Cannot compile XEmacs 21.5 (beta28) from latest changes (fwd)

Aidan Kehoe kehoea at parhasard.net
Tue Aug 26 18:44:35 EDT 2008


 Ar an séú lá is fiche de mí Lúnasa, scríobh robert delius royar: 

 > [I noticed that this mail sent earlier appeared to have been redirected 
 > to a different MX than the mail.xemacs.org greylister and never 
 > retried.]
 > 
 > If Lisp_Object is a union (not an int), then constructs such as
 >   assert (make_int (1) == mapped); /* line 2826 symbols.c */
 > need to be
 >   assert (EQ(make_int (1), mapped));
 > and
 >   Qzero != res  /* line 3250 symbols.c */
 > need to be
 >    !EQ(Qzero, res)
 > 
 > The changes suggested lead to a working XEmacs binary when unions are
 > used for Lisp_Objects.  I have not tested with ints used instead.

I’ve made your changes and tested with integers used for Lisp_Objects; it
works, thanks for the report, and I’ve committed them. How exactly do you
manage to regularly build using the union type without problems? For
example, tonight, with the union build, I had to compile data.c without
optimisation, otherwise the dump errored out in font.el, with
#'put-display-table confused about whether 247 is an integer that can be
converted to a valid character.

-- 
¿Dónde estará ahora mi sobrino Yoghurtu Nghé, que tuvo que huir
precipitadamente de la aldea por culpa de la escasez de rinocerontes?




More information about the XEmacs-Beta mailing list