[armedbear-devel] translate-logical-pathname and :wild-inferiors regression

[Alan Ruttenberg]

In 12389 just checked out and compiled just now.

CL-USER(4): (setf (logical-pathname-translations "ido")
'((#P"IDO:IDO-CORE;**;*.*"
#P"/Users/alanr/repos/infectious-disease-ontology/trunk/src/ontology/ido-core/**/*.*")
(#P"IDO:IMMUNOLOGY;**;*.*"
#P"/Users/alanr/repos/infectious-disease-ontology/trunk/src/ontology/immunology/**/*.*")
(#P"IDO:TOOLS;**;*.*"
#P"/Users/alanr/repos/infectious-disease-ontology/trunk/src/tools/**/*.*")
(#P"IDO:LIB;**;*.*"
#P"/Users/alanr/repos/infectious-disease-ontology/trunk/lib/**/*.*")))


CL-USER(5): (TRANSLATE-PATHNAME #P"IDO:IMMUNOLOGY;"
#P"IDO:IMMUNOLOGY;**;*.*"
#P"/Users/alanr/repos/infectious-disease-ontology/trunk/src/ontology/**/*.*")
Debugger invoked on condition of type SIMPLE-ERROR:
  Unsupported case in TRANSLATE-DIRECTORY-COMPONENTS.
Restarts:
  0: TOP-LEVEL Return to top level.

I patched it with the following, but don't know if it's the right thing


    (defun translate-directory-components-aux (src from to case)
      (cond
((and (null src) (null from) (null to))
 NIL)
((and to
      (not (member (car to) '(:wild :wild-inferiors))))
 (cons (casify (car to) case)
       (translate-directory-components-aux src from (cdr to) case)))
-->
((and (not src) (eq (car from) :wild-inferiors) (eq (car to)
:wild-inferiors))
 (translate-directory-components-aux src (cdr from) (cdr to) case))
-->
((not (and src from))
 ;; both are NIL --> TO is a wildcard which can't be matched
 ;; either is NIL --> SRC can't be fully matched against FROM, vice versa
 (throw 'failed-match))
...

I concur btw that having a function that generates a report to include for
the developers would be a good thing.
Svn version, possibly the svn diff wrt to that, java version, etc.

-Alan

On Mon, Jan 18, 2010 at 5:12 AM, Mark Evenson <evenson at panix.com> wrote:

> On 1/17/10 11:09 PM, Alan Ruttenberg wrote:
> > A recent commit (dec 2009) seems to have changed the behavior of
> > wild-inferiors.
> > Prior to the change :wild-inferiors would match 0 or more pathname
> > components. After the change 1 or more.
> >
> > e.g.
> >
> > (TRANSLATE-PATHNAME #P"IDO:IMMUNOLOGY;" #P"IDO:IMMUNOLOGY;**;*.*"
> >
> #P"/Users/alanr/repos/infectious-disease-ontology/trunk/src/ontology/**/*.*")
> >
> > Used to work, but now fails.
> >
> >      Unsupported case in TRANSLATE-DIRECTORY-COMPONENTS.
>
> In both the released abcl-0.18.0 and trunk (as of r12383), I don't see
> this bug. getting
>
> CL-USER> (TRANSLATE-PATHNAME #P"IDO:IMMUNOLOGY;"
> #P"IDO:IMMUNOLOGY;**;*.*"
>
> #P"/Users/alanr/repos/infectious-disease-ontology/trunk/src/ontology/**/*.*")
>
> #P"/Users/alanr/repos/infectious-disease-ontology/trunk/src/ontology/**/IDO:IMMUNOLOGY;"
>
> Could you please provide more information about which version of ABCL
> has this problem?
>
> Note to developers:  the task of determining which version of ABCL
> people report problems to us would be simplified if we included better
> versioning wrt. SVN.  There is a "version.src" Ant property that is
> stored in the abcl.jar file to keep track of versions, but when I last
> looked into it I never found a reasonable way to set it from Ant, as
> there is no built-in support in Ant for Subversion.  If we could depend
> on the existence of a 'svn' client binary in the PATH, we could parse
> its output, but this seems like a bad bet under Windows as a sizable
> number of people presumably use Explorer Shell extensions like
> TortoiseSVN.  It looks like a possible route would be to parse the
> contents of '.svn/entries' if it exists to take a stab at reporting
> which SVN version abcl.jar was built from.  Without access to an SVN
> client we couldn't tell reliably if the source has been modified from
> that SVN version, but it would give us a decent place to start.
>
> --
> "A screaming comes across the sky.  It has happened before, but there
> is nothing to compare to it now."
>
> _______________________________________________
> armedbear-devel mailing list
> armedbear-devel at common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/armedbear-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/armedbear-devel/attachments/20100118/528f676b/attachment.html>