From ungil at mac.com Wed Mar 9 00:44:11 2005 From: ungil at mac.com (Carlos Ungil) Date: Wed, 9 Mar 2005 01:44:11 +0100 Subject: [cl-gd-devel] clisp ffi Message-ID: <9fb2c29371319da6a6ec8a565b869ed7@mac.com> Hello, I've added support for clisp. It's only tested on Solaris, and not very much. The included test suite runs (almost) successfully, but I'm sure there will be bugs in the functionality not covered by these tests. Test 17 (ellipse filled in with tiles) fails, but it fails in the same way in SBCL so that might a problem with Solaris (or with my installation). Test 16 takes ages to complete, but I don't think I will optimize the code (as long as it is fast enough generating simple charts). I had no previous experience with clisp's FFI, so I'm sure there will be things that could have been done better, in particular about performance. The modified files can be found here http://ungil.com/cl-gd-0.4.2-clisp.tgz I've discovered as well some bugs, that I detail at the end of the message (only the first one is really relevant). Cheers, Carlos * [ colors-aux ] (defun resolve-c-color (color image) ......... ((or vector list) (setf (current-style image) (cdr color)) ;;; the first element of the list is dropped, that cdr shouldn't exist +styled+) ......... * [ gd-uffi ] (def-struct gd-image ......... (aal-x1 :int) (aal-y1 :int) (aal-x2 :int) (aal-y1 :int) ;;; this should be aal-y2, right? .......... * [ cl-gd-test ] (compare-files file) ;; the check is missing in test-012 From edi at agharta.de Wed Mar 9 15:17:11 2005 From: edi at agharta.de (Edi Weitz) Date: Wed, 09 Mar 2005 16:17:11 +0100 Subject: [cl-gd-devel] New release 0.4.3 (Was: clisp ffi) In-Reply-To: <9fb2c29371319da6a6ec8a565b869ed7@mac.com> (Carlos Ungil's message of "Wed, 9 Mar 2005 01:44:11 +0100") References: <9fb2c29371319da6a6ec8a565b869ed7@mac.com> Message-ID: On Wed, 9 Mar 2005 01:44:11 +0100, Carlos Ungil wrote: > I've added support for clisp. > > [...] > > The modified files can be found here > http://ungil.com/cl-gd-0.4.2-clisp.tgz > > I've discovered as well some bugs, that I detail at the end of the > message (only the first one is really relevant). Thanks. I've made a new release which fixes the bugs you found and adds a link to your CLISP port. I'm hesitant to add the port to the CL-GD distribution itself because it adds a large amount of code that I can't and won't maintain. I'd rather wait until CLISP is supported by UFFI... :) Download: Thanks again, Edi. From ungil at mac.com Wed Mar 9 19:09:10 2005 From: ungil at mac.com (Carlos Ungil) Date: Wed, 9 Mar 2005 20:09:10 +0100 Subject: [cl-gd-devel] Re: New release 0.4.3 (Was: clisp ffi) In-Reply-To: References: <9fb2c29371319da6a6ec8a565b869ed7@mac.com> Message-ID: Hello Edi, the link in the http://weitz.de/cl-gd page is wrong. Anyway, I've created a new package (including a README file) and changed the name to make it independent of the release number: http://ungil.com/cl-gd-clisp.tgz Also, I've noticed that you didn't change the images in test/orig. I've not tried, but I would expect some tests to fail (because of the bug that has been fixed). Cheers, Carlos On Mar 9, 2005, at 4:17 PM, Edi Weitz wrote: > On Wed, 9 Mar 2005 01:44:11 +0100, Carlos Ungil wrote: > >> I've added support for clisp. >> >> [...] >> >> The modified files can be found here >> http://ungil.com/cl-gd-0.4.2-clisp.tgz >> >> I've discovered as well some bugs, that I detail at the end of the >> message (only the first one is really relevant). > > Thanks. I've made a new release which fixes the bugs you found and > adds a link to your CLISP port. I'm hesitant to add the port to the > CL-GD distribution itself because it adds a large amount of code that > I can't and won't maintain. I'd rather wait until CLISP is supported > by UFFI... :) > > Download: > > Thanks again, > Edi. From edi at agharta.de Wed Mar 9 19:38:39 2005 From: edi at agharta.de (Edi Weitz) Date: Wed, 09 Mar 2005 20:38:39 +0100 Subject: [cl-gd-devel] New release 0.4.4 (Was: New release 0.4.3) In-Reply-To: (Carlos Ungil's message of "Wed, 9 Mar 2005 20:09:10 +0100") References: <9fb2c29371319da6a6ec8a565b869ed7@mac.com> Message-ID: On Wed, 9 Mar 2005 20:09:10 +0100, Carlos Ungil wrote: > the link in the http://weitz.de/cl-gd page is wrong. Argh! Sorry, I was in a hurry. This is fixed now (in 0.4.4). > Anyway, I've created a new package (including a README file) and > changed the name to make it independent of the release number: > > http://ungil.com/cl-gd-clisp.tgz > > Also, I've noticed that you didn't change the images in > test/orig. I've not tried, but I would expect some tests to fail > (because of the bug that has been fixed). I've done that now. However, I notice that test #17 still works with CMUCL but fails with AllegroCL, SBCL, and LispWorks. I wonder what the reason might be but I don't have enough time to investigate this at the moment. Any volunteers? Thanks, Edi. From edi at agharta.de Wed Mar 16 15:35:26 2005 From: edi at agharta.de (Edi Weitz) Date: Wed, 16 Mar 2005 16:35:26 +0100 Subject: [cl-gd-devel] New version 0.4.5 (Was: cl-gd:make-stream-fn) In-Reply-To: <16951.11860.397335.405939@zaphod.home.loc> (Walter C. Pelissero's message of "Tue, 15 Mar 2005 19:49:56 +0100") References: <16951.11860.397335.405939@zaphod.home.loc> Message-ID: [Cc to mailing list] On Tue, 15 Mar 2005 19:49:56 +0100, "Walter C. Pelissero" wrote: > I'm trying your cl-gd with PAServe and stumbled in a minor glitch in > MAKE-STREAM-FN. The COND first checks for: > > (subtypep (stream-element-type stream) '(unsigned-byte 8)) > > and then > > (subtypep (stream-element-type stream) 'character) > > Which both fail miserably at least on CMUCL with *HTML-STREAM* as > STREAM. The reason is that: > > (stream-element-type *HTML-STREAM*) => '(OR CHARACTER (UNSIGNED-BYTE 8)) > > (subtypep '(or character (unsigned-byte 8)) 'character) => NIL T > > (subtypep '(or character (unsigned-byte 8)) '(unsigned-byte 8)) => NIL T > > Apparently changing the order of the operands, makes the COND work > as expected, although, as far as I understand, the semantic of > subtypep isn't respected. No, I think CMUCL is right. This was a thinko in my code. Thanks for the report, I've uploaded a new version which fixes this. Cheers, Edi. From edi at agharta.de Sun Mar 20 21:11:07 2005 From: edi at agharta.de (Edi Weitz) Date: Sun, 20 Mar 2005 22:11:07 +0100 Subject: [cl-gd-devel] Re: CL-GD very minor typo In-Reply-To: <20050320203957.GA16708@ix.top> (Peter BARABAS's message of "Sun, 20 Mar 2005 21:39:57 +0100") References: <20050320203957.GA16708@ix.top> Message-ID: Hi! On Sun, 20 Mar 2005 21:39:57 +0100, Peter BARABAS wrote: > I'd like to report a simple typo in the latest CL-GD. It's in > images.lisp: > > (defmacro with-image* ((width height &optional true-color) &body body) > "Creates an image with size WIDTH x HEIGHT and executes BODY with > the image bound to *DEFAULT-IMAGE*. If TRUE-COLOR is true, creates a > true color image. The image is guaranteed to be destroyed before this > macro exits." > `(with-image (*default-image* ,width, height ,true-color) > , at body)) > > Note the placement of the comma before "height". Thanks. That's certainly a typo and will be fixed in the next release. > I know that the Lisp implementations are permissive (or the > HyperSpec?). I'm pretty sure this is legal Common Lisp although it was not what I intended... :) Cheers, Edi. From edi at agharta.de Thu Mar 31 09:39:42 2005 From: edi at agharta.de (Edi Weitz) Date: Thu, 31 Mar 2005 11:39:42 +0200 Subject: [cl-gd-devel] New version 0.4.6 Message-ID: Changelog: Version 0.4.6 2005-03-31 Fixed typo in WITH-IMAGE* (thanks to Peter Barabas) Handle CMUCL search lists correctly (thanks to Hans H?bner) Added -lc option to linker call and included makefile (thanks to Hans H?bner) Download: Cheers, Edi.