From dauclair at hotmail.com Tue Apr 2 01:10:13 2002 From: dauclair at hotmail.com (Douglas Michael Auclair) Date: Tue, 02 Apr 2002 01:10:13 Subject: unsealing causes issues? Message-ID: Hello, hackers, Hey, I'm a little confused here. I tried unsealing on my local machine in order to finish up sockets and now (as opposed to a year ago) I get the following error: Expected an instance of {Class } but got #f. Ugh! This doesn't change at all, no matter how many (all) methods on the newly opened class I unseal. This seems new, as I don't recall having this problem when I implemented most of the network library on 2.3.(4? 5?). Any hints/help? As it is, I'm abandoning making a subclass of , but that may me beaucoup code duplication from the streams library to the network library. Joy. Sincerely, Doug Auclair _________________________________________________________________ MSN Photos is the easiest way to share and print your photos: http://photos.msn.com/support/worldwide.aspx From housel at acm.org Tue Apr 2 20:08:19 2002 From: housel at acm.org (Peter S. Housel) Date: Tue, 2 Apr 2002 10:08:19 -0800 Subject: unsealing causes issues? References: Message-ID: <009201c1da71$5ecae780$7421010a@nb.mindspeed.com> Douglas Michael Auclair wrote: > Hey, I'm a little confused here. I tried unsealing on my local > machine in order to finish up sockets and now (as opposed to a year ago) I > get the following error: > > Expected an instance of {Class } but got #f. > > Ugh! This doesn't change at all, no matter how many (all) methods on the > newly opened class I unseal. > > This seems new, as I don't recall having this problem when I implemented > most of the network library on 2.3.(4? 5?). > > Any hints/help? How do I reproduce this problem? Cheers, -Peter- From dauclair at hotmail.com Wed Apr 3 21:00:09 2002 From: dauclair at hotmail.com (Douglas Michael Auclair) Date: Wed, 03 Apr 2002 21:00:09 Subject: unsealing causes issues? Message-ID: >From: "Peter S. Housel" [snip sealing] >How do I reproduce this problem? Dear Peter and all, Add the 'open' keyword in common/streams/file-streams.dylan on the definer macro. When that didn't work for me, I also commented out the 'sealed domain' declarations on everything that was connected to . That didn't work either. As soon as the compiler tries to make a PROGRAM, like parsergen or melange, it throws the #f instead of error. Sincerely, Doug Auclair _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp. From bfulgham at debian.org Sun Apr 7 10:17:37 2002 From: bfulgham at debian.org (Brent A. Fulgham) Date: Sun, 07 Apr 2002 00:17:37 -0800 Subject: Macro Problems :-( Message-ID: <20020407081737.GA6431@pacbell.net> Can someone smart with Macros help me debug this problem? I'm trying to use the "dynamic-bind" macro in common-dylan, but I keep getting weird compile errors. Attached is a lid file, source file, and exports file for my test case. If I take the text of the macro from "fun-dev-compat.dylan" and paste it into my program (under the name dynamic-bind2) everything works. If I expand the macro by hand and enter that text into my program, it works. But if I use the macro as exported from common-dylan, it won't compile. Questions: 1. Can anyone see what's wrong? 2. Is there some way to force d2c to write out a macro-expanded version of the source file so I can see what it's doing wrong? Arg! :-( -Brent -------------- next part -------------- Module: when Synopsis: Contains definitions for Testworks test harness Author: Shri Amit, Andrew Armstrong Copyright: Original Code is Copyright (c) 1996-2000 Functional Objects, Inc. All rights reserved. License: Functional Objects Library Public License Version 1.0 Dual-license: GNU Lesser General Public License Warranty: Distributed WITHOUT WARRANTY OF ANY KIND // dynamic-bind // shadow the given variables within a block we add define macro dynamic-bind2 { dynamic-bind2 (?var:name = ?val:expression) ?:body end } => { begin let old-value = ?var; block () ?var := ?val; ?body cleanup ?var := old-value; end block; end } { dynamic-bind2 (?var:name = ?val:expression, ?others:*) ?:body end } => { begin let old-value = ?var; block () ?var := ?val; dynamic-bind (?others) ?body end; cleanup ?var := old-value; end block; end } { dynamic-bind (?:name(?arg:expression) ?eq:token ?val:expression) ?:body end } => { ?name ## "-dynamic-binder"(?val, method() ?body end, ?arg) } { dynamic-bind (?:name(?arg:expression) ?eq:token ?val:expression,?others:*) ?:body end } => { ?name ## "-dynamic-binder"(?val, method () dynamic-bind (?others) ?body end; end, ?arg) } end macro dynamic-bind2; /// Perform component define function main() let a = 0; let b = 2; format-out ("a = %d, b = %d\n", a, b); /* // I work perfectly! begin let old-value = b; block () b := 3; format-out ("a = %d, b = %d", a, b); cleanup b := 2; end block; end; */ /* // I work perfectly, too! dynamic-bind2 (b = 3) format-out ("a = %d, b = %d\n", a, b); end; */ dynamic-bind (b = 3) format-out ("a = %d, b = %d", a, b); end; format-out ("a = %d, b = %d\n", a, b); end function main; main(); -------------- next part -------------- module: dylan-user define library when use common-dylan; use io; end library; define module when use common-dylan; use format-out; end module; -------------- next part -------------- library: when executable: when files: when-exports when -------------- next part -------------- when: when.lid when.dylan when-exports.dylan d2c when.lid clean: -rm -f *.o *.s *.a *.c *.mak *~ when -rm -rf .libs From david at lichteblau.com Sun Apr 7 10:36:40 2002 From: david at lichteblau.com (David Lichteblau) Date: Sun, 7 Apr 2002 10:36:40 +0200 Subject: Macro Problems :-( In-Reply-To: <20020407081737.GA6431@pacbell.net> References: <20020407081737.GA6431@pacbell.net> Message-ID: <20020407103640.F25287@lichteblau.com> Quoting Brent A. Fulgham (bfulgham at debian.org): Version 1 (attached): > Attached is a lid file, source file, and exports file for my test case. > If I take the text of the macro from "fun-dev-compat.dylan" and paste it > into my program (under the name dynamic-bind2) everything works. If I > expand the macro by hand and enter that text into my program, it works. Version 2 (not attached): > But if I use the macro as exported from common-dylan, it won't compile. Hmm, version 1 as attached is supposed to work? It does not, because you forgot to rename dynamic-bind to -bind2 at some places. Version 2 does not work? I cannot find version 2 in the attachment... Please provide a testcase! David -------------- next part -------------- --- when/when.dylan Sun Apr 7 10:26:15 2002 +++ when-fixed/when.dylan Sun Apr 7 10:26:52 2002 @@ -7,7 +7,7 @@ Dual-license: GNU Lesser General Public License Warranty: Distributed WITHOUT WARRANTY OF ANY KIND -// dynamic-bind +// dynamic-bind2 // shadow the given variables within a block we add define macro dynamic-bind2 @@ -30,23 +30,23 @@ let old-value = ?var; block () ?var := ?val; - dynamic-bind (?others) ?body end; + dynamic-bind2 (?others) ?body end; cleanup ?var := old-value; end block; end } - { dynamic-bind (?:name(?arg:expression) ?eq:token ?val:expression) + { dynamic-bind2 (?:name(?arg:expression) ?eq:token ?val:expression) ?:body end } => { ?name ## "-dynamic-binder"(?val, method() ?body end, ?arg) } - { dynamic-bind (?:name(?arg:expression) ?eq:token ?val:expression,?others:*) + { dynamic-bind2 (?:name(?arg:expression) ?eq:token ?val:expression,?others:*) ?:body end } => { ?name ## "-dynamic-binder"(?val, method () - dynamic-bind (?others) + dynamic-bind2 (?others) ?body end; end, @@ -78,7 +78,7 @@ end; */ - dynamic-bind (b = 3) + dynamic-bind2 (b = 3) format-out ("a = %d, b = %d", a, b); end; -------------- next part -------------- --- when-fixed/when.dylan Sun Apr 7 10:26:52 2002 +++ when-simplified-recursion/when.dylan Sun Apr 7 10:28:03 2002 @@ -11,18 +11,10 @@ // shadow the given variables within a block we add define macro dynamic-bind2 - { dynamic-bind2 (?var:name = ?val:expression) + { dynamic-bind2 () ?:body end } - => { begin - let old-value = ?var; - block () - ?var := ?val; - ?body - cleanup - ?var := old-value; - end block; - end } + => { ?body } { dynamic-bind2 (?var:name = ?val:expression, ?others:*) ?:body end } @@ -35,12 +27,6 @@ ?var := old-value; end block; end } - { dynamic-bind2 (?:name(?arg:expression) ?eq:token ?val:expression) - ?:body - end } - => { ?name ## "-dynamic-binder"(?val, - method() ?body end, - ?arg) } { dynamic-bind2 (?:name(?arg:expression) ?eq:token ?val:expression,?others:*) ?:body end } From dauclair at hotmail.com Sun Apr 7 13:35:07 2002 From: dauclair at hotmail.com (Douglas Michael Auclair) Date: Sun, 07 Apr 2002 13:35:07 Subject: each-subclass initializers! libtool on Mac OS-X (10.1)? Message-ID: [snip my request to Gabor to add each-subclass initializers] >From: pubcvs at berlin.ccc.de >allow keyword init of each-subclass slots too. >Init values/functions for each-subclass already worked (magic!) Dear Gabor, You are so fast! Dear Mac experts (including Gabor, Rob and Gareth), Okay, so I'm on Mac OS-X (10.1), working on rebuilding d2c, and libtool has so far refused to build, even after I kicked the sources in some ways recommended in fink.sourceforge.net. I guess libtool 1.3.5 isn't good enough? What did y'all do to get d2c build for darwin? Sincerely, Doug Auclair _________________________________________________________________ Join the world?s largest e-mail service with MSN Hotmail. http://www.hotmail.com From gabor at mac.com Mon Apr 8 01:07:04 2002 From: gabor at mac.com (Gabor Greif) Date: Mon, 08 Apr 2002 01:07:04 +0200 Subject: each-subclass initializers! libtool on Mac OS-X (10.1)? Message-ID: <200204072306.BAA17906@lucky.symbiose.com> ---------- >From: "Douglas Michael Auclair" >To: gabor at mac.com, gd-hackers at gwydiondylan.org >Subject: each-subclass initializers! libtool on Mac OS-X (10.1)? >Date: Son, 7. Apr 2002 15:35 Uhr > > Dear Gabor, > > You are so fast! Rapidly delivering flawed solutions :-( Serious link problems surface if there are init-values in the library for any class-allocated slot. I hope to have this fixed this week. Gabor From bfulgham at debian.org Mon Apr 8 01:33:44 2002 From: bfulgham at debian.org (Brent A. Fulgham) Date: Sun, 07 Apr 2002 16:33:44 -0700 Subject: Macro Problems :-( In-Reply-To: <20020407103640.F25287@lichteblau.com> References: <20020407081737.GA6431@pacbell.net> <20020407103640.F25287@lichteblau.com> Message-ID: <20020407233344.GA669@pacbell.net> On Sun, Apr 07, 2002 at 10:36:40AM +0200, David Lichteblau wrote: > Version 1 (attached): > > Attached is a lid file, source file, and exports file for my test case. > > If I take the text of the macro from "fun-dev-compat.dylan" and paste it > > into my program (under the name dynamic-bind2) everything works. If I > > expand the macro by hand and enter that text into my program, it works. > > Version 2 (not attached): > > But if I use the macro as exported from common-dylan, it won't compile. > > Hmm, version 1 as attached is supposed to work? It does not, because > you forgot to rename dynamic-bind to -bind2 at some places. > Huh. Well, I was really just trying to mirror the dynamic-bind that's in common-dylan so that I could investigate why this was happening. > Version 2 does not work? I cannot find version 2 in the attachment... > Please provide a testcase! All three versions were in the original 'when.dylan' file. I've revised things a bit to be more clear. Case 1: A simple hand-expansion of the macro. Case 2: The local version of the macro. Case 3: Using the macro as exported by common-dylan. This fails to compile, so I have it commented out. Try for yourself! Can anyone verify if they have the same trouble? Thanks, -Brent -------------- next part -------------- Module: when Synopsis: Contains definitions for Testworks test harness Author: Shri Amit, Andrew Armstrong Copyright: Original Code is Copyright (c) 1996-2000 Functional Objects, Inc. All rights reserved. License: Functional Objects Library Public License Version 1.0 Dual-license: GNU Lesser General Public License Warranty: Distributed WITHOUT WARRANTY OF ANY KIND // dynamic-bind // shadow the given variables within a block we add define macro dynamic-bind2 { dynamic-bind2 (?var:name = ?val:expression) ?:body end } => { begin let old-value = ?var; block () ?var := ?val; ?body cleanup ?var := old-value; end block; end } { dynamic-bind2 (?var:name = ?val:expression, ?others:*) ?:body end } => { begin let old-value = ?var; block () ?var := ?val; dynamic-bind2 (?others) ?body end; cleanup ?var := old-value; end block; end } { dynamic-bind2 (?:name(?arg:expression) ?eq:token ?val:expression) ?:body end } => { ?name ## "-dynamic-binder"(?val, method() ?body end, ?arg) } { dynamic-bind2 (?:name(?arg:expression) ?eq:token ?val:expression,?others:*) ?:body end } => { ?name ## "-dynamic-binder"(?val, method () dynamic-bind2 (?others) ?body end; end, ?arg) } end macro dynamic-bind2; /// Perform component define function main() let a = 0; let b = 2; format-out ("a = %d, b = %d\n", a, b); // Case 1: I work perfectly! begin let old-value = b; block () b := 3; format-out ("Case 1: a = %d, b = %d", a, b); cleanup b := 2; end block; end; // Case 2: I work perfectly, too! dynamic-bind2 (b = 3) format-out ("Case 2: a = %d, b = %d\n", a, b); end; // Case 3: I fail to compile! //dynamic-bind (b = 3) // format-out ("Case 3: a = %d, b = %d", a, b); //end; format-out ("a = %d, b = %d\n", a, b); end function main; main(); -------------- next part -------------- library: when executable: when files: when-exports when -------------- next part -------------- module: dylan-user define library when use common-dylan; use io; end library; define module when use common-dylan; use format-out; end module; -------------- next part -------------- when: when.lid when.dylan when-exports.dylan d2c when.lid clean: -rm -f *.o *.s *.a *.c *.mak *~ when -rm -rf .libs From david at lichteblau.com Mon Apr 8 09:44:13 2002 From: david at lichteblau.com (David Lichteblau) Date: Mon, 8 Apr 2002 09:44:13 +0200 Subject: Macro Problems :-( In-Reply-To: <20020407233344.GA669@pacbell.net> References: <20020407081737.GA6431@pacbell.net> <20020407103640.F25287@lichteblau.com> <20020407233344.GA669@pacbell.net> Message-ID: <20020408094413.G25287@lichteblau.com> Quoting Brent A. Fulgham (bfulgham at debian.org): > All three versions were in the original 'when.dylan' file. I've revised > things a bit to be more clear. Oh, sorry. > Can anyone verify if they have the same trouble? Hmm, dynamic-bind is not exported from common-dylan here, but from functional-extensions. David --- dynamic-bind/when-exports.dylan Mon Apr 8 09:35:17 2002 +++ dynamic-bind-extensions/when-exports.dylan Mon Apr 8 09:41:05 2002 @@ -7,5 +7,6 @@ define module when use common-dylan; + use functional-extensions; use format-out; end module; From brent.fulgham at xpsystems.com Mon Apr 8 18:27:04 2002 From: brent.fulgham at xpsystems.com (Brent Fulgham) Date: Mon, 8 Apr 2002 09:27:04 -0700 Subject: Macro Problems :-( Message-ID: David, Thanks for your help: > > Can anyone verify if they have the same trouble? > > Hmm, dynamic-bind is not exported from common-dylan here, but from > functional-extensions. > I believe you are probably using Functional Developer. I was trying to test our (Gwydion Dylan's) implementation of dynamic-bind (which is currently exported by common-dylan). We should probably move it to a new library (functional-extensions) to support Functional Developer better. However, it just seems to have problems for me, so I'd like to get that resolved before doing any further changes. Do you have access to a copy of Gwydion Dylan you could use to test? -Brent From david at lichteblau.com Mon Apr 8 19:58:37 2002 From: david at lichteblau.com (David Lichteblau) Date: Mon, 8 Apr 2002 19:58:37 +0200 Subject: Macro Problems :-( In-Reply-To: References: Message-ID: <20020408195836.D2326@lichteblau.com> Quoting Brent Fulgham (brent.fulgham at xpsystems.com): > > Hmm, dynamic-bind is not exported from common-dylan here, but from > > functional-extensions. > > I believe you are probably using Functional Developer. I was trying to > test our (Gwydion Dylan's) implementation of dynamic-bind (which is > currently exported by common-dylan). We should probably move it to a > new library (functional-extensions) to support Functional Developer better. > > However, it just seems to have problems for me, so I'd like to get that > resolved before doing any further changes. > > Do you have access to a copy of Gwydion Dylan you could use to test? Well, I am probably still just too confused to find out what your question is... ;-) Anyway, I am still assuming that we are talking about common-dylan from current Gwydion Dylan sources, which exports dynamic-bind from module functional-extensions (libraries common-extensions as well as common-dylan), and that is what the one-liner diff shown in my last mail imports. d. From brent.fulgham at xpsystems.com Mon Apr 8 20:08:06 2002 From: brent.fulgham at xpsystems.com (Brent Fulgham) Date: Mon, 8 Apr 2002 11:08:06 -0700 Subject: Macro Problems :-( Message-ID: > Well, I am probably still just too confused to find out what your > question is... ;-) Anyway, I am still assuming that we are talking > about common-dylan from current Gwydion Dylan sources, which exports > dynamic-bind from module functional-extensions (libraries > common-extensions as well as common-dylan), and that is what the > one-liner diff shown in my last mail imports. > No -- you solved my problem: I was expecting to get an error from d2c that "dynamic-bind" was not defined or valid (since I had not imported the functional-extensions module). Instead, what I was getting was this: Parsing when.dylan "when.dylan", line 104, character 1 through line 152, character 4: define function main() ^^^^^^^^^^^^^^^^^^^^^^ through end; ^^^^^^^^^^^ Error: Syntax error in ``define function'' macro. None of the main rules matched. skipping rest of when.dylan seeding representations Which made me think there was a syntax error or something in the macro definition. By importing the module (as you suggested), I find that it works as expected. So you have resolved my problem. Thanks, -Brent From hugh_greene at yahoo.com Mon Apr 8 23:05:51 2002 From: hugh_greene at yahoo.com (Hugh Greene) Date: Mon, 08 Apr 2002 22:05:51 +0100 Subject: Problems with 2.3.9pre1-x86-win32 for MS VC [was: Re: Revision] In-Reply-To: <87zo0eumvn.fsf@andreas.org> References: Message-ID: <5.0.2.1.0.20020408214718.01c6a620@pop.mail.yahoo.com> [Moved back on-list] At 17:21 2002/04/08 +0000, Andreas Bogk wrote: >"Brent Fulgham" writes: > > Please try moving [gwydion-dylan-2.3.9pre1-x86-win32.zip] again > > and let me know if it is corrupted or not. ... > >Seems to be good on the server this time. Thanks Brent, this build unzipped okay for me, too; but I'm still having problems using it. We'll get there (or rather, here ;-) yet, I'm sure ... I've remembered to set DYLANDIR this time, but now I get the output below, followed by a system popup dialog telling me that the DLL "msys-1.0.dll" couldn't be found for "cmp.exe" (which is in the gwydion bin directory). That DLL is nowhere on my system. ---------------------------------------------------------------- H:\Users\hughg\Dylan\projects\hello-world>d2c hello-world.lid Compiling library hello-world [Loading library Dylan...] Parsing library.dylan Creating cc-hello-world-files.mak Processing library.dylan .. 'mv' is not recognized as an internal or external command, operable program or batch file. mv failed? abnormal program termination H:\Users\hughg\Dylan\projects\hello-world> ---------------------------------------------------------------- I've worked around this for now by putting my Cygwin install on the path before the Gwydion directories. The next problem is ... ---------------------------------------------------------------- H:\Users\hughg\Dylan\projects\hello-world>d2c hello-world.lid Compiling library hello-world [Loading library Dylan...] Parsing library.dylan make -f cc-hello-world-files.mak cl -nologo -Ox -DWIN32 -I. -IG:\Programs\gwydion-dylan-2.3.9pre1-x86-win32/include -c inits.c -o inits.obj inits.c inits.c(1) : fatal error C1083: Cannot open include file: 'runtime.h': No such file or directory make: *** [inits.obj] Error 2 gmake failed? abnormal program termination ---------------------------------------------------------------- The file "H:\gwydion\gwydion-dylan-2.3.9pre1-x86-win32\include\runtime.h" certainly exists. I wondered if it might be because the "cl" line has a "/" instead of a "\" in the second "-I" directive path. However, I tried hacking CCFLAGS in cc-hello-world-files.mak to change this and I still get he same problem. atdhvaannkcse for any clues, Hugh _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com From brent.fulgham at xpsystems.com Mon Apr 8 23:35:23 2002 From: brent.fulgham at xpsystems.com (Brent Fulgham) Date: Mon, 8 Apr 2002 14:35:23 -0700 Subject: Problems with 2.3.9pre1-x86-win32 for MS VC [was: Re: Revision] Message-ID: > I've remembered to set DYLANDIR this time, but now I get the > output below, followed by a system popup dialog telling me > that the DLL "msys-1.0.dll" couldn't be found for "cmp.exe" > (which is in the gwydion bin directory). That DLL is nowhere > on my system. > This is my bad -- I copied the "cmp.exe" from the Mingw distribution (Msys-1.0) into the ZIP file, but neglected the msys-1.0.dll file. I'll correct that in a future upload. For now you can download the Msys-1.0 stuff from http://mingw.sourceforge.net. > 'mv' is not recognized as an internal or external command, > operable program or batch file. > mv failed? > > abnormal program termination Okay -- another dependency. I'll add 'mv.exe' to the stuff I provide in the ZIP file. > H:\Users\hughg\Dylan\projects\hello-world> > ---------------------------------------------------------------- > > I've worked around this for now by putting my Cygwin install > on the path > before the Gwydion directories. The next problem is ... > > ---------------------------------------------------------------- > H:\Users\hughg\Dylan\projects\hello-world>d2c hello-world.lid > Compiling library hello-world > [Loading library Dylan...] > Parsing library.dylan > > make -f cc-hello-world-files.mak > cl -nologo -Ox -DWIN32 -I. > -IG:\Programs\gwydion-dylan-2.3.9pre1-x86-win32/include -c inits.c -o > inits.obj > inits.c > inits.c(1) : fatal error C1083: Cannot open include file: > 'runtime.h': No > such file or directory > make: *** [inits.obj] Error 2 > gmake failed? > Hrm! Could you send me the makefile (cc-hello-world-files.mak) so I could review it? Thanks for your testing! -Brent From hugh_greene at yahoo.com Wed Apr 10 11:03:54 2002 From: hugh_greene at yahoo.com (Hugh Greene) Date: Wed, 10 Apr 2002 10:03:54 +0100 Subject: Problems with 2.3.9pre1-x86-win32 for MS VC [was: Re: Revision] In-Reply-To: Message-ID: <5.0.2.1.0.20020410095900.01c5ffc0@pop.mail.yahoo.com> At 14:35 2002/04/08 -0700, Brent Fulgham wrote: > > I've remembered to set DYLANDIR this time, but now I get the > > output below, followed by a system popup dialog telling me > > that the DLL "msys-1.0.dll" couldn't be found for "cmp.exe" ... > >This is my bad -- I copied the "cmp.exe" from the Mingw distribution >(Msys-1.0) into the ZIP file, but neglected the msys-1.0.dll file. >I'll correct that in a future upload. For now you can download >the Msys-1.0 stuff from http://mingw.sourceforge.net. Okay, got it. I had to update "rm.exe" and "cmp.exe" as the latest version there is slightly newer than the one you included, but that works. > > [...] > > 'mv' is not recognized as an internal or external command, > > operable program or batch file. [...] > >Okay -- another dependency. I'll add 'mv.exe' to the stuff I provide >in the ZIP file. It looks like you're missing "touch.exe" as well. > > [...] The next problem is ... > > > > ---------------------------------------------------------------- > > H:\Users\hughg\Dylan\projects\hello-world>d2c hello-world.lid > > Compiling library hello-world > > [Loading library Dylan...] > > Parsing library.dylan > > > > make -f cc-hello-world-files.mak > > cl -nologo -Ox -DWIN32 -I. > > -IG:\Programs\gwydion-dylan-2.3.9pre1-x86-win32/include -c inits.c -o > > inits.obj > > inits.c > > inits.c(1) : fatal error C1083: Cannot open include file: > > 'runtime.h': No > > such file or directory > > make: *** [inits.obj] Error 2 > > gmake failed? > > > >Hrm! Could you send me the makefile (cc-hello-world-files.mak) >so I could review it? Actually, now that I've patched up the msys utilities, it seems to work fine. Maybe the above was a result of the "mv" failing, or because I was using the Cygwin make instead of the msys one? In any case, I've now got my very own ~1Mb "Hello world" executable :-) Thanks, Hugh _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com From carlgay at attbi.com Fri Apr 12 17:34:14 2002 From: carlgay at attbi.com (Carl Gay) Date: Fri, 12 Apr 2002 11:34:14 -0400 Subject: FTP sites out of date Message-ID: <5.1.0.14.0.20020412112840.02404d90@mail.attbi.com> FYI, the list of download sites on downloading.phtml needs a bit of updating: The Federated - says latest version is 2.3.6. Basis Technology Corp - down (was down several days ago also) KDD - down Infodrom - down From dauclair at hotmail.com Fri Apr 12 17:08:45 2002 From: dauclair at hotmail.com (Douglas Michael Auclair) Date: Fri, 12 Apr 2002 17:08:45 Subject: FTP sites out of date Message-ID: >From: Carl Gay >FYI, the list of download sites on downloading.phtml needs a bit of >updating: (sort of OT) Also, the latest source updates stopped in February ... are they now archived elsewhere? Sincerely, Doug Auclair _________________________________________________________________ MSN Photos is the easiest way to share and print your photos: http://photos.msn.com/support/worldwide.aspx From gabor at mac.com Fri Apr 12 23:11:39 2002 From: gabor at mac.com (Gabor Greif) Date: Fri, 12 Apr 2002 23:11:39 +0200 Subject: each-subclass initializers! libtool on Mac OS-X (10.1)? Message-ID: <200204122111.XAA28091@lucky.symbiose.com> Finally I have the feeling each-subclass stuff begins to work out. Inherited each-subclass slots with init-values/functions should be ok now too. Thanks to Martin Kaeske, who reported problems and so helped to iron out some bugs. Remember, I am writing software, not testing it, so it is now your turn! ;-) Have fun, Gabor ---------- >From: "Gabor Greif" >To: gd-hackers at gwydiondylan.org >Subject: Re: each-subclass initializers! libtool on Mac OS-X (10.1)? >Date: Mon, 8. Apr 2002 1:07 Uhr > > > > ---------- >>From: "Douglas Michael Auclair" >>To: gabor at mac.com, gd-hackers at gwydiondylan.org >>Subject: each-subclass initializers! libtool on Mac OS-X (10.1)? >>Date: Son, 7. Apr 2002 15:35 Uhr >> > >> Dear Gabor, >> >> You are so fast! > > > Rapidly delivering flawed solutions :-( > > Serious link problems surface if there are init-values in the library for > any class-allocated slot. > > I hope to have this fixed this week. > > Gabor From robmyers at mac.com Fri Apr 12 23:40:56 2002 From: robmyers at mac.com (Rob Myers) Date: Fri, 12 Apr 2002 22:40:56 +0100 Subject: Melange and long longs Message-ID: Is melange meant to handle long longs OK? Particularly long long function parameters. I think this may be causing a problem in parsing the Carbon headers. - Rob. From ggreif at lucent.com Sat Apr 13 14:57:32 2002 From: ggreif at lucent.com (Gabor Greif) Date: Sat, 13 Apr 2002 14:57:32 +0200 Subject: Mailing list archive unaccessible Message-ID: <3CB82B3C.8FCFC840@lucent.com> When visiting http://www.gwydiondylan.org/list-archive/gd-hackers/threads.html and clicking on an article I get: #################################################################################### Forbidden You don't have permission to access /list-archive/gd-hackers/msg04613.html on this server. Apache/1.3.14 Server at www.gwydiondylan.org Port 80 #################################################################################### Any idea why? Thanks, Gabor From robmyers at mac.com Sat Apr 13 22:12:35 2002 From: robmyers at mac.com (Rob Myers) Date: Sat, 13 Apr 2002 21:12:35 +0100 Subject: RFC: Melange long long parameter hack, Was Re: Melange and long longs In-Reply-To: Message-ID: On Friday, April 12, 2002, at 10:40 PM, Rob Myers wrote: > Is melange meant to handle long longs OK? Particularly long long > function parameters. I think this may be causing a problem in parsing > the Carbon headers. Melange doesn't handle long long parameters. It seems to be happy with long long pointer parameters and long longs in structs, I'd like to modify melange to handle long longs as longs: C: long long fun(long long a); => Dylan: call-out("fun", long:, long: a); This is obviously broken, but the MacOS Carbon headers have several functions that take long long parameters, and I'd like to be able to handle them in some way. I'll add a comment that this should be fixed when call-out can handle long-long: parameters. Please can you comment on this proposal. If anyone has a better proposal I'd very much like to hear it. :-). - Rob. From andreas at andreas.org Sun Apr 14 20:32:52 2002 From: andreas at andreas.org (Andreas Bogk) Date: 14 Apr 2002 18:32:52 +0000 Subject: Mailing list archive unaccessible In-Reply-To: <3CB82B3C.8FCFC840@lucent.com> References: <3CB82B3C.8FCFC840@lucent.com> Message-ID: <87sn5yyvu3.fsf@andreas.org> Gabor Greif writes: > You don't have permission to access > /list-archive/gd-hackers/msg04613.html on this server. Hm, *some* files had weird permissions. And the mhonarc database is broken, and fails to add new files. *sigh* And since orbz.org is offline, spam is taking over rapidly. I'll try to do something as soon as I can. Andreas -- "In my eyes it is never a crime to steal knowledge. It is a good theft. The pirate of knowledge is a good pirate." (Michel Serres) From andreas at andreas.org Sun Apr 14 20:37:29 2002 From: andreas at andreas.org (Andreas Bogk) Date: 14 Apr 2002 18:37:29 +0000 Subject: FTP sites out of date In-Reply-To: <5.1.0.14.0.20020412112840.02404d90@mail.attbi.com> References: <5.1.0.14.0.20020412112840.02404d90@mail.attbi.com> Message-ID: <87ofgmyvme.fsf@andreas.org> Carl Gay writes: > FYI, the list of download sites on downloading.phtml needs a bit of updating: Thanks for notifying, done. Andreas -- "In my eyes it is never a crime to steal knowledge. It is a good theft. The pirate of knowledge is a good pirate." (Michel Serres) From andreas at andreas.org Sun Apr 14 20:38:57 2002 From: andreas at andreas.org (Andreas Bogk) Date: 14 Apr 2002 18:38:57 +0000 Subject: FTP sites out of date In-Reply-To: References: Message-ID: <87k7rayvjy.fsf@andreas.org> "Douglas Michael Auclair" writes: > (sort of OT) Also, the latest source updates stopped in February The location of the FTP server on the directory changed, and I didn't update the script. Fixed. Andreas -- "In my eyes it is never a crime to steal knowledge. It is a good theft. The pirate of knowledge is a good pirate." (Michel Serres) From andreas at andreas.org Sun Apr 14 20:43:52 2002 From: andreas at andreas.org (Andreas Bogk) Date: 14 Apr 2002 18:43:52 +0000 Subject: each-subclass initializers! In-Reply-To: <200204122111.XAA28091@lucky.symbiose.com> References: <200204122111.XAA28091@lucky.symbiose.com> Message-ID: <87g01yyvbr.fsf@andreas.org> "Gabor Greif" writes: > Finally I have the feeling each-subclass stuff begins to work out. Wohoo! Andreas -- "In my eyes it is never a crime to steal knowledge. It is a good theft. The pirate of knowledge is a good pirate." (Michel Serres) From bfulgham at debian.org Mon Apr 15 07:10:16 2002 From: bfulgham at debian.org (Brent A. Fulgham) Date: Sun, 14 Apr 2002 22:10:16 -0700 Subject: Testing Framework Message-ID: <20020415051016.GA20547@pacbell.net> As you may know, the fd-sources code that Andreas added to the archive a month or so ago contains a very useful test framework. I think it would be very helpful if we made use of it in GD so we could have tighter control over the effects of source changes, and much-needed confidence that updates did not break anything. I propose we start making use of it as soon as possible. Andreas has already gotten it building on GD, and my simple testing seems to indicate that his modifications have not broken anything substantial. FD's use of the framework provides a 'test' directory in each library's source directory. GD currently has a single test directory. I don't much care which way we go, but I'd like to start making some changes to the repository, and I'd like to get your buy-in as to which route we take. I'm somewhat inclined to put the tests down in the source directories, to keep them close to the implementations. Does anyone have any strong feelings one way or the other? Thanks, -Brent From bruce at hoult.org Mon Apr 15 07:23:15 2002 From: bruce at hoult.org (Bruce Hoult) Date: Mon, 15 Apr 2002 17:23:15 +1200 Subject: Testing Framework In-Reply-To: <20020415051016.GA20547@pacbell.net> References: <20020415051016.GA20547@pacbell.net> Message-ID: At 10:10 PM -0700 14/4/02, Brent A. Fulgham wrote: >FD's use of the framework provides a 'test' directory in each library's >source directory. GD currently has a single test directory. I don't >much care which way we go, but I'd like to start making some changes to >the repository, and I'd like to get your buy-in as to which route we >take. We need a bit of both. Unit tests should go in the source directory for that library. But we also need sanity tests of the compiler as a whole (the existing test directory, pretty much), and also we should permanently keep a test case for each bug that is reported and fixed. It would also be good if all these tests could be built either as a single executable (much faster and less disk space wasted) or as individual tests (for working on bug fixes, and for ones that are so bad they actually crash the compiler and/or runtime). -- Bruce From housel at acm.org Mon Apr 15 07:26:32 2002 From: housel at acm.org (Peter S. Housel) Date: 14 Apr 2002 23:26:32 -0600 Subject: RFC: Melange long long parameter hack, Was Re: Melange and long longs In-Reply-To: References: Message-ID: <1018848392.5163.16.camel@ip68-4-88-233.oc.oc.cox.net> On Sat, 2002-04-13 at 14:12, Rob Myers wrote: > On Friday, April 12, 2002, at 10:40 PM, Rob Myers wrote: > > > Is melange meant to handle long longs OK? Particularly long long > > function parameters. I think this may be causing a problem in parsing > > the Carbon headers. > > Melange doesn't handle long long parameters. It seems to be happy with > long long pointer parameters and long longs in structs, > > I'd like to modify melange to handle long longs as longs: > > C: long long fun(long long a); > => > Dylan: call-out("fun", long:, long: a); > > This is obviously broken, but the MacOS Carbon headers have several > functions that take long long parameters, and I'd like to be able to > handle them in some way. I'll add a comment that this should be fixed > when call-out can handle long-long: parameters. > > Please can you comment on this proposal. If anyone has a better proposal > I'd very much like to hear it. :-). I think there is some machinery in melange to handle long long, it's just that there's no type for it to map to. I don't think the long long => long hack will carry us very far, either. We need to have a native type for that in the compiler. I'm sure it wouldn't take too long to do, and in fact it probably needs to be done anyway for C-FFI, but for the next month or two I'm swamped with Ph.D. research work. -Peter- From bruce at hoult.org Mon Apr 15 07:48:36 2002 From: bruce at hoult.org (Bruce Hoult) Date: Mon, 15 Apr 2002 17:48:36 +1200 Subject: RFC: Melange long long parameter hack, Was Re: Melange and long longs In-Reply-To: <1018848392.5163.16.camel@ip68-4-88-233.oc.oc.cox.net> References: <1018848392.5163.16.camel@ip68-4-88-233.oc.oc.cox.net> Message-ID: At 11:26 PM -0600 14/4/02, Peter S. Housel wrote: >On Sat, 2002-04-13 at 14:12, Rob Myers wrote: >> On Friday, April 12, 2002, at 10:40 PM, Rob Myers wrote: >> >> > Is melange meant to handle long longs OK? Particularly long long >> > function parameters. I think this may be causing a problem in parsing >> > the Carbon headers. >> >> Melange doesn't handle long long parameters. It seems to be happy with >> long long pointer parameters and long longs in structs, >> >> I'd like to modify melange to handle long longs as longs: >> >> C: long long fun(long long a); >> => >> Dylan: call-out("fun", long:, long: a); >> >> This is obviously broken, but the MacOS Carbon headers have several >> functions that take long long parameters, and I'd like to be able to >> handle them in some way. I'll add a comment that this should be fixed >> when call-out can handle long-long: parameters. >> >> Please can you comment on this proposal. If anyone has a better proposal >> I'd very much like to hear it. :-). > >I think there is some machinery in melange to handle long long, it's >just that there's no type for it to map to. We don't necessarily need a type for it to map to. Melange knows how many bytes it takes in a struct. That's easy. Then you need *some* way to access the value. A pair of accessors such as HiWord and LoWord would be fine. Any Dylan programs that care can then use to manipulate long long values. -- Bruce From gabor at mac.com Mon Apr 15 20:46:31 2002 From: gabor at mac.com (Gabor Greif) Date: Mon, 15 Apr 2002 20:46:31 +0200 Subject: Mailing list archive unaccessible Message-ID: <200204151845.UAA06295@lucky.symbiose.com> When I tried today, the newer articles were still inaccessible :-( Gabor ---------- >From: Andreas Bogk >To: "gd-hackers at gwydiondylan.org" >Subject: Re: Mailing list archive unaccessible >Date: Son, 14. Apr 2002 20:32 Uhr > > Gabor Greif writes: > >> You don't have permission to access >> /list-archive/gd-hackers/msg04613.html on this server. > > Hm, *some* files had weird permissions. And the mhonarc database is > broken, and fails to add new files. *sigh* > > And since orbz.org is offline, spam is taking over rapidly. > > I'll try to do something as soon as I can. > > Andreas > > -- > "In my eyes it is never a crime to steal knowledge. It is a good > theft. The pirate of knowledge is a good pirate." > (Michel Serres) From gabor at mac.com Mon Apr 15 23:42:57 2002 From: gabor at mac.com (Gabor Greif) Date: Mon, 15 Apr 2002 23:42:57 +0200 Subject: each-subclass initializers! Message-ID: <200204152142.XAA07745@lucky.symbiose.com> The current CVS version seems to run the Airport program from "Dylan Programming" :-) Gabor PS: I found a bunch of bugs with inherited slots, also with instance slots. Incrementally eradicating them in the next days. ---------- >From: Andreas Bogk >To: gd-hackers at gwydiondylan.org >Subject: Re: each-subclass initializers! >Date: Son, 14. Apr 2002 20:43 Uhr > > "Gabor Greif" writes: > >> Finally I have the feeling each-subclass stuff begins to work out. > > Wohoo! > > Andreas > > -- > "In my eyes it is never a crime to steal knowledge. It is a good > theft. The pirate of knowledge is a good pirate." > (Michel Serres) From dauclair at hotmail.com Tue Apr 16 05:01:41 2002 From: dauclair at hotmail.com (Douglas Michael Auclair) Date: Tue, 16 Apr 2002 05:01:41 Subject: each-subclass initializers! Message-ID: >From: "Gabor Greif" >The current CVS version seems to run the Airport program from "Dylan >Programming" :-) Dear Gabor, *Wow* That's amazing! So: define abstract class () // ... each-subclass slot standard-size :: ; end; // ... define class () // ... inherited slot standard-size, init-value: make(, length: 30, width: 10, height: 10); end; works? Included the inherited each-subclass slot with an initial value? If that's so, you've given me everything I need to transform the xml-parser to work with each-subclass slots. Thankyouthankyouthankyou! Now all I need is a working d2c on a working computer here to try it out. I'll give my 200 MHz x86-solaris box a go. Sincerely, Doug Auclair P.S. I just scanned pp. 57-71 of the DRM, but I didn't see anywhere a description of using \= as a slot initializer. Is it equivalent to the init-value: keyword? Could I write slot foo = 5; instead of slot foo, init-value: 5; with no change in semantics/meaning? Could I use \= in the inherited slot (in the body of this email) instead of the init-value: keyword? _________________________________________________________________ Send and receive Hotmail on your mobile device: http://mobile.msn.com From gabor at mac.com Tue Apr 16 08:55:30 2002 From: gabor at mac.com (Gabor Greif) Date: Tue, 16 Apr 2002 08:55:30 +0200 Subject: each-subclass initializers! Message-ID: <200204160654.IAA23116@lucky.symbiose.com> >From: "Douglas Michael Auclair" >To: gabor at mac.com, gd-hackers at gwydiondylan.org >Subject: Re: each-subclass initializers! >Date: Die, 16. Apr 2002 7:01 Uhr > > > > >>From: "Gabor Greif" >>The current CVS version seems to run the Airport program from "Dylan >>Programming" :-) > > Dear Gabor, > *Wow* That's amazing! So: > > define abstract class () > // ... > each-subclass slot standard-size :: ; > end; > > // ... > > define class () > // ... > inherited slot standard-size, > init-value: make(, length: 30, width: 10, height: 10); > end; > > works? Included the inherited each-subclass slot with an initial value? If Sure! (Modulo bugs I mean... but I squashed a dozen of them lately so just tell if you find one). > that's so, you've given me everything I need to transform the xml-parser to > work with each-subclass slots. Thankyouthankyouthankyou! Now all I need is Start right now. > a working d2c on a working computer here to try it out. I'll give my 200 > MHz x86-solaris box a go. > > Sincerely, > Doug Auclair > > P.S. I just scanned pp. 57-71 of the DRM, but I didn't see anywhere a > description of using \= as a slot initializer. Is it equivalent to the It was added to the language when DRM (these chapters) was almost completely written. > init-value: keyword? Could I write > > slot foo = 5; > > instead of > > slot foo, init-value: 5; > > with no change in semantics/meaning? Could I use \= in the inherited slot > (in the body of this email) instead of the init-value: keyword? In this particular example: yes. The transformation you get is slot foo = EXPRESSION; ----- > slot foo, init-function: method() EXPRESSION end; but d2c is smart enough to perform it in the way you suggest if EXPRESSION is "harmless" enough. I.e. literal constant I think. Hope this helps :-) Btw. "Dylan Programming" is much more explicit on them. Gabor > > _________________________________________________________________ > Send and receive Hotmail on your mobile device: http://mobile.msn.com > From andreas at andreas.org Tue Apr 16 11:06:12 2002 From: andreas at andreas.org (Andreas Bogk) Date: 16 Apr 2002 09:06:12 +0000 Subject: Mailing list archive unaccessible In-Reply-To: <200204151845.UAA06295@lucky.symbiose.com> References: <200204151845.UAA06295@lucky.symbiose.com> Message-ID: <871ydgm2rf.fsf@andreas.org> "Gabor Greif" writes: > When I tried today, the newer articles were still inaccessible :-( Hm, looks like the UMASK of the process archiving the mails changed. Now where do I fix this... *scratch* Good news is that it seems we have less of a problem with spam now. The combination of ordb.org and spamcop.net seems to help. Andreas -- "In my eyes it is never a crime to steal knowledge. It is a good theft. The pirate of knowledge is a good pirate." (Michel Serres) From dauclair at hotmail.com Wed Apr 17 00:12:43 2002 From: dauclair at hotmail.com (Douglas Michael Auclair) Date: Tue, 16 Apr 2002 22:12:43 +0000 Subject: Mindy weirdness with latest sources Message-ID: Hey, all, with the very latest sources from the repository on x86-solaris-gcc I get the following error from mindy running the tests: DYLANPATH=../common/transcendental:../mindy/libraries/dylan:../common/collection-extensions:../common/table-ext:../common/string-ext:../common/standard-io:../common/streams:../mindy/libraries/random:../common/print:../common/format:../common/matrix::../common/format-out:../common/time:../common/regular-expressions:../common/getopt; export DYLANPATH ; /export/home/dauclair/dylan/gd/src/mindy/interp/mindy -f ./format-out-test-lib.dbc Warning: the following variables are undefined: in library Dylan: in module Extensions: invoke-debugger [/export/home/dauclair/dylan/gd/src/mindy/libraries/dylan/ext.dylan, line 90] (but then the test runs fine) Anybody else have this problem on other/same platforms? Sincerely, Doug Auclair _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp. From gabor at mac.com Wed Apr 17 00:21:51 2002 From: gabor at mac.com (Gabor Greif) Date: Wed, 17 Apr 2002 00:21:51 +0200 Subject: Mindy weirdness with latest sources Message-ID: <200204162221.AAA10990@lucky.symbiose.com> Looks like I forgot to check in some file :-( Maybe tomorrow... Gabor ---------- >From: "Douglas Michael Auclair" >To: gd-hackers at gwydiondylan.org >Subject: Mindy weirdness with latest sources >Date: Mit, 17. Apr 2002 0:12 Uhr > > Hey, all, > > with the very latest sources from the repository on x86-solaris-gcc I get > the following error from mindy running the tests: > > DYLANPATH=../common/transcendental:../mindy/libraries/dylan:../common/collec tion-exte > nsions:../common/table-ext:../common/string-ext:../common/standard-io:../com mon/strea > ms:../mindy/libraries/random:../common/print:../common/format:../common/matr ix::../co > mmon/format-out:../common/time:../common/regular-expressions:../common/getopt; > export DYLANPATH ; /export/home/dauclair/dylan/gd/src/mindy/interp/mindy -f > ./format-out-test-lib.dbc > Warning: the following variables are undefined: > in library Dylan: > in module Extensions: > invoke-debugger > [/export/home/dauclair/dylan/gd/src/mindy/libraries/dylan/ext.dylan, line > 90] > > (but then the test runs fine) > > Anybody else have this problem on other/same platforms? > > Sincerely, > Doug Auclair > > _________________________________________________________________ > Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp. > From dauclair at hotmail.com Wed Apr 17 04:47:47 2002 From: dauclair at hotmail.com (Douglas Michael Auclair) Date: Wed, 17 Apr 2002 04:47:47 Subject: ESC-slots working great! New release? Message-ID: Dear all, I tested out each-subclass slots with initialization on the xml-parser (in file gd/examples/xml-parser/interface.dylan, specifically): define abstract class () constant each-subclass slot after-open = ""; constant each-subclass slot before-close = ""; end; // ... define class () inherited slot after-open = "!ENTITY "; end class ; // ... and so forth .. It worked like a charm (OT: also found a logic error in the xml-parser for when entities are eliminated by substitution; fixed it on my local machine). Thanks again, Gabor! As soon as I get my internet connection back on my CVS machine, I'll upload the each-subclass slot changes on the xml-parser. So, the addition of each-subclass slots is pretty significant, isn't it? After Gabor does his clean-ups, should we do a new release? Sincerely, Doug Auclair _________________________________________________________________ Chat with friends online, try MSN Messenger: http://messenger.msn.com From gabor at mac.com Wed Apr 17 08:48:23 2002 From: gabor at mac.com (Gabor Greif) Date: Wed, 17 Apr 2002 08:48:23 +0200 Subject: ESC-slots working great! New release? Message-ID: <200204170647.IAA26274@lucky.symbiose.com> I think it _is_ signifiant, it can comple the most widespread example program "Airport" (next to "Hello world!" :-) finally! So we should bump to pre2 (anyone?) and wait some weeks till we are stabilized to release (to prevent the 2.3.6 debacle). I am going into bug-hunting mode now. (Not as if I weren't into it all the time :-) Cheers, Gabor PS: Andreas, can you mail me privately about the news-adder.phtml again? Lost it... ---------- >From: "Douglas Michael Auclair" >To: gabor at mac.com, gd-hackers at gwydiondylan.org >Subject: ESC-slots working great! New release? >Date: Mit, 17. Apr 2002 6:47 Uhr > > So, the addition of each-subclass slots is pretty significant, isn't it? > After Gabor does his clean-ups, should we do a new release? > > Sincerely, > Doug Auclair From robmyers at mac.com Wed Apr 17 20:39:18 2002 From: robmyers at mac.com (Rob Myers) Date: Wed, 17 Apr 2002 19:39:18 +0100 Subject: RFC: Melange long long parameter hack, Was Re: Melange and long longs In-Reply-To: <1018848392.5163.16.camel@ip68-4-88-233.oc.oc.cox.net> Message-ID: <6CDA27E0-5232-11D6-BF80-000A27A8BAE0@mac.com> On Monday, April 15, 2002, at 06:26 AM, Peter S. Housel wrote: > I think there is some machinery in melange to handle long long, it's > just that there's no type for it to map to. I don't think the long long > => long hack will carry us very far, either. We need to have a native > type for that in the compiler. > > I'm sure it wouldn't take too long to do, and in fact it probably needs > to be done anyway for C-FFI I've changed my approach and used a Perl script to convert all long long typedefs in the Carbon headers to struct long_long. This works fine but melange thinks the structs should be passed by reference. With a couple of other trivial Perl hacks, the Carbon headers are now parseable by melange! - Rob. From bfulgham at debian.org Thu Apr 18 08:04:34 2002 From: bfulgham at debian.org (Brent A. Fulgham) Date: Wed, 17 Apr 2002 23:04:34 -0700 Subject: More Eric Kidd Activities... Message-ID: <3CBE61F2.8040704@debian.org> In case you had not seen yet, the GOO folks have seen fit to release the sources. A quick Debian package can be obtained from http://people.debian.org/~bfulgham/goo. Enjoy! Goo is pretty fun! -Brent From andreas at andreas.org Thu Apr 18 13:25:26 2002 From: andreas at andreas.org (Andreas Bogk) Date: 18 Apr 2002 11:25:26 +0000 Subject: ESC-slots working great! New release? In-Reply-To: <200204170647.IAA26274@lucky.symbiose.com> References: <200204170647.IAA26274@lucky.symbiose.com> Message-ID: <87elhdi6zd.fsf@andreas.org> "Gabor Greif" writes: > So we should bump to pre2 (anyone?) and wait some weeks till we are > stabilized to release (to prevent the 2.3.6 debacle). Good plan. I suggest feature-freezing the compiler, we can still work on Common Dylan compliance. Andreas -- "In my eyes it is never a crime to steal knowledge. It is a good theft. The pirate of knowledge is a good pirate." (Michel Serres) From gabor at mac.com Thu Apr 18 21:42:45 2002 From: gabor at mac.com (Gabor Greif) Date: Thu, 18 Apr 2002 21:42:45 +0200 Subject: CVS: dauclair Message-ID: <200204182308.BAA20352@lucky.symbiose.com> XSL... sounds like a technology capable of removing the unwanted elements from gd/src/d2c/runtime/Macintosh/d2c-runtime-CW.mcp.xml ... Is this true? And if yes, how do I do it? Thanks in advance, Gabor ---------- >From: pubcvs at berlin.ccc.de >To: gd-chatter at gwydiondylan.org, update at gwydiondylan.org >Subject: CVS: dauclair >Date: Don, 18. Apr 2002 11:37 Uhr > > dauclair > gd/examples/xml-parser interface.dylan library.dylan printing.dylan > productions.dylan transform.dylan > Thu Apr 18 11:37:50 CEST 2002 > Update of /home/cvsroot/gd/examples/xml-parser > In directory berlin.ccc.de:/var/tmp/cvs-serv84047 > > Modified Files: > interface.dylan library.dylan printing.dylan productions.dylan > transform.dylan > Log Message: > > Changed so that now it also holds processing-instructions and > comments. With that change, I needed to add a new gf: prepare-document > because before-transform on is now called multiple times. > > I've also tested the xml-parser and xml-test on multiple xsl documents ... > works like a charm. > From dauclair at hotmail.com Fri Apr 19 10:30:32 2002 From: dauclair at hotmail.com (Douglas Michael Auclair) Date: Fri, 19 Apr 2002 08:30:32 +0000 Subject: XSL and the xml-parser lib Message-ID: >From: "Gabor Greif" >XSL... sounds like a technology capable of removing the unwanted elements >from gd/src/d2c/runtime/Macintosh/d2c-runtime-CW.mcp.xml ... > >Is this true? And if yes, how do I do it? Dear Gabor, Hehehe! You are asking a teacher a question at the end of the lecture! Yes, XSL can remove elements from an XML doc. XML to XML transformations are easy with XSL. I'm finishing a tutorial on just that topic using xstlproc, an XSL tool available from http://www.libxml.org/XSLT. The tutorial is at http://www.geocities.com/xml_pages/movies/layered.html. Of course, to learn XSL is to learn a whole new language. Instead you can use the Dylan xml-parser lib with a very simple xform program. Let's say, for example, you want to eliminate entire 's from the CW XML document. The program would be: module: xformer-for-Gabor define class () constant slot remove-groups :: , required-init-keyword: remove:; end; // for pretty-printing define method before-transform(xml :: , state :: , depth :: , stream :: ) format(stream, "\n%s", make(, size: depth, fill: ' ')); end method before-transform; define method transform(elt :: , tag-name == #"group", state :: , stream :: ) unless(member?(elt["name"].text, state.remove-groups, test: \=)) next-method(); end unless; end method transform; define function main(app-name, app-args) => () let remover = make(, remove: copy-sequence(app-args, start: 1)); with-open-file(in = app-args[0], direction: #"input-output") transform-document(parse-document(stream-contents(in, clear-contents?: #f), state: remover); end with-open-file; end function main; main(application-name(), application-arguments()); when called with $ xformer-for-gabor gd/src/d2c/runtime/Macintosh/d2c-runtime-CW.mcp.xml gc > res.xml writes the new XML file (without a gc ) to res.xml (BTW that xml file is very, very big!). You need the following libraries (all located at gd/examples): anaphora, multimap, meta and xml-parser ... the xml-parser needs a d2c that compiles each-subclass slots, and I've made some very recent changes to the xml-parser library 1) to ensure capitalization of tags is preserved, and 2) to improve the printing facility. You'll note in the transformed document (res.xml), that there's no DTD ... if CW complains, you can copy and paste the 40 or so lines into res.xml (one of the reasons for the delay in my answer is that I started to implement DTD's containing element-decls, etc ... it was a bit more than I could manage for right now). Sincerely, Doug Auclair _________________________________________________________________ Chat with friends online, try MSN Messenger: http://messenger.msn.com From dauclair at hotmail.com Tue Apr 23 08:40:37 2002 From: dauclair at hotmail.com (Douglas Michael Auclair) Date: Tue, 23 Apr 2002 06:40:37 +0000 Subject: Problem compiling with each-subclass d2c Message-ID: Dear Gabor and all, I'm having a problem compiling an executable that uses a library (xml-parser) with each-subclass slots. The compiler quits when it reaches a file that has a class that inherits from a class in the library that has each-subclass slots with the following message: Accessing uninitialized slot associated-meta-slot in { instance} The simplest way to reproduce this problem is to compile the following files: // foo.lid executable: foo library: foo files: library.dylan foo.dylan // library.dylan is the usual import and "use xml-parser;" in the library and module declarations // foo.dylan Module: foo define class () end; with: d2c -L ../anaphora -L ../multimap -L ../meta -L ../xml-parser foo.dylan Any ideas what's happening? Sincerely, Doug Auclair _________________________________________________________________ Join the world?s largest e-mail service with MSN Hotmail. http://www.hotmail.com From gabor at mac.com Tue Apr 23 21:52:40 2002 From: gabor at mac.com (Gabor Greif) Date: Tue, 23 Apr 2002 21:52:40 +0200 Subject: Problem compiling with each-subclass d2c Message-ID: <200204231951.VAA11725@lucky.symbiose.com> Yeah, they are probably not correctly externalized into the .du file. (Or read in.) Looking at it. Gabor ---------- >From: "Douglas Michael Auclair" >To: gd-hackers at gwydiondylan.org, gabor at mac.com >Subject: Problem compiling with each-subclass d2c >Date: Die, 23. Apr 2002 8:40 Uhr > > Any ideas what's happening? > > Sincerely, > Doug Auclair From dauclair at hotmail.com Tue Apr 23 22:15:29 2002 From: dauclair at hotmail.com (Douglas Michael Auclair) Date: Tue, 23 Apr 2002 20:15:29 +0000 Subject: Problem compiling with each-subclass d2c Message-ID: >From: "Gabor Greif" >Yeah, they are probably not correctly externalized into the .du file. (Or >read in.) Looking at it. Dear Gabor, Thanks! I must say, the problem I've encountered is an obscure one, because: // lib define abstract class (