Preliminary analysis of existing ASDF practices
in [Lisp]
|
Prev: Is LISP divine?
Next: keywords & macros
From: Mario S. Mommer on 20 Apr 2010 15:13 Rainer Joswig <joswig(a)lisp.de> writes: > The dependency description of mk:defsystem is based on the structure > of the modules and systems. To answer your question, it lacks these > features. > > The Symbolics defsystem dependencies are based on operations. To > define the operation dependencies the Symbolics defsystem provides a > short form (:serial, :parallel, ...) and a long form, where these > dependencies between actions can be written in detail. > > ... > (:in-order-to (:load :compile) (:load defs)) > (:in-order-to :compile (:load macros)) > ... > Amazing. Why did that not live on? Is the mechanism incompatible with current CL? Are there some documents about it, and maybe even some source??
From: Joshua Taylor on 20 Apr 2010 15:39 On 2010.04.20 3:13 PM, Mario S. Mommer wrote: > > Rainer Joswig<joswig(a)lisp.de> writes: >> The dependency description of mk:defsystem is based on the structure >> of the modules and systems. To answer your question, it lacks these >> features. >> >> The Symbolics defsystem dependencies are based on operations. To >> define the operation dependencies the Symbolics defsystem provides a >> short form (:serial, :parallel, ...) and a long form, where these >> dependencies between actions can be written in detail. >> >> ... >> (:in-order-to (:load :compile) (:load defs)) >> (:in-order-to :compile (:load macros)) >> ... >> > > Amazing. Why did that not live on? Is the mechanism incompatible with > current CL? > > Are there some documents about it, and maybe even some source?? LispWorks' still has a similar DEFSYSTEM. E.g., from http://www.lispworks.com/documentation/lw445/LWRM/html/lwref-273.htm there's an example: (defsystem clos-sys (:default-pathname "/usr/users/clc/defsys/" :default-type :lsp-file :package defsystem) :members (("defsys-macros" :type :system :root-module nil) "class" "time-methods" ("scm-pathname" :source-only t) "execute-plan" "file-types" "make-system" "conv-defsys") :rules ((:in-order-to :compile ("class" "time-methods") (:caused-by (:compile "defsys-macros")) (:requires (:load "defsys-macros"))) (:in-order-to :compile ("time-methods" "execute-plan") (:requires (:load "class"))))) //JT
From: Rainer Joswig on 20 Apr 2010 17:20 In article <87aasy7x9f.fsf(a)padme.localdomain>, m_mommer(a)yahoo.com (Mario S. Mommer) wrote: > Rainer Joswig <joswig(a)lisp.de> writes: > > The dependency description of mk:defsystem is based on the structure > > of the modules and systems. To answer your question, it lacks these > > features. > > > > The Symbolics defsystem dependencies are based on operations. To > > define the operation dependencies the Symbolics defsystem provides a > > short form (:serial, :parallel, ...) and a long form, where these > > dependencies between actions can be written in detail. > > > > ... > > (:in-order-to (:load :compile) (:load defs)) > > (:in-order-to :compile (:load macros)) > > ... > > > > Amazing. Why did that not live on? Is the mechanism incompatible with > current CL? The implementation is extensive and Symbolics Genera specific. > Are there some documents about it, and maybe even some source?? It is documented in the 'Program Development Utilities' manual. A lower-quality PDF version of the manual can be found at bitsavers.org . The source is Symbolics specific and comes with Genera. -- http://lispm.dyndns.org/
From: Rainer Joswig on 20 Apr 2010 17:29 In article <hqkvt5$urr$1(a)news.eternal-september.org>, Joshua Taylor <tayloj(a)cs.rpi.edu> wrote: > On 2010.04.20 3:13 PM, Mario S. Mommer wrote: > > > > Rainer Joswig<joswig(a)lisp.de> writes: > >> The dependency description of mk:defsystem is based on the structure > >> of the modules and systems. To answer your question, it lacks these > >> features. > >> > >> The Symbolics defsystem dependencies are based on operations. To > >> define the operation dependencies the Symbolics defsystem provides a > >> short form (:serial, :parallel, ...) and a long form, where these > >> dependencies between actions can be written in detail. > >> > >> ... > >> (:in-order-to (:load :compile) (:load defs)) > >> (:in-order-to :compile (:load macros)) > >> ... > >> > > > > Amazing. Why did that not live on? Is the mechanism incompatible with > > current CL? > > > > Are there some documents about it, and maybe even some source?? > > LispWorks' still has a similar DEFSYSTEM. E.g., from More similar to the Symbolics DEFSYSTEM is the Franz Allegro CL DEFSYSTEM: http://www.franz.com/support/documentation/current/doc/defsystem.htm Though I'd guess that the Allegro CL defsystem is a bit simpler. > > http://www.lispworks.com/documentation/lw445/LWRM/html/lwref-273.htm > > there's an example: > > (defsystem clos-sys > (:default-pathname "/usr/users/clc/defsys/" > :default-type :lsp-file > :package defsystem) > :members > (("defsys-macros" :type :system :root-module nil) > "class" > "time-methods" > ("scm-pathname" :source-only t) > "execute-plan" > "file-types" > "make-system" > "conv-defsys") > :rules > ((:in-order-to :compile ("class" "time-methods") > (:caused-by (:compile "defsys-macros")) > (:requires > (:load "defsys-macros"))) > (:in-order-to :compile > ("time-methods" "execute-plan") > (:requires (:load "class"))))) > > //JT A syntactic difference is that the rules in LispWorks have their own main clause under :RULES. -- http://lispm.dyndns.org/
From: o.jasper on 21 Apr 2010 14:01
I didn't read other people's posts nearly well enough before replying.. Sorry for that. Most of my opinion still stands, except i noticed an issue with how i suggested to have backwards-compatibility; just starting asdf when there is no way around it isn't enough.. one has to stop it, and convince asdf it has the dependencies that were obtained from outside it. |