I tagged ASDF 1.636. * Can some ABCL user try it? It doesn't have asdf-method-combination and magic asdf:around anymore. * James Anderson, can you re-test it, particularly with allegro? * Juanjo, does it satisfy you wrt output-files protocol?
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The only way to have a friend is to be one. — Ralph Waldo Emerson
most everything now has equivalent results.[1] abcl now loads asdf, but the test fails anomalously. looking...
--- [1] : http://ec2-204-236-204-89.compute-1.amazonaws.com/test/ 20100315T221624/
On 2010-03-15, at 22:53 , Faré wrote:
I tagged ASDF 1.636.
- Can some ABCL user try it? It doesn't have asdf-method-combination
and magic asdf:around anymore.
- James Anderson, can you re-test it, particularly with allegro?
- Juanjo, does it satisfy you wrt output-files protocol?
[ François-René ÐVB Rideau | Reflection&Cybernethics | http:// fare.tunes.org ] The only way to have a friend is to be one. — Ralph Waldo Emerson
asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
most everything now has equivalent results.[1] abcl now loads asdf, but the test fails anomalously. looking...
abcl fails in connection with a make-pathname operation of the sort
(make-pathname :directory '(:relative) :name "file" :type :unspecific :host nil :device nil)
which i suspect is the first component relative pathname from the first file component in the first system definition, and i've not (yet) understood how handler-bind is intended to behave with abcl, so that pretty much aborts the tests.
i also can't say much about the context, as the stack trace is opaque.
anyway...
--- [1] : http://ec2-204-236-204-89.compute-1.amazonaws.com/test/ 20100315T221624/
On 2010-03-15, at 22:53 , Faré wrote:
I tagged ASDF 1.636.
- Can some ABCL user try it? It doesn't have asdf-method-combination
and magic asdf:around anymore.
- James Anderson, can you re-test it, particularly with allegro?
- Juanjo, does it satisfy you wrt output-files protocol?
[ François-René ÐVB Rideau | Reflection&Cybernethics | http:// fare.tunes.org ] The only way to have a friend is to be one. — Ralph Waldo Emerson
asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
most everything now has equivalent results.[1] abcl now loads asdf, but the test fails anomalously. looking...
abcl fails in connection with a make-pathname operation of the sort
(make-pathname :directory '(:relative) :name "file" :type :unspecific :host nil :device nil)
which i suspect is the first component relative pathname from the first file component in the first system definition, and i've not (yet) understood how handler-bind is intended to behave with abcl, so that pretty much aborts the tests.
i also can't say much about the context, as the stack trace is opaque.
anyway...
abcl has a distinct dislike for :unspecific pathname components.
if the 1.636 code for merge-pathnames* is patched for abcl to always use #'ununspecific, abcl constructs all systems and succeeds with the same 1280/2080 directory/file matches as most of the other implementations.
--- [1] : http://ec2-204-236-204-89.compute-1.amazonaws.com/test/ 20100315T221624/
On 2010-03-15, at 22:53 , Faré wrote:
I tagged ASDF 1.636.
- Can some ABCL user try it? It doesn't have asdf-method-combination
and magic asdf:around anymore.
- James Anderson, can you re-test it, particularly with allegro?
- Juanjo, does it satisfy you wrt output-files protocol?
[ François-René ÐVB Rideau | Reflection&Cybernethics | http:// fare.tunes.org ] The only way to have a friend is to be one. — Ralph Waldo Emerson
asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
_______________________________________________ asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
On 15 March 2010 18:10, james anderson james.anderson@setf.de wrote:
abcl has a distinct dislike for :unspecific pathname components.
I have disabled :unspecific for clisp and abcl.
Can you try again with 1.639?
I also have integrated your test-suite into the normal ASDF test suite, at rpg's suggestion, since it works for SBCL and CCL.
Now to identify the cases where it fails for clisp, abcl, allegro & al., and find fixes.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] If mice were the ultimate input device, humans would be born with one arm and three fingers.
On 2010-03-16, at 00:46 , Faré wrote:
On 15 March 2010 18:10, james anderson james.anderson@setf.de wrote:
abcl has a distinct dislike for :unspecific pathname components.
I have disabled :unspecific for clisp and abcl.
Can you try again with 1.639?
with pseudo namestrings : http:// ec2-204-236-204-89.compute-1.amazonaws.com/test/20100315T235855.txt
without pseudo namestrings : http:// ec2-204-236-204-89.compute-1.amazonaws.com/test/20100316T000221.txt
I also have integrated your test-suite into the normal ASDF test suite, at rpg's suggestion, since it works for SBCL and CCL.
Now to identify the cases where it fails for clisp, abcl, allegro & al., and find fixes.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http:// fare.tunes.org ] If mice were the ultimate input device, humans would be born with one arm and three fingers.
On 15 March 2010 19:14, james anderson james.anderson@setf.de wrote:
On 2010-03-16, at 00:46 , Faré wrote:
On 15 March 2010 18:10, james anderson james.anderson@setf.de wrote:
abcl has a distinct dislike for :unspecific pathname components.
I have disabled :unspecific for clisp and abcl.
Can you try again with 1.639?
with pseudo namestrings : http:// ec2-204-236-204-89.compute-1.amazonaws.com/test/20100315T235855.txt
without pseudo namestrings : http:// ec2-204-236-204-89.compute-1.amazonaws.com/test/20100316T000221.txt
What are pseudo-namestrings?
What are homogeneous failures?
I have the hardest time debugging the least failure with CLISP.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] An apple every eight hours will keep three doctors away.
On 2010-03-16, at 01:26 , Faré wrote:
On 15 March 2010 19:14, james anderson james.anderson@setf.de wrote:
On 2010-03-16, at 00:46 , Faré wrote:
On 15 March 2010 18:10, james anderson james.anderson@setf.de wrote:
abcl has a distinct dislike for :unspecific pathname components.
I have disabled :unspecific for clisp and abcl.
Can you try again with 1.639?
with pseudo namestrings : http:// ec2-204-236-204-89.compute-1.amazonaws.com/test/20100315T235855.txt
without pseudo namestrings : http:// ec2-204-236-204-89.compute-1.amazonaws.com/test/20100316T000221.txt
What are pseudo-namestrings?
this is your term from earlier mail:
On 2010-03-09, at 01:35 , Faré wrote:
if someone would enumerate the cases which are supported. i will try again. as i understand your demurral, relative to the tested enumeration[2], one should eliminate the cases which involve string designators for logical pathnames. should anything else be removed? are that any additional cases for which support should be tested?
Non-logical pathnames are out. Namestrings without #p"..." are out. Pathnames made with (make-pathname ...) are in. /-separated pseudo-namestrings relative paths are in. Tests should include cases where the source-file-type is "lisp", NIL or :directory.
in a separate message, you indicated that a complete absolute pathname namestring should also appear in some test.
What are homogeneous failures?
configurations which are either all logical pathnames or include no logical pathname.[1] at one point, i thought that might (inversely) correlate with failure, but that appears to not be the case.
I have the hardest time debugging the least failure with CLISP.
i have yet to try to understand the causes for anything but the most basic errors. i would like to hear first, whether the test configurations are correct and/or complete.
--- [1] http://github.com/lisp/de.setf.asdf.x/blob/master/test/asdf- pathname-test.lisp#L242
What are pseudo-namestrings?
this is your term from earlier mail:
Oh, OK.
in a separate message, you indicated that a complete absolute pathname namestring should also appear in some test.
Yes, although that would probably only work on Unix, so should be made #+unix or some such.
What are homogeneous failures?
configurations which are either all logical pathnames or include no logical pathname.[1] at one point, i thought that might (inversely) correlate with failure, but that appears to not be the case.
At least I'm glad that SBCL, CCL and LispWorks pass all tests. There is *some* sanity in the world.
I have the hardest time debugging the least failure with CLISP.
i have yet to try to understand the causes for anything but the most basic errors. i would like to hear first, whether the test configurations are correct and/or complete.
Correct, they look like they are. Complete, I'll have to think way more, but I suppose you could add an untype static file just for kicks.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The penalty for laughing in a courtroom is six months in jail; if it were not for this penalty, the jury would never hear the evidence. — H. L. Mencken
On 3/16/10 12:10 AM, james anderson wrote:
most everything now has equivalent results.[1] abcl now loads asdf, but the test fails anomalously. looking...
abcl fails in connection with a make-pathname operation of the sort
(make-pathname :directory '(:relative) :name"file" :type :unspecific :host nil :device nil)
I just fixed the failure of that form in [abcl trunk in r122549][1].
[1]: http://trac.common-lisp.net/armedbear/changeset/12549
[…]
abcl has a distinct dislike for :unspecific pathname components.
if the 1.636 code for merge-pathnames* is patched for abcl to always use #'ununspecific, abcl constructs all systems and succeeds with the same 1280/2080 directory/file matches as most of the other implementations.
I'll try to fix what I can in ABCL's implementation of :UNSPECIFC to get ASDF running.
good morning;
[it appears as if this never went out.]
given the parity when tested with exclusively pathname- instance :pathname arguments, it seemed appropriate to test the consequences when pseudo namestrings are included in the configurations. the pseudo namestring values[0,1,2] follow earlier discussions on this list and are intended to follow the guidelines in the texinfo documentation.[3] despite this, the results are neither uniform nor positive.[4]
--- [0] : http://github.com/lisp/de.setf.asdf.x/blob/master/test/asdf- pathname-test.lisp#L63 [1] : http://github.com/lisp/de.setf.asdf.x/blob/master/test/asdf- pathname-test.lisp#L73 [2] : http://github.com/lisp/de.setf.asdf.x/blob/master/test/asdf- pathname-test.lisp#L101 [3] : nb. of all the various failure modes lurking in these tests, the purported non-portability of parse-namestring seems to be the least likely source. each configuration set contains a dual definition ( make-pathname v/s parse-namestring ), which pair evidently yielded identical results across the implementations which succeeded on all configurations when tested without pseudo namestrings. as it were. [4] http://ec2-204-236-204-89.compute-1.amazonaws.com/test/ 20100315T231205.txt
On 2010-03-15, at 22:53 , Faré wrote:
I tagged ASDF 1.636.
- Can some ABCL user try it? It doesn't have asdf-method-combination
and magic asdf:around anymore.
- James Anderson, can you re-test it, particularly with allegro?
- Juanjo, does it satisfy you wrt output-files protocol?
[ François-René ÐVB Rideau | Reflection&Cybernethics | http:// fare.tunes.org ] The only way to have a friend is to be one. — Ralph Waldo Emerson
asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
_______________________________________________ asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel _______________________________________________ asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
-
Faré -
james anderson -
Mark Evenson