8:35 p.m.
The Internals manual says:
,-------------------
| The C code is actually written in a dialect of C called "Clean C",
| meaning that it can be compiled, warning-free, with either a C or C++
| compiler.
`-------------------
This doesn't seem to be true for lwlib at least (not that I care about
lwlib specifically -- I just heppened to have used my standard
configure command line). Has anyone actually compiled XEmacs with a
C++ compiler recently and has a recipe that I could borrow?
Thanks
Marcus
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:42 p.m.
>>>> Marcus <marcus.harnisch(a)xemacs.org> writes:
Has anyone actually compiled XEmacs with a C++ compiler recently and
has a recipe that I could borrow?
The forth configuration used by the buildbot is a g++
configuration. We have x86 and osx builds using that config.
- http://www.lidell.nu:8010/builders/imac%20osx%20cfg3
- http://www.lidell.nu:8010/builders/x86%20gentoo%20cfg3
Whether that is a good configuration to use is another deal. What
problems do you see when using a C++ compiler? I hope using C++ with
more or less any configuration would work fine!?
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
7:10 p.m.
Hi Mats
Sorry for the late response, I had to sort some SMTP issues first.
Mats Lidell <matsl(a)xemacs.org> writes:
Whether that is a good configuration to use is another deal. What
problems do you see when using a C++ compiler? I hope using C++ with
more or less any configuration would work fine!?
Thanks. One difference was `--with-xemacs-compiler=g++', whereas I
used the naive CC=g++. But there seem to be some remaining issues
related to lwlib and Xaw.
,--------------------
| lwlib-Xaw.c: In function 'void xaw_scrollbar_scroll(Widget, XtPointer,
XtPointer)':
| lwlib-Xaw.c:616:35: error: cast from 'XtPointer {aka void*}' to 'int'
loses precision [-fpermissive]
| event_data.slider_value = (int) call_data;
| ^
| lwlib-Xaw.c:619:13: error: cast from 'XtPointer {aka void*}' to 'int'
loses precision [-fpermissive]
| if ((int) call_data > 0)
| ^
| make[1]: *** [lwlib-Xaw.o] Error 1
`--------------------
Below is the configuration that I am using.
,--------------------
| ./configure --with-xemacs-compiler=g++ --with-sound=alsa
| --enable-menubars=lucid --enable-scrollbars=athena
| --enable-dialogs=athena --enable-widgets=athena --with-athena=3d
| --without-gpm --without-ldap --without-postgresql --with-mule
| --without-xim --without-canna --with-xft=emacs,tabs,gauges,menubars
| --with-zlib --without-bignum --with-kkcc --with-newgc
| --without-error-checking --with-modules --with-database=berkdb
| --with-ncurses
`--------------------
For my own experiments I don't really care about any toolkit, so I might just
copy one of the working configs.
Cheers
Marcus
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
3:25 a.m.
On Mon, Jan 5, 2015 at 12:10 PM, Marcus <marcus.harnisch(a)xemacs.org> wrote:
,--------------------
| lwlib-Xaw.c: In function 'void xaw_scrollbar_scroll(Widget, XtPointer,
XtPointer)':
| lwlib-Xaw.c:616:35: error: cast from 'XtPointer {aka void*}' to 'int'
loses precision [-fpermissive]
| event_data.slider_value = (int) call_data;
| ^
| lwlib-Xaw.c:619:13: error: cast from 'XtPointer {aka void*}' to 'int'
loses precision [-fpermissive]
| if ((int) call_data > 0)
| ^
| make[1]: *** [lwlib-Xaw.o] Error 1
`--------------------
Does the compile succeed if you cast to long first, then to int? That is
Line 616:
event_data.slider_value = (int) (long) call_data;
Line 619:
if ((int) (long) call_data > 0)
It would be better to cast to intptr_t, instead of long, but then we'd
have to know what an intptr_t is.
--
Jerry James
http://www.jamezone.org/
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4:19 p.m.
The following message is a courtesy copy of an article
that has been posted to gmane.emacs.xemacs.beta as well.
Jerry James <james(a)xemacs.org> writes:
Does the compile succeed if you cast to long first, then to int?
Indeed, it does. The secrets of "why" are part of the C++ standard, I
reckon. I am not even getting a warning despite the fact that casting
long to int is lossy.
Many thanks
Marcus
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
5:18 p.m.
Marcus writes:
Jerry James <james(a)xemacs.org> writes:
> Does the compile succeed if you cast to long first, then to int?
Indeed, it does. The secrets of "why" are part of the C++ standard, I
reckon. I am not even getting a warning despite the fact that casting
long to int is lossy.
The casts among ints are allowed, otherwise you'd need a new syntax
for what is a very common idiom (especially in C, which does a fair
amount of implicit coercion of function arguments to wider or narrower
types). It makes sense to warn about pointers vs integer types,
though, because the implicit assumption that pointers are the same
width as some integer type is much more likely to be violated (as I
just found out...), and less likely to be recognized since something
other than width (signedness, for example) is very likely the
conscious reason for the cast.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
5:31 a.m.
Marcus writes:
,--------------------
| lwlib-Xaw.c: In function 'void xaw_scrollbar_scroll(Widget, XtPointer,
XtPointer)':
| lwlib-Xaw.c:616:35: error: cast from 'XtPointer {aka void*}' to 'int'
loses precision [-fpermissive]
What architecture is your host? In most architectures, pointers are
the same size as long ints which are the same size as ints (32 bits or
64 bits). That's why others aren't seeing this (and one reason why
the X11 devils chose XtPointer to be the "AnyType" type).
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
3:44 p.m.
Stephen
What architecture is your host?
Linux x86_64. Not exactly uncommon I think.
In most architectures, [...] long ints [...] are the same size as
ints (32 bits or 64 bits).
I beg your pardon, but Linux has always used LP64 (ie. sizeof(void *)
!= sizeof(int)), and according to Wikipedia "most" relevant unixoid
systems have:
"Most Unix and Unix-like systems, e.g. Solaris, Linux, BSD, and OS X;
z/OS"
(http://en.wikipedia.org/wiki/64-bit_computing#64-bit_data_models)
--
Marcus
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4:22 p.m.
The following message is a courtesy copy of an article
that has been posted to gmane.emacs.xemacs.beta as well.
Marcus <marcus.harnisch(a)xemacs.org> writes:
sizeof(void *) != sizeof(int)
Sorry, that was meant to read
sizeof(long) != sizeof(int)
--
Marcus
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
5:07 p.m.
Marcus writes:
Stephen
> What architecture is your host?
Linux x86_64. Not exactly uncommon I think.
I wonder why this hasn't been reported before, then. I haven't done
any builds with C++ for several years, but I did at one time. I'm
pretty sure I was already using x86_64 then.... Maybe gcc (?) was
conforming to an older, more permissive standard then.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
3608
days inactive
3623
days old
9 comments
4 participants
participants (4)
-
Jerry James -
Marcus Harnisch -
Stephen J. Turnbull -
The XEmacs Package Smoketest