- asdf-devel - mailman3.common-lisp.net

[asdf-devel] [PROPOSAL] Easier configuration, installation, and removal of asdf systems
by Krzysztof Drewniak 19 Nov '10

19 Nov '10

Rationale: To make asdf easier to use by providing a standard interface for configuring, installing, and removing systems Details: To facilitate easier configuration, installation, and removal of asdf systems, I propose 2 new operations and 2 new functions be added. The operations would be install-op and remove-op, and the functions would be configure and config-of These new components would work as defined below: * New function: configure (defmethod configure (&rest options) ...) configure would, given a plist of options, print them to a file in the current directory (which is assumed to contain the asdf system to configure), for example `config.asdf'. If config.asdf already exists, it should be moved to config.asdf.bak before the mew config.asdf is written. * New operation: install-op install-op would, given a directory in which `configure' had been run, use the options given to configure to install the system. The options involved would be :sources and :registry . :sources is the pathname of a directory which is to be created, after which the contents of the current directory is copied into the directory named by :sources . After the files are copied, all .asd files in :sources are linked into the pathname located in :registry. It is an error to run install-op without first having called configure. * New function config-get (defmethod config-get (system config-key &optional default) ...) config-get will, given a system and config-key retrieve the value of config-key located in the config.asdf file located in the same director as the .asd file of the system (after dereferencing symlinks). This method is meant so that library/program authors can add additional configure options which the library/program can later access (Ex. a graphics library with an optional #d component could instruct users to pass `:3d t' or `:3d nil' depending on whether they want the 3d component. The user's choice would be accessed through config-get). The functions will return 2 values, the primary being the result of the lookup or default if the lookup fails, and a secondary which is T if the lookup succeeded and NIL otherwise. * New operation: remove-op remove-op would undo the operations preformed by install-op, first removing the symlinks in :registry then deleting :sources. Comments, suggestions, questions, etc. are appreciated. Krzysztof Drewniak [P.S. I will not be available between 09/22/2010 and 09/29/2010 due to a vacation resort in Mexico which has an 80% chance of no free Internet] -- X-Real-Email-With-Antispam: krzysdrewniak at gmail dot com pgp key on keyserver.ubuntu.com 94F02AE8

3 2

[asdf-devel] Directly loading "/home/etc/etc/system.asd"
by Theam Yong Chew 09 Nov '10

09 Nov '10

Hi all, Is there a way to load a system directly, leaving everything else (source-registry configuration, *central-registry*, or what not) alone? I hacked up something to do this, (asdf:load-system #P"c:/blah/blah/cl-ppcre-1.2.19/cl-ppcre.asd") so that there's no need to worry about specific versions or setting up paths to a specific library. If the library being loaded depends on other modules, then we return to searching the source-registry as if by typing this before loading our system: (asdf:load-system <lib-depended-upon>) Does asdf already support this? Will such a feature be useful? Yong.

2 1

[asdf-devel] ccl:rebuild-ccl and asdf
by R. Matthew Emerson 08 Nov '10

08 Nov '10

We've gotten a couple of bug reports from users who are having trouble doing (ccl:rebuild-ccl :clean t) when (require 'asdf) is in their init files. http://trac.clozure.com/ccl/ticket/765 http://trac.clozure.com/ccl/ticket/773 While it's undoubtedly best to start up without an init file when you're planning to rebuild the lisp (by using "ccl -n"), I'm wondering if there would be some way to avoid this particular issue involving ASDF. During the process of rebuilding ccl, we compile (but do not load) ccl:tools;asdf.lisp. Perhaps we could simply stop compiling ASDF as part of ccl:rebuild-ccl. That way, the compile-time side effects that seem to be causing the trouble would be avoided.

2 1

[asdf-devel] 2.010.1
by Faré 31 Oct '10

31 Oct '10

I renamed 2.147 into 2.010.1. We were going back and forth in version numbers (according to version-satisfies) between 2.1xx and 2.0xx, and at least that's stopped. We now have quite a few unstable 2.1xx releases in git with the "wrong" order, but presumably the few people who use asdf from git will know to upgrade from git and won't have trouble with it. Maybe I should have been more daring and have used 2.0.10.1, but that would have meant out-of-order versioning amongst stable releases, too, and that would have been more problematic. News: as compared to 2.010, 2.010.1 mostly works fine on GCL 2.7, except that GCL 2.7 has problems with condition handling (at least in the version I have from debian). [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] Clothes make the man. Naked people have little or no influence on society. — Mark Twain

1 0

[asdf-devel] New ASDF 2.010
by Faré 29 Oct '10

29 Oct '10

I just release ASDF 2.010 (same as 2.146), and I'd like all CL implementation vendors to test it and include it on their platforms so one could universally (require :asdf). As compared to 2.009: * much faster and more ported recursive groveling through directories looking for .asd's (notably fixes issues with ABCL) * portability tweaks for CCL, SCL, CLISP * eagerly error out when trying to use : in portable relative pathname designator strings * have compile-file* heed :output-file keyword * be backward-compatible with client fasl's using ASDF symbols from the obsolete ASDF-UTILITIES package that existed in ASDF2 up till 2.005 * for those who care about internals, in a comment, explanation of do-first vs in-order-to Actively supported and tested by me (on Linux amd64): ccl clisp sbcl ecl allegro abcl scl Supported but untested by me: lispworks (lacking a license), allegromodern, cmucl, any of the previous on windows Known to have many issues: gcl (still largely broken in 2.010, and still doesn't pass all tests in 2.147.) Never replied to my contact attempts: corman These implementations include asdf2 in their release and you can (require :asdf): ccl sbcl ecl abcl scl cmucl These don't yet, but probably should: allegro lispworks clisp gcl [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] Be inconsistent — but don't do it all the time. — Unknown

1 0

[asdf-devel] Feature-conditional components
by Faré 24 Oct '10

24 Oct '10

1- ASDF has no such mechanism yet. Many people have asked for it (at least nikodemus, manic12, and myself recently). 2- the feature feature was never working before I fixed and cerror'ed it, and only does conditional *dependency* on another component, that has to independently exist or conditionally exist, for which we have no non-ugly mechanism. 3- the :if-component-dep-fails mechanism may or may not be working well enough to do that right now, but man it's ugly, non-intuitive, non-documented, and gross even if it were documented. 4- the very structure of ASDF objects doesn't allow a clean way of expressing conditionally-included components, and this can't be added as an extension (at least, not unless ASDF is extended to provide a proper API, anyway). Doing anything more sophisticated than #+foo or :if-component-dep-fails is going to be "interesting" and require major work. 5- XCVB has such a mechanism, and I propose that if ASDF is to adopt non-trivial conditionals, we should import XCVB's mini-language for conditionals, if not the rest of it. 6- Hey - why not develop XCVB some more? Like, port it to your favorite platform? I'd need a bit of help to make standalone backends for Windows, including a committed hacker to port relevant parts of IOLib to Windows. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] Invent a clever saying, and your name shall live forever. — Anonymous

4 7

[asdf-devel] Error writing some output translation
by Mitch Berkson 18 Oct '10

18 Oct '10

I am encountering the error below when trying to load-system from directory d:/foo/: Error: Failed to MAKE-DIRECTORY file C:\Documents and Settings\mitch\Application Data\common-lisp\cache\lispworks-6.0.1-windows-x86\d\foo\d:\: The filename, directory name, or volume label syntax is incorrect.(123). I am able to load-system systems I have downloaded (e.g., lw-add-ons, cl-pdf) which are in a different directory (d:/lisp/). I have fooled around with a .conf file in the d:/foo/.config/common-lisp/asdf-output-translations.conf.d/ directory but that hasn't helped. I'd appreciate any suggestions. Mitch Berkson

2 4

[asdf-devel] find-system
by Lou Vanek 18 Oct '10

18 Oct '10

I am totally confused with the asdf:find-system method. The symbol is exported but it isn't bound to a function, at least on my system. About 10 different packages call asdf:find-system but it cannot be called because it isn't bound. Asdf:find-system IS declared and defined and exported in asdf.lisp version 2.009, but when a package tries to call it I get an error. Yet I can see it's defined in asdf.lisp as: (defmethod find-system (name &optional (error-p t)) (find-system (coerce-name name) error-p)) (defmethod find-system ((name string) &optional (error-p t)) (catch 'find-system ... CL-USER> (apropos "find-system") ASDF:FIND-SYSTEM ASDF::FIND-SYSTEM-FALLBACK, Def: FUNCTION ; No value CL-USER> (asdf::find-system :usocket) ; Evaluation aborted on #<CCL::UNDEFINED-FUNCTION-CALL #x18634CFE>. CL-USER> asdf::*asdf-version* "2.009" CL-USER> (lisp-implementation-type) "Clozure Common Lisp" CL-USER> (lisp-implementation-version) "Version 1.6-dev-r14367M-trunk (LinuxX8632)" What am I missing? Thanks, Lou Vanek

3 8

[asdf-devel] How to have fasls go to a fasl directory under the source directory?
by Mitch Berkson 17 Oct '10

17 Oct '10

Is there a general way to have the fasls for files in a source directory default to being written to a fasl subdirectory of the source directory? So if I have the source in d:/lisp/lw-add-ons, I'd like the fasls to be written to d:/lisp/lw-add-ons. I have done this by adding a d:/lisp/.config/common-lisp/asdf-output-translations.conf.d/lw-add-ons-asdf.conf file containing:("d:/lisp/lw-add-ons/" "d:/lisp/lw-add-ons/fasl/") But is there a way to make this the default behavior so that d:/lisp/other-system/ will have their fasls written to d:/lisp/other-system/fasl? Alternatively, could they be written to a central fasl directory but in a sub-directory named for the package? So d:/lisp/fasl/other-system/? This is on Win XP. Thanks for any help. Mitch Berkson

1 1

[asdf-devel] What do I win?
by Nikodemus Siivola 15 Oct '10

15 Oct '10

Congratulations, you're the first ever user of FEATURE dependencies! Please contact the asdf-devel mailing-list. [Condition of type SIMPLE-ERROR] Cheers, -- Nikodemus

5 12