guile-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: gnome-guile broken with guile 1.5 (gg used guile-snarf).


From: Thien-Thi Nguyen
Subject: Re: gnome-guile broken with guile 1.5 (gg used guile-snarf).
Date: Mon, 11 Mar 2002 21:25:34 -0800

   From: Rob Browning <address@hidden>
   Date: Mon, 11 Mar 2002 11:21:47 -0600

   Hmm.  I'm not sure if it's related to your work, a problem in
   gnome-guile, or some (expected) drift in 1.5 vs 1.4, but gnome-guile
   doesn't seem to like the new guile-snarf.  Though I'm not very
   familiar with the snarfing process, it looks like the final .x include
   isn't being ignored at snarf time as it should be (since the snarf
   command is being run to actually generate that file).

libguile/guile-snarf.in is different between 1.4 and 1.6:

 ! ${CPP} -DSCM_MAGIC_SNARFER "$@" > ${temp} || exit $?
 ! < ${temp} grep "^ *SCM__I" | sed -e "s/^ *SCM__I//" -e 's/SCM__D.*$//g'
   
  -------
   
 ! ${CPP} -DSCM_MAGIC_SNARF_INITS "$@" > ${temp} || exit $?
 ! < ${temp} grep "^ *\^ *\^" | sed -e "s/^ *\^ *\^//"
  
so that's the likely cause of the problems.  the term "expected drift"
sounds really weird to me, btw -- have you been talking to marketing?

i suppose what i'll do is port 1.4 functionality into 1.6, add
"--compat=1.4" (or whatever, i'm open to suggestions) option handling,
and document these things.  :-/

that is, if possible.  diff 1.4 1.6 shows:

 < SCM_VCELL (c_name, scheme_name)
 < SCM_GLOBAL_VCELL (c_name, scheme_name)
 < SCM_VCELL_INIT (c_name, scheme_name, init_val)
 < SCM_GLOBAL_VCELL_INIT (c_name, scheme_name, init_val)
 < SCM_CONST_LONG (c_name, scheme_name, value)
 ---
 > SCM_VARIABLE (c_name, scheme_name)
 > SCM_GLOBAL_VARIABLE (c_name, scheme_name)
 > SCM_VARIABLE_INIT (c_name, scheme_name, init_val)
 > SCM_GLOBAL_VARIABLE_INIT (c_name, scheme_name, init_val)

for the most part compat mode means defining some aliases and
supporting SCM_CONST_LONG.  no big deal.

the hair arrives because usage is different also.  crux of the matter is
1.4 does not require #ifndef SCM_MAGIC_SNARFER in source (like 1.6), so
two suggested Makefile frags need to be documented.

so overall, i think this can be done.  any objections?

   and at build time we get:

     guile-gtk.c:2784: guile-gtk.x: No such file or directory

this means the call to guile-snarf failed in some way, but only the
Makefile knows for sure...  what does guile-snarf invocation look like
there?

thi



reply via email to

[Prev in Thread] Current Thread [Next in Thread]