ripoll 2012-05-05T09:03:17 http://comments.gmane.org/gmane.lisp.ecl.general/9070 <pre>ECL as it is in git/CVS seems to pass all of MOP-FEATURES tests http://common-lisp.net/project/closer/features.html Presumably this means Closer-to-mop is not needed any more with ECL. Juanjo </pre> Juan Jose Garcia-Ripoll 2012-04-28T09:47:52 http://comments.gmane.org/gmane.lisp.ecl.general/9069 <pre>Ever since the introduction of EXT:RUN-PROGRAM, ECL was experiencing random deadlocks when executing other programs. The reason for that was that the external process ended before ECL was able of calling sigwait() and retrieve its exit code --- the SIGPIPE arrived, but ECL had not yet finished constructing the external process structure and got lost. I have finally solved this stupid problem by synchronizing the child and the parent processes: using a pipe, the parent holds the child until the process structure is recorded with ECL. From that time on, ECL can keep track of the SIGPIPE POSIX signal that announces the end of the child process, providing the exit code. This fixes one of the most annoying bugs I ever faced -- 1 year and a half to come up with a solution, wow. Juanjo </pre> Juan Jose Garcia-Ripoll 2012-04-26T20:00:04 http://comments.gmane.org/gmane.lisp.ecl.general/9065 <pre>The following list is based on a similar one in closer-to-mop, explaining the problems in ECL which are solved by that library. It clarifies the status of MOP-compliance as of today -- very few things remain to be fixed, but a lot of testing is still required. Fixed in git/CVS - The functions ADD-DIRECT-METHOD and REMOVE-DIRECT-METHOD don't exist, and thus are also not called. - ADD-METHOD doesn't call FIND-METHOD or REMOVE-METHOD to remove potentially replaced methods, but uses some internal implementation instead. - SPECIALIZER-DIRECT-GENERIC-FUNCTIONS and SPECIALIZER-DIRECT-METHODS don't exist, and thus are also not updated by ADD-METHOD and REMOVE-METHOD. - VALIDATE-SUPERCLASS doesn't exist, and thus isn't called to determine whether two classes in a direct superclass relationship have compatible metaclasses. - COMPUTE-APPLICABLE-METHODS and COMPUTE-EFFECTIVE-METHOD are not generic functions. - COMPUTE-APPLICABLE-METHODS-USING-CLASSES doesn't exist. - DEFMETHOD do</pre> Juan Jose Garcia-Ripoll 2012-04-24T22:20:39 http://comments.gmane.org/gmane.lisp.ecl.general/9063 <pre>Dear Juanjo, is it OK with you if unbundle asdf-ecl from asdf.asd, and instead create a somewhat portable asdf extension called asdf-bundle.asd ? Impact: when the unbundling is complete, ecl users will be encouraged to (require :asdf-bundle) instead or in addition to (require :asdf) if they want to use the extensions currently provided in asdf-ecl. But it will make the bundle-op operations available to other implementations in a more portable way. —♯ƒ • François-René ÐVB Rideau •Reflection&amp;Cybernethics• http://fare.tunes.org To authorize all commercial relations between consenting adults. ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ____________</pre> Faré 2012-04-24T02:38:45 http://comments.gmane.org/gmane.lisp.ecl.general/9062 <pre>The AMOP mandates that every generic function and method be registered with the specializers they use. However, as far as I know it does not mandate any way to "delete" function objects, so that the following sequence leaves a trail of non-existent function objects in a class: * Define the class FOO * Define a method for function FAA specialized on FOO * Apply FMAKUNBOUND on 'FAA At the end class FOO still keeps a list of functions and methods that were registered with it. Juanjo </pre> Juan Jose Garcia-Ripoll 2012-04-23T18:38:50 http://comments.gmane.org/gmane.lisp.ecl.general/9061 <pre>* Both have been implemented and are used internally by ECL when dispatching. * A new function INTERN-EQL-SPECIALIZER is created. * This is accompanied by a new class, METAOBJECT. * This forced some fixes to MAKE-LOAD-FORM and FIND-METHOD. </pre> Juan Jose Garcia-Ripoll 2012-04-23T18:03:37 http://comments.gmane.org/gmane.lisp.ecl.general/9058 <pre>I had to add COMPUTE-APPLICABLE-METHODS-USING-CLASSES to the ECL MOP implementation. This was needed by Slime. In passing I also turned COMPUTE-EFFECTIVE-METHOD and COMPUTE-APPLICABLE-METHODS into generic function. This implies that certain code relying on Closer to MOP might break. In order to minimize the impact, I am trying to add the remaining features to ECL -- computation of method functions, specializer classes, etc. Please report if these changes impact your code or render ECL unusable for some reason. Best, Juanjo </pre> Juan Jose Garcia-Ripoll 2012-04-21T08:18:54 http://comments.gmane.org/gmane.lisp.ecl.general/9053 <pre>The current one relies on timed spinlocks. It is not great, but it works. I also solved some problems that prevented MSVC from building ECL. Juanjo </pre> Juan Jose Garcia-Ripoll 2012-04-18T20:45:51 http://comments.gmane.org/gmane.lisp.ecl.general/9052 <pre>I have committed yesterday a set of fixes for Windows. Among them is a first attempt at porting the new implementation of multithreading and synchronization objects -- it is not yet optimal, as it relies on timed spinlocks. Juanjo </pre> Juan Jose Garcia-Ripoll 2012-04-18T20:03:36 http://comments.gmane.org/gmane.lisp.ecl.general/9050 <pre>The code as it appears in the tree is considered feature complete. This means I do not want to introduce any further properties or arguments to the functions in the MP package and now the priority is shared among these goals: * Implement the spinlock-free wait queue in Windows. I believe this is doable with a simple SleepEx() and our current interrupt signaling mechanism -- both combined would behave like sigsuspend() in our POSIX code. * Further stress testing. I have added three files to src/tests/bugs/, covering locks, semaphores and mailboxes. Next is condition variables and trying to enlarge the set of tests to hit corner cases. * Analyzing "formally" the code for the queue. I have simplified the code and tried to specify it further in src/threads/queue.d, and it would be nice if some more knowledgeable people could have a look at this abstraction to see whether it might hit some ugly corner cases. * There are some hidden problems with the garbage collector and interrupts. I have not managed to rep</pre> Juan Jose Garcia-Ripoll 2012-04-15T21:33:11 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.lisp.ecl.general