3.2rc on MKCL (Was: ASDF 3.1.7.43 is 3.2 release candidate)
On Thu, Dec 15, 2016 at 7:51 PM, Faré <fahree@gmail.com> wrote:
"Our greatest release ever!"
:-) I just tried ASDF master head on my latest MKCL with this small (attached) edit and I got this as result: -#--------------------------------------- Using ../../../bin/mkcl Ran 63 tests: 63 passing and 0 failing all tests apparently successful -#--------------------------------------- This is on Ubuntu Linux x86_64. So I guess it looks like good to go on that configuration. I will test on Win32 and Win64 soon. Cheers, JCB
On 12/16/16 Dec 16 -1:58 AM, Jean-Claude Beaudoin wrote:
On Thu, Dec 15, 2016 at 7:51 PM, Faré <fahree@gmail.com <mailto:fahree@gmail.com>> wrote:
"Our greatest release ever!"
:-)
I just tried ASDF master head on my latest MKCL with this small (attached) edit and I got this as result:
-#--------------------------------------- Using ../../../bin/mkcl Ran 63 tests: 63 passing and 0 failing all tests apparently successful -#---------------------------------------
This is on Ubuntu Linux x86_64. So I guess it looks like good to go on that configuration. I will test on Win32 and Win64 soon.
Just to check: this means that MKCL now *can* fmakunbound a SETF method? Cheers, r
On Fri, Dec 16, 2016 at 12:15 PM, Robert Goldman <rpgoldman@sift.net> wrote:
Just to check: this means that MKCL now *can* fmakunbound a SETF method?
I have not been able to reproduce this presumed "fmakunbound on SETF" problem on any MKCL 1.1.X (from X=0 onward). So I don't know where this issue came from. BTW, the current ASDF code (with this edit) does not compile on MKCL 1.1.6 or before due to various other bugs internal to MKCL itself; it compiles correctly on MKCL 1.1.8 or later. (There is no MKCL 1.1.7).
On 12/16/16 Dec 16 -4:16 PM, Jean-Claude Beaudoin wrote:
On Fri, Dec 16, 2016 at 12:15 PM, Robert Goldman <rpgoldman@sift.net <mailto:rpgoldman@sift.net>> wrote:
Just to check: this means that MKCL now *can* fmakunbound a SETF method?
I have not been able to reproduce this presumed "fmakunbound on SETF" problem on any MKCL 1.1.X (from X=0 onward). So I don't know where this issue came from.
BTW, the current ASDF code (with this edit) does not compile on MKCL 1.1.6 or before due to various other bugs internal to MKCL itself; it compiles correctly on MKCL 1.1.8 or later. (There is no MKCL 1.1.7).
Thanks! I merged your patch -- well actually, I just removed the line, instead of commenting it out as you did -- so it will be in 3.2 when released. Thanks again for testing. best, r
Testing a fresh ASDF repo clone (master head) with latest MKCL on Vista (Win32) and Windows 7 (Win64) both result in only 2 failures: test-program.script test-run-program.script Will investigate shortly...
On 12/17/16 Dec 17 -4:59 AM, Jean-Claude Beaudoin wrote:
Testing a fresh ASDF repo clone (master head) with latest MKCL on Vista (Win32) and Windows 7 (Win64) both result in only 2 failures:
test-program.script test-run-program.script
Will investigate shortly...
Thanks. Please keep us posted. I don't have a Windows tool chain, so while I can test on Windows, I can't build lisp implementations there. Best, r
On Sat, Dec 17, 2016 at 4:44 PM, Robert Goldman <rpgoldman@sift.net> wrote:
On 12/17/16 Dec 17 -4:59 AM, Jean-Claude Beaudoin wrote:
Testing a fresh ASDF repo clone (master head) with latest MKCL on Vista (Win32) and Windows 7 (Win64) both result in only 2 failures:
test-program.script test-run-program.script
Will investigate shortly...
Thanks. Please keep us posted.
Here is what I came up with: 1) test-program.script fails trying to find a file named .../build/fasls/mkcl-1.1.9-xxx/asdf/test/hello-world- example--all-systems.image (where ... and xxx are place holders you can easily substitute) what exists instead is .../build/fasls/mkcl-1.1.9-xxx/asdf/test/hello-world- example--all-systems.exe which seems to be the expected executable corresponding to the ???--all-systems.image on unix. (BTW, why is there such an "image" extension on unix/linux?) 2) test-run-program.script .../uiop/launch-program.lisp is incomplete for MKCL on windows and needs to be patched as per attached file to be in any way functional in that context. Once patched accordingly and applied to very latest MKCL repo master head this test now passes. Cheers, JCB
On Mon, Dec 19, 2016 at 2:32 AM, Jean-Claude Beaudoin <jean.claude.beaudoin@gmail.com> wrote:
On Sat, Dec 17, 2016 at 4:44 PM, Robert Goldman <rpgoldman@sift.net> wrote:
On 12/17/16 Dec 17 -4:59 AM, Jean-Claude Beaudoin wrote:
Testing a fresh ASDF repo clone (master head) with latest MKCL on Vista (Win32) and Windows 7 (Win64) both result in only 2 failures:
test-program.script test-run-program.script
Will investigate shortly...
Thanks. Please keep us posted.
Here is what I came up with:
1) test-program.script fails trying to find a file named
.../build/fasls/mkcl-1.1.9-xxx/asdf/test/hello-world-example--all-systems.image
(where ... and xxx are place holders you can easily substitute)
what exists instead is
.../build/fasls/mkcl-1.1.9-xxx/asdf/test/hello-world-example--all-systems.exe
which seems to be the expected executable corresponding to the ???--all-systems.image on unix. (BTW, why is there such an "image" extension on unix/linux?)
The "image" extension allows to distinguish an image meant for continued development from an application that has its own entry-point. If it looks for the wrong file, there's a bug in my code. Sigh.
2) test-run-program.script
.../uiop/launch-program.lisp is incomplete for MKCL on windows and needs to be patched as per attached file to be in any way functional in that context. Once patched accordingly and applied to very latest MKCL repo master head this test now passes.
+ #+(or mkcl) (list "cmd" (strcat "/c " command)) The list seems very wrong, unless MKCL doesn't quote its arguments to the win32 spawn function, which is probably wrong in its own right. Can you confirm how to execute a raw CMD command line with MKCL? —♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org When a pamphlet was published entitled "100 Authors Against Einstein", Einstein retorted "If I were wrong, one would be enough."
On Mon, Dec 19, 2016 at 3:36 AM, Faré <fahree@gmail.com> wrote:
On Mon, Dec 19, 2016 at 2:32 AM, Jean-Claude Beaudoin <jean.claude.beaudoin@gmail.com> wrote:
(...)
2) test-run-program.script
.../uiop/launch-program.lisp is incomplete for MKCL on windows and needs to be patched as per attached file to be in any way functional in that context. Once patched accordingly and applied to very latest MKCL repo master head this test now passes.
+ #+(or mkcl) (list "cmd" (strcat "/c " command)) The list seems very wrong, unless MKCL doesn't quote its arguments to the win32 spawn function, which is probably wrong in its own right.
You lost me entirely here. It does look like something that works pretty well for being so wrong... No "spawn" function involved here, just a very native and basic Win32 CreateProcessW() call.
Can you confirm how to execute a raw CMD command line with MKCL?
Are you looking for something like (mkcl:run-program "cmd" '("/c echo ok 1") :output t) or (mkcl:run-program "cmd" '("/c" "echo" "ok" "1") :output t) both produce the same output (sent to stdout here by the ":output t" pair, adjust to your need). Any slicing or concatenation of the arguments produce the same result.
Oh. Is it the case that you don't even try to properly escape commands as with uiop:escape-windows-command or GetCommandLineW? Grepping your source code, it looks like it has a lot of bogus escaping, on Unix as well as on Windows :-( http://msdn.microsoft.com/en-us/library/bb776391(v=vs.85).aspxhttp://msdn.mi... mkcl needs an interface that works, and then asdf can use it. —♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org My knowledge is finite. My ignorance is infinite. — Faré On Mon, Dec 19, 2016 at 7:08 AM, Jean-Claude Beaudoin <jean.claude.beaudoin@gmail.com> wrote:
On Mon, Dec 19, 2016 at 3:36 AM, Faré <fahree@gmail.com> wrote:
On Mon, Dec 19, 2016 at 2:32 AM, Jean-Claude Beaudoin <jean.claude.beaudoin@gmail.com> wrote:
(...)
2) test-run-program.script
.../uiop/launch-program.lisp is incomplete for MKCL on windows and needs to be patched as per attached file to be in any way functional in that context. Once patched accordingly and applied to very latest MKCL repo master head this test now passes.
+ #+(or mkcl) (list "cmd" (strcat "/c " command)) The list seems very wrong, unless MKCL doesn't quote its arguments to the win32 spawn function, which is probably wrong in its own right.
You lost me entirely here. It does look like something that works pretty well for being so wrong...
No "spawn" function involved here, just a very native and basic Win32 CreateProcessW() call.
Can you confirm how to execute a raw CMD command line with MKCL?
Are you looking for something like
(mkcl:run-program "cmd" '("/c echo ok 1") :output t) or (mkcl:run-program "cmd" '("/c" "echo" "ok" "1") :output t)
both produce the same output (sent to stdout here by the ":output t" pair, adjust to your need). Any slicing or concatenation of the arguments produce the same result.
participants (3)
-
Faré -
Jean-Claude Beaudoin -
Robert Goldman