[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Guile + SWIG
From: |
Paul Emsley |
Subject: |
Re: Guile + SWIG |
Date: |
Thu, 08 Apr 2021 08:13:34 +0100 |
User-agent: |
Evolution 3.34.1-2 |
Hi Matt and Linas,
Thanks for your feedback.
I was indeed confused by thinking the warning was a terminal error, because I
also neglected to add a newline
after using display (so it seemed to me as if it was hanging (after years of
using python, I expected the
prompt to be shown on pressing "Return", so when it didn't, I thought that
there was something wrong)).
So guile 2.2 is in the bag - and I'm pleased about that. Soon I will be looking
at converting guile-gtk scripts to
guile-gnome.
Regards,
Paul.
On Mon, 2021-04-05 at 16:51 -0700, Matt Wette wrote:
> When scripts with load-extension are compiled, the compiler does not know
> what symbols are defined in the extension so you may get warnings. Are you
> getting an error or just the warning?
>
> On 4/5/21 10:44 AM, Paul Emsley wrote:
> > Hi Linas,
> >
> > Thanks for your reply.
> >
> > Let me clarify if I can. I have many C++ functions that I would like to use
> > from scheme. I am using
> > SWIG to generate wrappers for these functions, and these get added
> > at the "root module"
> >
> > http://www.swig.org/Doc3.0/Guile.html
> >
> > The default linkage is the simplest; nothing special is done. In this
> > case
> > the function SWIG_init() is exported. Simple linkage can be used in
> > several ways:
> >
> > Embedded Guile, no modules. You want to embed a Guile interpreter into
> > your
> > program; all bindings made by SWIG shall show up in the root module.
> > Then call
> > SWIG_init() in the inner_main() function
> >
> > So that is what I was trying to do. My understanding was that
> > enhanced-ligand-coot-p
> > and its friends would be added so that their usage would be no different to
> > list, string?,
> > map and any number of basic scheme functions. However something is amiss.
> >
> > If I try this:
> > ./coot
> > Enter `,help' for help.
> > scheme@(guile-user)> (load "test-embedding.scm")
> > (55555555555555 #f)
> >
> > Now, if I install that in $prefix/share/guile/site
> > If I try this:
> > ./coot
> > Enter `,help' for help.
> > scheme@(guile-user)> (use-modules (test-embedding))
> > While compiling expression:
> > In procedure module-lookup: Unbound variable: enhanced-ligand-coot-p
> > scheme@(guile-user)>
> >
> > Is that because the "compiler" doesn't know where enhanced-ligand-coot-p
> > is? Maybe the compiler
> > is not my application, so that's why it's not there?
> >
> > If that is the case, then can I not compile my .scm files? How to do that?
> > Or can I somehow say
> > "don't worry that it's not there now, when the application runs, it will be
> > there"?
> >
> > Thanks,
> >
> > Paul.
> >
> > p.s. It's quite likely that Coot was used in the analysis of spike proteins
> > from coronavirus
> > that lead to 2P and hexapro stabilization mutants, and they have been taken
> > up by the pharmaceutical
> > companies and embedded into the mRNA or more conventional preps. So you can
> > have some guile-1.8.8-based
> > technology squirted into your arm next week.
> >
> >
> > On Mon, 2021-04-05 at 10:27 -0500, Linas Vepstas wrote:
> > > Hi Paul,
> > >
> > > Conventionally, a guile module would look something like
> > >
> > >
> > > (define-module (test-embedding))
> > > (use-modules (srfi srfi-1)) ; .. etc
> > >
> > > (define (enhanced-ligand-coot-p) #t) ; etc.
> > >
> > > (export enhanced-ligand-coot-p)
> > >
> > >
> > > In your example, you never explained how "enhanced-ligand-coot-p"
> > > magically showed up in your environment.
> > >
> > > When a module is defined, it starts with a new environment. The module
> > > cannot access stuff you've defined previously; you would have to import
> > > it somehow.
> > >
> > > --linas
> > >
> > >
> > > On Mon, 2021-04-05 at 11:24 +0100, Paul Emsley wrote:
> > > > Hi,
> > > >
> > > > I am trying to bring my swigged Gtk2 + guile 1.8.8 + guile-gtk
> > > > application up to date and I'd like some help please.
> > > >
> > > > My SWIG usage wraps many function, including
> > > > enhanced_coot_ligand_p(). I am confused between the difference
> > > > between
> > > > functions that are available from C/C++ and those from a scheme
> > > > script:
> > > >
> > > > if my inner_main looks like this:
> > > >
> > > > void inner_main(void *closure, int argc, char **argv) {
> > > >
> > > > SWIG_init();
> > > > std::string handler_string = "(lambda (key . args) ";
> > > > handler_string += "(display (list \"Error in proc:\" key \" args:
> > > > \" args)) (newline))";
> > > > SCM handler = scm_c_eval_string(handler_string.c_str());
> > > > std::string thunk_str = "(lambda() (display (list
> > > > 444444444444444444 (enhanced-ligand-coot-p))) (newline))\n";
> > > > SCM thunk = scm_c_eval_string(thunk_str.c_str());
> > > > scm_catch(SCM_BOOL_T, thunk, handler);
> > > > gtk_main();
> > > >
> > > > }
> > > >
> > > > then I get on the terminal what I expected:
> > > >
> > > > (444444444444444444 #f)
> > > >
> > > > (i.e. enhanced-ligand-coot-p is evaluated)
> > > >
> > > >
> > > > if my inner_main looks like this:
> > > > void inner_main(void *closure, int argc, char **argv) {
> > > >
> > > > SWIG_init();
> > > > std::string handler_string = "(lambda (key . args) ";
> > > > handler_string += "(display (list \"Error in proc:\" key \" args:
> > > > \" args)) (newline))";
> > > > SCM handler = scm_c_eval_string(handler_string.c_str());
> > > > std::string thunk_str = "(use-modules (test-
> > > > embedding))\n"; //////////// different ///////////////
> > > > SCM thunk = scm_c_eval_string(thunk_str.c_str());
> > > > scm_catch(SCM_BOOL_T, thunk, handler);
> > > > gtk_main();
> > > >
> > > > }
> > > >
> > > > then I get:
> > > >
> > > > ;; compiling /home/paule/autobuild/Linux-pen-pre-release-
> > > > gtk3/share/guile/site/test-embedding.scm
> > > > ;;; test-embedding.scm:21:30: warning: possibly unbound variable
> > > > `enhanced-ligand-coot-p'
> > > > ;;; compiled /home/paule/.cache/guile/ccache/2.2-LE-8-
> > > > 3.A/home/paule/autobuild/Linux-pen-pre-release-
> > > > gtk3/share/guile/site/test-embedding.scm.go
> > > > ------------------- test embedding! -----------------------
> > > > Backtrace:
> > > > 18 (apply-smob/1 #<catch-closure 7f98d1436b20>)
> > > > In ice-9/boot-9.scm:
> > > > 2312:4 17 (save-module-excursion _)
> > > > In ice-9/eval-string.scm:
> > > > 38:6 16 (read-and-eval #<input: string 7f98d08acbd0> #:lang _)
> > > > In ice-9/eval.scm:
> > > > 721:20 15 (primitive-eval _)
> > > > In ice-9/psyntax.scm:
> > > > 1262:36 14 (expand-top-sequence _ _ _ #f _ _ _)
> > > > 1209:24 13 (parse _ (("placeholder" placeholder)) ((top) #(# # …))
> > > > …)
> > > > 285:10 12 (parse _ (("placeholder" placeholder)) (()) _ c&e (eval)
> > > > …)
> > > > In ice-9/boot-9.scm:
> > > > 3377:20 11 (process-use-modules _)
> > > > 222:17 10 (map1 (((test-embedding))))
> > > > 3378:31 9 (_ ((test-embedding)))
> > > > 2800:17 8 (resolve-interface (test-embedding) #:select _ #:hide _
> > > > …)
> > > > In ice-9/threads.scm:
> > > > 390:8 7 (_ _)
> > > > In ice-9/boot-9.scm:
> > > > 2726:13 6 (_)
> > > > In ice-9/threads.scm:
> > > > 390:8 5 (_ _)
> > > > In ice-9/boot-9.scm:
> > > > 2994:20 4 (_)
> > > > 2312:4 3 (save-module-excursion _)
> > > > 3014:26 2 (_)
> > > > In unknown file:
> > > > 1 (primitive-load-path "test-embedding" #<procedure
> > > > 7f98d…>)
> > > > In test-embedding.scm:
> > > > 19:0 0 (_)
> > > >
> > > > test-embedding.scm:4:0: In procedure module-lookup: Unbound variable:
> > > > enhanced-ligand-coot-p
> > > >
> > > >
> > > >
> > > > test-embedding.scm looks like this and is installed in
> > > > $prfx/share/guile/site
> > > >
> > > > (display "------------- test embedding! ------------\n")
> > > > (display (list 55555555555555 (enhanced-ligand-coot-p)))
> > > > (newline)
> > > >
> > > >
> > > > So, inner_main() knows that enhanced-ligand-coot-p is available, but
> > > > the scheme script does not. I don't know how to
> > > > resolve this.
> > > >
> > > > How do I get a list of the available functions?
> > > >
> > > > Thanks in advance for your help.
> > > >
> > > > Paul.
> > > >
> > > >
> > > >
>
>
Re: Guile + SWIG, Linas Vepstas, 2021/04/07