January 2010 - asdf-devel - mailman3.common-lisp.net

[asdf-devel] dependency bug?
by Robert Brown 26 Jan '10

26 Jan '10

Thanks very much for the suggestions. My example code was a simplification of a more complicated system definition, part of an project where I'm implementing Google's protocol buffers in Lisp. http://github.com/brown/protobuf/blob/master/protobuf.asd Protocol description files (suffix .proto) need to be converted into Lisp source, which is then compiled and loaded. I'm trying to get the dependencies right in one of my COMPONENT-DEPENDS-ON methods. Anyway, back to my simple example. When I use (:cl-source-file "bug2" :depends-on ("bug1")) the ASDF component for bug2 ends up with the following IN-ORDER-TO slot: ((ASDF:LOAD-OP (ASDF:LOAD-OP "bug1")) (ASDF:COMPILE-OP (ASDF:COMPILE-OP "bug1"))) I don't understand this. I believe it should be: ((ASDF:LOAD-OP (ASDF:LOAD-OP "bug1")) (ASDF:COMPILE-OP (ASDF:LOAD-OP "bug1"))) and in fact when do # touch bug1.lisp (asdf:operate 'asdf:compile-op 'bug) component bug1 is loaded (the source is compiled and then the fasl is loaded), not just compiled, before bug2 is compiled. My goal is to understand what the ASDF operations mean. LOAD-OP means "compile and load my code". However, COMPILE-OP seems to have two different meanings. COMPILE-OP means "compile" when it's a goal to perform, but it means "compile and load" when it's a dependency of another goal. Does that sound right? It's confusing me because we already have an operator that functions as "compile and load", namely LOAD-OP. bob ===================== On Fri, Jan 22, 2010 at 9:05 PM, Robert Goldman <rpgoldman(a)sift.info> wrote: > On 1/22/10 Jan 22 -6:20 PM, Robert Brown wrote: >> ASDF is behaving in a way I find mysterious. Maybe I've found a >> bug, but more likely I just don't understand ASDF dependencies. >> >> To demonstrate the problem, create the three files below. >> >> ========== >> bug.asd >> ========== >> >> (cl:in-package #:common-lisp-user) >> >> (defpackage #:bug-system >> (:use #:common-lisp #:asdf)) >> >> (in-package #:bug-system) >> >> (defsystem bug >> :components >> ((:cl-source-file "bug1") >> (:cl-source-file "bug2" >> :in-order-to ((load-op (load-op "bug1")) >> (compile-op (load-op "bug1")))))) >> >> ========== >> bug1.lisp >> ========== >> >> (defun bug1 ()) >> >> ========== >> bug2.lisp >> ========== >> >> (defun bug2()) >> >> ========== >> >> Execute the following: >> >> (load "bug.asd") >> (let ((*load-verbose* t)) (asdf:operate 'asdf:load-op 'bug)) >> >> Next, exit your Lisp session, start up your Lisp environment >> again and execute the two lines above for a second time. ASDF >> recompiles bug2.lisp. I expect it to instead just load >> bug2.fasl. >> >> Can someone explain what's going on? >> >> My example is intended to mirror the situation where you need to >> have some in-line functions or macros loaded into the Lisp system >> before another file can be compiled or loaded. Maybe I should be >> expressing this dependency differently to ASDF. > > Try this with no :in-order-to, just > > (:file "bug2" :depends-on ("bug1")) > > and see if that does what you expect.... If not, get back to us, ok? > > cheers, > r

4 8

[asdf-devel] [patch] markdown documentation tweaks
by Daniel Herring 26 Jan '10

26 Jan '10

The attached patch fixes the markdown files in asdf. Git command to apply it: # git am 0001-markdown-documentation-tweaks.patch Later, Daniel

3 5

[asdf-devel] possible clisp/win32 incompatibility
by Daniel Herring 25 Jan '10

25 Jan '10

Hi all, I'm trying to get ABLE running on mswin and hit multiple issues at once... For one, it appears the a recent version of ASDF broke compatibility with clisp on windows. If *enable-asdf-binary-locations* is t, I get "invalid :directory (clisp-2.48...)"; if it is nil, ASDF is trying to load sources from the wrong location. This could be a configuration issue on my side, but it affects both clisp 2.48 and clisp 2.45, and this configuration works fine with ccl 1.3 and 1.4. Both clisp versions work fine with the same configuration using a ~5 month old version of ASDF... I'm focusing on other bugs now. Later, Daniel

5 9

[asdf-devel] loading platform-dependent code
by Tamas Papp 24 Jan '10

24 Jan '10

Hi, I am the author of the cl-cairo2 library. Cairo itself is a graphics library, and my library provides a Lisp interface. Some of the features (eg drawing in X11, Windows or Quartz) are platform dependent. Currently, the approach of cl-cairo2 is to provide extra system definitions for these, eg an X11 user would load the systems cl-cairo2 and cl-cairo2-x11 manually. I am wondering if there is a better way to do platform-dependent loading. The solution I was thinking of was to try loading the platform dependent shared C libraries using CFFI, and set some flag (or push something to *features*) if they succeed, but if they fail, capture the condition with handler-case and just proceed with a warning. Then I could load platform-dependent code conditionally on this --- but I don't know how to do that in ASDF. Please tell me what is the "right way" to do these things (ie stuff that I put in features). Examples of libraries that load platform-dependent code conditionally with ASDF would also help. Thanks, Tamas

3 2

[asdf-devel] dependency bug?
by Robert Brown 23 Jan '10

23 Jan '10

ASDF is behaving in a way I find mysterious. Maybe I've found a bug, but more likely I just don't understand ASDF dependencies. To demonstrate the problem, create the three files below. ========== bug.asd ========== (cl:in-package #:common-lisp-user) (defpackage #:bug-system (:use #:common-lisp #:asdf)) (in-package #:bug-system) (defsystem bug :components ((:cl-source-file "bug1") (:cl-source-file "bug2" :in-order-to ((load-op (load-op "bug1")) (compile-op (load-op "bug1")))))) ========== bug1.lisp ========== (defun bug1 ()) ========== bug2.lisp ========== (defun bug2()) ========== Execute the following: (load "bug.asd") (let ((*load-verbose* t)) (asdf:operate 'asdf:load-op 'bug)) Next, exit your Lisp session, start up your Lisp environment again and execute the two lines above for a second time. ASDF recompiles bug2.lisp. I expect it to instead just load bug2.fasl. Can someone explain what's going on? My example is intended to mirror the situation where you need to have some in-line functions or macros loaded into the Lisp system before another file can be compiled or loaded. Maybe I should be expressing this dependency differently to ASDF. bob

3 2

[asdf-devel] ASDF configuration
by Faré 19 Jan '10

19 Jan '10

Dear ASDF-interested hackers, I've cleaned up my code for ASDF configuration and committed it to my repo. http://common-lisp.net/gitweb?p=projects/xcvb/asdf.git Documentation is currently in README.source-registry there. While discussing with Peter Van Eynde and James Knight regarding debian common-lisp-controller bug 560781 I came to the conclusion that we probably want something similar for ASDF-BINARY-LOCATIONS. If someone who has hacked ABL more than I understands the issues, I'd be grateful to discuss with him and flesh out an implementation of such a configuration (maybe refactoring the code to share stuff with the current parser). Are there any other parts of ASDF that require similar configuration? In any case, in my repository, I bumped the version to 1.500 to signal the significance of recent changes. I will merge and release next week if no one sees any objection. When all the configuration we need is ready and debugged, I'd like to release a version 2.0 and start peddling ASDF upgrades to all Lisp implementation vendors. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The probability of the people in power being individuals who would dislike the possession and exercise of power is on a level with the probability that an extremely tender-hearted person would get the job of whipping-master on a slave plantation. — Frank H. Knight

1 0

[asdf-devel] asdf.texinfo version 1.375 has errors
by Peter Van Eynde 19 Jan '10

19 Jan '10

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Hi all, While trying to build a new Debian package I get the following errors: > $ makeinfo --verbose asdf.texinfo > makeinfo (GNU texinfo) 4.13 > Making info file `asdf.info' from `asdf.texinfo'. > Formatting node Top... > Formatting node Introduction... > Formatting node Using asdf to load systems... > Formatting node Defining systems with defsystem... > Formatting node The defsystem form... > Formatting node A more involved example... > Formatting node The defsystem grammar... > Formatting node Other code in .asd files... > Formatting node The object model of asdf... > Formatting node Operations... > Formatting node Predefined operations of asdf... > Formatting node Creating new operations... > Formatting node Components... > Formatting node Common attributes of components... > Formatting node Pre-defined subclasses of component... > Formatting node Creating new component types... > Formatting node Controlling where ASDF saves compiled files... > Formatting node Error handling... > Formatting node Compilation error and warning handling... > Formatting node Miscellaneous additional functionality... > Formatting node Getting the latest version... > Formatting node FAQ... > Formatting node TODO list... > Formatting node missing bits in implementation... > Formatting node Inspiration... > Formatting node Concept Index... > Formatting node Function and Class Index... > Formatting node Variable Index... > /home/pvaneynd/fakeroot/clean/git/cl-asdf//asdf.texinfo:1301: Prev field of node `FAQ' not pointed to. > /home/pvaneynd/fakeroot/clean/git/cl-asdf//asdf.texinfo:1451: This node (missing bits in implementation) has the bad Next. > /home/pvaneynd/fakeroot/clean/git/cl-asdf//asdf.texinfo:1281: Node `Getting the latest version' lacks menu item for `FAQ' despite being its Up target. > /home/pvaneynd/fakeroot/clean/git/cl-asdf//asdf.texinfo:1281: Next field of node `Getting the latest version' not pointed to (perhaps incorrect sectioning?). > /home/pvaneynd/fakeroot/clean/git/cl-asdf//asdf.texinfo:1301: This node (FAQ) has the bad Prev. This is fixed with the trivial patch: > diff --git a/asdf.texinfo b/asdf.texinfo > index 3d379e4..ddefe01 100644 > --- a/asdf.texinfo > +++ b/asdf.texinfo > @@ -1298,7 +1298,7 @@ You will find the above referenced tags in this repository. > Discussion of ASDF development is conducted on the mailing list > @kbd{asdf-devel@@common-lisp.net}. > > -@node FAQ, TODO list, missing bits in implementation, Getting the latest version, Top > +@node FAQ, TODO list, Getting the latest version, Top > @comment node-name, next, previous, up > @chapter FAQ See also http://git.debian.org/?p=pkg-common-lisp/cl-asdf.git;a=commit;h=ea3aa53b623… Groetjes, Peter - -- signature -at- pvaneynd.mailworks.org http://www.livejournal.com/users/pvaneynd/ "God, root, what is difference?" Pitr | "God is more forgiving." Dave Aronson| -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQIcBAEBCAAGBQJLVVNrAAoJEB4VNT76eIlawy4QAKUEa/QOLfuKZCvCthVmPj0p 6PuLzfmpGM7jJ+7fsx0WQEAHgE/9RkFL9emC8diAPRiAXCBdB1ypZ9uciniYtc1Y x7/5QL2vpJH9NYmMlK6xKE2l0doYAe2sgSSt6BzvFpvXdnv7mt1Il5CmN93RtJQI 9RUm7aei+uqN1n+tojNw37OMa0YC2x/hSdjPdqTyDqm3gDa0kUYXDfTHomlcBBRP tavFqwLvqfZ3LlbIiObJqCu6eh35gpCW6O1AmdG/dXAGfzjoKV7tlR8nvfIoCBsu P6XUi32P0K0H2fGfJ55ETjszJSwb735aWvAlaNHuZV6Z7xa5uLKJ9FTraY+4K2i6 j7swacO/NJla3m8eOVaZF+6wZEuiWZpODSH0u6eo0JSlCX6OpGO41q4uQ6iJA7Pz nj5e0kEejnamk9IfLVAFPL8kpz7HfxreEVGr83iP4/At6JKOAO3YbOfwcL0/po51 2b+eqkZNsZqmolrJK0hQqIN5A0LhuyQ4CVr/o9xulx08MZkHCnNYC7t2cLBMWBjG hYx5exF2iqlnOlF7IZ74qdjRye9ykZ8aOs24zvSX6gWwzfyniLZoEYuH8sNrEWD1 queXBogs4Wqmda50gICYw6ADmplcCpQfnI2/f+DfjH6JSnxv4dbYcCD7hqwrlkS3 KDuVd4nSxQ3BhfQ2NSi3 =0l7y -----END PGP SIGNATURE-----

1 0

[asdf-devel] ASDF configuration through files
by Faré 18 Jan '10

18 Jan '10

Dear ASDF hackers, my experimental repository at http://common-lisp.net/gitweb?p=projects/xcvb/asdf.git contains some (fugly) code to address the configurability issue: https://bugs.launchpad.net/asdf/+bug/485918 I won't release the code until I clean it up somewhat, but I'm already requesting comments. The idea is that you should configure your paths through user and system files, more or less as described in a previous mailing-list post: http://common-lisp.net/pipermail/asdf-devel/2009-December/000580.html This hooks into the asdf sysdef search mechanism in a way that takes precedence over the current *central-registry* but falls back to it when systems are not found. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The difference between a programmer and a user, is that the programmer knows there is no difference between using and programming. — Faré

2 3

[asdf-devel] git resources
by Daniel Herring 17 Jan '10

17 Jan '10

Hi all, I just found this list of resources. http://techbase.kde.org/Getting_Started/Sources/Amarok_Git_Tutorial#Recomme… It might be good to add some of them to ASDF's resource list. - Daniel

1 0

[asdf-devel] tab-blocking commit hook
by dherring＠tentpost.com 15 Jan '10

15 Jan '10

The attached script should prevent tabs in *.asd, *.lisp, or *.script files from being committed to a git repository. Existing tabs will not be affected, unless they appear in (the context of) another modification. Here's an example stopped commit. "" Error: This git repository disallows the #\Tab character in lisp files. Only spaces should be used for indenting lisp code. In emacs, 'M-x untabify' and (setq indent-tabs-mode nil) may help. First offending file and lines: test.asd + this is a tab "" To install, simply copy this script to .git/hooks/pre-commit and set the executable bit. It might also be good to put a copy somewhere permanent so others can install it to their local repositories. Later, Daniel

4 5