[asdf-devel] question about relative pathname to point to parent directory
What's the right syntax for specifying a relative pathname for a module whose files should live in the parent directory instead of in a subdirectory? E.g., if I have (:module foo :components ((:file "bar") (:file "baz")) :pathname ???? ) what goes in there if I do NOT want the files to live in foo/, but with all the other files? I could put a FAQ about this in the manual... best, r
I'm not sure what you want, but try "" for current directory, ".." for up, etc. ".." is actually not recognized by ASDF but passed to the implementation, and from it to the OS. Paths are relative to the parent's pathname's directory. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] Theists think all gods but theirs are false. Atheists simply don't make an exception for the last one. On 25 May 2010 11:24, Robert Goldman <rpgoldman@sift.info> wrote:
What's the right syntax for specifying a relative pathname for a module whose files should live in the parent directory instead of in a subdirectory?
E.g., if I have
(:module foo :components ((:file "bar") (:file "baz")) :pathname ???? )
what goes in there if I do NOT want the files to live in foo/, but with all the other files?
I could put a FAQ about this in the manual...
best, r
On 5/25/10 May 25 -11:11 AM, Faré wrote:
I'm not sure what you want, but try "" for current directory, ".." for up, etc. ".." is actually not recognized by ASDF but passed to the implementation, and from it to the OS.
I added a test that verifies that "" works as a pathname component. I wasn't sure that it would work for ACL, and wanted to verify for my satisfaction. I'd like to push the new test (no change to asdf.lisp). Is that OK with you? thanks, r
Paths are relative to the parent's pathname's directory.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] Theists think all gods but theirs are false. Atheists simply don't make an exception for the last one.
On 25 May 2010 11:24, Robert Goldman <rpgoldman@sift.info> wrote:
What's the right syntax for specifying a relative pathname for a module whose files should live in the parent directory instead of in a subdirectory?
E.g., if I have
(:module foo :components ((:file "bar") (:file "baz")) :pathname ???? )
what goes in there if I do NOT want the files to live in foo/, but with all the other files?
I could put a FAQ about this in the manual...
best, r
On 26 May 2010 09:18, Robert Goldman <rpgoldman@sift.info> wrote:
On 5/25/10 May 25 -11:11 AM, Faré wrote:
I'm not sure what you want, but try "" for current directory, ".." for up, etc. ".." is actually not recognized by ASDF but passed to the implementation, and from it to the OS.
I added a test that verifies that "" works as a pathname component. I wasn't sure that it would work for ACL, and wanted to verify for my satisfaction.
I'd like to push the new test (no change to asdf.lisp). Is that OK with you?
Yes. Please commit new tests that you have. Pending ASDF 2 release, I'd like to discourage code changes, but to encourage test and documentation improvements. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] If you could ask a unique question to a computer during a Turing test, what would you ask? — Douglas Hofstadter, Metamagical Themas
On 5/26/10 May 26 -9:35 AM, Faré wrote:
On 26 May 2010 09:18, Robert Goldman <rpgoldman@sift.info> wrote:
On 5/25/10 May 25 -11:11 AM, Faré wrote:
I'm not sure what you want, but try "" for current directory, ".." for up, etc. ".." is actually not recognized by ASDF but passed to the implementation, and from it to the OS.
I added a test that verifies that "" works as a pathname component. I wasn't sure that it would work for ACL, and wanted to verify for my satisfaction.
I'd like to push the new test (no change to asdf.lisp). Is that OK with you?
Yes. Please commit new tests that you have.
Pending ASDF 2 release, I'd like to discourage code changes, but to encourage test and documentation improvements.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] If you could ask a unique question to a computer during a Turing test, what would you ask? — Douglas Hofstadter, Metamagical Themas
OK, I have just pushed the new test definition (I'd like to encourage everyone who can to run it and make sure it passes --- as much to test for bugs in the test as in ASDF), and also some mods to the pathnames discussion in the texinfo file. cheers, r
On 5/25/10 May 25 -11:11 AM, Faré wrote:
I'm not sure what you want, but try "" for current directory, ".." for up, etc. ".." is actually not recognized by ASDF but passed to the implementation, and from it to the OS.
Question: if ".." is handled by the OS, does that mean it may not be portable (especially on Windows)? If so, do we need a native-to-ASDF way to specify a (:relative :up) or (:relative :back) pathname with a string? thanks, r
Paths are relative to the parent's pathname's directory.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] Theists think all gods but theirs are false. Atheists simply don't make an exception for the last one.
On 25 May 2010 11:24, Robert Goldman <rpgoldman@sift.info> wrote:
What's the right syntax for specifying a relative pathname for a module whose files should live in the parent directory instead of in a subdirectory?
E.g., if I have
(:module foo :components ((:file "bar") (:file "baz")) :pathname ???? )
what goes in there if I do NOT want the files to live in foo/, but with all the other files?
I could put a FAQ about this in the manual...
best, r
On 26 May 2010 09:20, Robert Goldman <rpgoldman@sift.info> wrote:
On 5/25/10 May 25 -11:11 AM, Faré wrote:
I'm not sure what you want, but try "" for current directory, ".." for up, etc. ".." is actually not recognized by ASDF but passed to the implementation, and from it to the OS.
Question: if ".." is handled by the OS, does that mean it may not be portable (especially on Windows)? If so, do we need a native-to-ASDF way to specify a (:relative :up) or (:relative :back) pathname with a string?
It would be good if ".." were handled specially by ASDF like "" is, but it would have made the code more complex than it already is for a feature that is both more rarely needed, and more portably supported as "..". It would probably be nice to add ASDF-side parsing of ".." in ASDF 2.1, but at this point, I'm wary to change anything not necessary in the code before ASDF 2. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] Funny how the so-called money-disparaging anti-materialist are always so prompt to demand other people's money in the form of taxes.
participants (2)
-
Faré -
Robert Goldman