7:19 p.m.
I can no longer build from source. . .
I took a clean clone of the repo just now (5715:68f8d295be49),
configured as usual:
./configure --with-pdump=yes --with-modules=no --with-mule=yes
--with-ncurses=yes --with-msw=yes --with-png --with-xpm --with-jpeg
Tried to build:
make
No joy:
. . .
gcc -c -I/usr/include/noX -Wall -Wno-switch -Wundef -Wsign-compare
-Wno-char-subscripts -Wpacked -Wpointer-arith -Wshadow
-Wmissing-declarations -Wmissing-prototypes -Wstrict-prototypes
-Wdeclaration-after-statement -Wunused-parameter -g -Demacs
-I. -I/usr/local/src/xemacs-21.5-scratch/src -DHAVE_CONFIG_H
-fno-caller-saves alloc.c
In file included from sysfile.h:95:0,
from alloc.c:61:
syswindows.h:446:16: error: redefinition of 'struct
tagNMDATETIMEFORMATA'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4120:18:
note: originally defined here
syswindows.h:453:3: error: conflicting types for 'NMDATETIMEFORMATA'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4126:5:
note: previous declaration of 'NMDATETIMEFORMATA' was here
syswindows.h:453:28: error: conflicting types for
'LPNMDATETIMEFORMATA'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4126:24:
note: previous declaration of 'LPNMDATETIMEFORMATA' was here
syswindows.h:455:16: error: redefinition of 'struct
tagNMDATETIMEFORMATW'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4128:18:
note: originally defined here
syswindows.h:462:3: error: conflicting types for 'NMDATETIMEFORMATW'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4134:5:
note: previous declaration of 'NMDATETIMEFORMATW' was here
syswindows.h:462:28: error: conflicting types for
'LPNMDATETIMEFORMATW'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4134:24:
note: previous declaration of 'LPNMDATETIMEFORMATW' was here
In file included from syswindows.h:634:0,
from sysfile.h:95,
from alloc.c:61:
intl-auto-encap-win32.h:63:36: error: expected declaration specifiers
or '...' before 'PCVOID'
intl-auto-encap-win32.h:2303:37: error: expected declaration
specifiers or '...' before 'PCVOID'
alloc.c: In function 'Fmake_byte_code':
alloc.c:2257:5: warning: value computed is not used
alloc.c: In function 'malloced_storage_size':
alloc.c:3841:11: warning: declaration of 'log2' shadows a global
declaration
alloc.c: In function 'tree_memory_usage_1':
alloc.c:4562:7: warning: value computed is not used
GNUmakefile:102: recipe for target `alloc.o' failed
make[1]: *** [alloc.o] Error 1
make[1]: Leaving directory `/usr/local/src/xemacs-21.5-scratch/src'
GNUmakefile:97: recipe for target `src' failed
make: *** [src] Error 2
cygcheck -s
. . .
Cygwin DLL version: 1.7.17
. . .
w32api 9999-1 OK
w32api-headers 3.0b_svn5496-1 OK
w32api-runtime 3.0b_svn5496-1 OK
A fair amount of poking around doesn't shed any light. The above
setup worked for mercurial tip on 23 August 2012. Suggestions?
ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ht(a)inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
8:04 p.m.
ht writes:
I can no longer build from source. . .
. . .
> cygcheck -s
. . .
Cygwin DLL version: 1.7.17
. . .
w32api 9999-1 OK
w32api-headers 3.0b_svn5496-1 OK
w32api-runtime 3.0b_svn5496-1 OK
I should also have said -- Windows 7 Pro 64-bit, same as last
successful build.
ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ht(a)inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:24 p.m.
>>>> Henry S Thompson <ht(a)inf.ed.ac.uk> writes:
> I can no longer build from source. . .
> [...]
I should also have said -- Windows 7 Pro 64-bit, same as last
successful build.
I've had the same problem on a 32-bit Vista machine. (Must confess I
bailed out after some initial error checking and installed a Virtual
Machine with guest OS linux for my XEmacs needs...)
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
8:16 p.m.
On Fri, Feb 1, 2013 at 12:19 PM, Henry S. Thompson <ht(a)inf.ed.ac.uk> wrote:
I can no longer build from source. . .
I don't do Windows. (My work and home machines both run Linux.) But
the cause of this seems pretty clear. In src/Changelog I see:
2002-11-17 Ben Wing <ben(a)xemacs.org>
...
* syswindows.h:
* syswindows.h (struct):
* syswindows.h (tagNMDATETIMEFORMATW):
* syswindows.h (tagNMTTDISPIFNOA):
* syswindows.h (tagNMTTDISPINFOW):
Fix Cygwin complaints now that some missing structs have been added.
So, a decade later, it looks like a couple of those structs are now
declared in <w32api/commctrl.h>. Somebody besides me will have to
figure out how it should really be fixed (maybe with an appropriate
#if W32API_INSTALLED_VER < W32API_VER(x,y) for some values of x and
y?), but this patch may let you keep moving for now:
diff -r 68f8d295be49 src/syswindows.h
--- a/src/syswindows.h Wed Jan 23 11:48:46 2013 -0700
+++ b/src/syswindows.h Fri Feb 01 13:14:35 2013 -0700
＠＠ -443,6 +443,7 ＠＠
#endif /* not NMCBEDRAGBEGIN */
+#if 0
typedef struct tagNMDATETIMEFORMATA
{
NMHDR nmhdr;
＠＠ -460,6 +461,7 ＠＠
LPCWSTR pszDisplay;
WCHAR szDisplay[64];
} NMDATETIMEFORMATW, FAR * LPNMDATETIMEFORMATW;
+#endif
#if W32API_INSTALLED_VER < W32API_VER(2,2)
Regards,
--
Jerry James
http://www.jamezone.org/
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:35 p.m.
>>>> Jerry James <james(a)xemacs.org> writes:
I don't do Windows. (My work and home machines both run Linux.)
But the cause of this seems pretty clear. In src/Changelog I see:
[...]
So, a decade later, it looks like a couple of those structs are now
declared in <w32api/commctrl.h>.
Yes, I stumbled on this problem a few months ago, and I came to this
conclusion as well.
Somebody besides me will have to figure out how it should really be
fixed [...]
I got lost when trying to do that. With really no time available, at
that time, I switched to a virtual machine and linux for getting my
hands on XEmacs. Its a little bit slow but works.
Anyway, I think I have the XEmacs workspace left on that machine so I
give this a spin next week and see what more pops up. At least I can
try the patch.
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:36 p.m.
Jerry James writes:
On Fri, Feb 1, 2013 at 12:19 PM, Henry S. Thompson
<ht(a)inf.ed.ac.uk> wrote:
> I can no longer build from source. . .
I don't do Windows. (My work and home machines both run Linux.) But
the cause of this seems pretty clear. In src/Changelog I see:
. . .
Ah, thanks, I didn't look that far back.
That patch helps, but still leaves me with
intl-auto-encap-win32.h:63:36: error: expected declaration specifiers
or '...' before 'PCVOID'
intl-auto-encap-win32.h:2303:37: error: expected declaration
specifiers or '...' before 'PCVOID'
Web searching suggests (maybe -- not sure about this one. . .) that
that should be LPCVOID -- trying that gets alloc.c to compile, and a
lot more, but finally:
event-msw.c: In function 'mswindows_wnd_proc':
event-msw.c:3742:10: error: 'WM_IME_CHAR' undeclared (first use in
this function)
event-msw.c:3742:10: note: each undeclared identifier is reported only
once for each function it appears in
event-msw.c:3744:10: error: 'WM_IME_STARTCOMPOSITION' undeclared
(first use in this function)
event-msw.c:3748:10: error: 'WM_IME_COMPOSITION' undeclared (first use
in this function)
These are defined in /usr/include/w32api/winuser.h, but must be being
ifdefed out -- it's late and I can't work through the nested ifdefs
now -- Vin, if you can see where this is going wrong, that would be
great.
ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ht(a)inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
8:48 p.m.
Hi, Henry -
On Fri, Feb 1, 2013 at 2:19 PM, Henry S. Thompson <ht(a)inf.ed.ac.uk> wrote:
I can no longer build from source. . .
I guess you're building on cygwin, right? I suspect you're trying to
build with a different version of gcc than you used previously. The
w32api files have also been upgraded fairly recently in cygwin, so
that could be a source of trouble.
What version of windows? I pretty regularly run a Cygwin/XP build,
but it has probably been a few months. I'll start up a new build
sometime this weekend.
- Vin
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
2:42 p.m.
>>>> Vin Shelton <acs(a)alumni.princeton.edu> writes:
What version of windows? I pretty regularly run a Cygwin/XP build,
but it has probably been a few months. I'll start up a new build
sometime this weekend.
I just tested on a Windows XP, SP3, 32-bit, machine and with cygwin,
cygcheck (cygwin) 1.7.15, and I get this (I have a slight feeling of a
dejavu, right!)
(I just saw configure preferred g++ for some reason but still it is an
obvious clash with system header files. So removing these structs is
likely the right thing and then we'll see what happens ;-))
----------------------------------------------------------------------
make[1]: Entering directory `/cygdrive/w/src/xemacs/work/windows-a/src'
g++ -c -Wall -Wno-switch -Wundef -Wsign-compare -Wno-char-subscripts -Wpacked
-Wpointer-arith -Weffc++ -g -fno-strict-aliasing -Demacs -I.
-I/cygdrive/w/src/xemacs/work/windows-a/src -DHAVE_CONFIG_H -fno-caller-saves alloc.c
In file included from sysfile.h:95:0,
from alloc.c:61:
syswindows.h:446:16: error: redefinition of 'struct tagNMDATETIMEFORMATA'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4120:39: error:
previous definition of 'struct tagNMDATETIMEFORMATA'
syswindows.h:453:20: error: invalid type in declaration before ',' token
syswindows.h:453:20: error: conflicting declaration 'typedef int
NMDATETIMEFORMATA'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4126:5: error:
'NMDATETIMEFORMATA' has a previous declaration as 'typedef struct
tagNMDATETIMEFORMATA NMDATETIMEFORMATA'
syswindows.h:453:28: error: conflicting declaration 'typedef int*
LPNMDATETIMEFORMATA'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4126:24: error:
'LPNMDATETIMEFORMATA' has a previous declaration as 'typedef struct
tagNMDATETIMEFORMATA* LPNMDATETIMEFORMATA'
syswindows.h:455:16: error: redefinition of 'struct tagNMDATETIMEFORMATW'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4128:39: error:
previous definition of 'struct tagNMDATETIMEFORMATW'
syswindows.h:462:20: error: invalid type in declaration before ',' token
syswindows.h:462:20: error: conflicting declaration 'typedef int
NMDATETIMEFORMATW'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4134:5: error:
'NMDATETIMEFORMATW' has a previous declaration as 'typedef struct
tagNMDATETIMEFORMATW NMDATETIMEFORMATW'
syswindows.h:462:28: error: conflicting declaration 'typedef int*
LPNMDATETIMEFORMATW'
/usr/lib/gcc/i686-pc-cygwin/4.5.3/../../../../include/w32api/commctrl.h:4134:24: error:
'LPNMDATETIMEFORMATW' has a previous declaration as 'typedef struct
tagNMDATETIMEFORMATW* LPNMDATETIMEFORMATW'
In file included from syswindows.h:634:0,
from sysfile.h:95,
from alloc.c:61:
intl-auto-encap-win32.h:63:36: error: 'PCVOID' has not been declared
intl-auto-encap-win32.h:2303:37: error: 'PCVOID' has not been declared
GNUmakefile:101: recipe for target `alloc.o' failed
make[1]: *** [alloc.o] Error 1
make[1]: Leaving directory `/cygdrive/w/src/xemacs/work/windows-a/src'
GNUmakefile:97: recipe for target `src' failed
make: *** [src] Error 2
----------------------------------------------------------------------
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4:04 p.m.
Mats Lidell writes:
(I just saw configure preferred g++ for some reason but still it is
an
obvious clash with system header files. So removing these structs is
likely the right thing and then we'll see what happens ;-))
That assumes that people are using reasonably recent Cygwin, which
isn't necessarily the case. Remember that not everybody can upgrade
free software because some people live in environments where either
they can't for bureaucratic reasons, or they have other software
sensitive to the version of Cygwin installed and can't easily upgrade.
If the versions of Cygwin where this doesn't happen are all "old
enough", OK, we can do this and deal with fallout later. But I would
tend to think "old enough" is at least 3 years and preferably 5.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:48 p.m.
>>>> Stephen J Turnbull <stephen(a)xemacs.org> writes:
That assumes that people are using reasonably recent Cygwin, which
isn't necessarily the case.
Thanks for correcting my sloppy natural language. By remove I was
referring to "removing them for this particular version". As Jerry
suggest in his patch that is achieved by proper usage of a pre
processor #if-statement.
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
9:09 a.m.
Mats Lidell writes:
>>>>> Stephen J Turnbull <stephen(a)xemacs.org>
writes:
> That assumes that people are using reasonably recent Cygwin, which
> isn't necessarily the case.
Thanks for correcting my sloppy natural language. By remove I was
referring to "removing them for this particular version". As Jerry
suggest in his patch that is achieved by proper usage of a pre
processor #if-statement.
Oh, sorry. I get nervous when people talk about removing code that is
known to be needed to support some older platforms, that's all.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:30 a.m.
I am in the process of getting the Cygwin build to compile. The right
path involves getting make-mswin-unicode.pl to work again (it had
bit-rotted) and running it against an updated intl-encap-win32.c. I
should have a patch ready in a day or two, modulo DAY JOB.
I actually have a patch that seems to work, but in the process I have
broken the native Windows build.
- Vin
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
1:30 p.m.
Hi,
Seems like the Windows build is getting up to speed!
I tried my patch on my Vista machine at work and there "make install"
worked fine. So I'm up and running there! Strange that it fails on XP.
I'll include my patch for completeness. Looking forward to the right fix!
Yours
2013/2/4 Vin Shelton <acs(a)alumni.princeton.edu>
I am in the process of getting the Cygwin build to compile. The
right
path involves getting make-mswin-unicode.pl to work again (it had
bit-rotted) and running it against an updated intl-encap-win32.c. I
should have a patch ready in a day or two, modulo DAY JOB.
I actually have a patch that seems to work, but in the process I have
broken the native Windows build.
- Vin
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
12:13 p.m.
OK, attached are the hacks I needed to get a complete build. Only one
has been discussed here in detail (the ifdef in syswindows.h), the
rest are probably _way_ too brutal and unprincipled -- I'm sending
them not for adoption, but as documentation to help those more
competent than I to narrow things down.
In particular, I remain completely baffled by the need for the
#defines of WM_IME_... in event-msw.c - all my efforts with gcc -E
left me baffled as to why the #defines in w32api/winuser.h were not
getting through.
I'm guessing that the PCVOID -> LPCVOID change is needed because
something is now going wrong with how intl-auto-encap-win32.c includes
syswindows.h, which has the following from Ben in 2010:
/* Some types that show up in Cygwin headers but not in Visual Studio
headers,
and cause problems if we used Cygwin headers to generate
intl-auto-encap-win32.[ch]. */
typedef LPCVOID PCVOID;
Hope this helps!
ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ht(a)inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
6:07 p.m.
OOps -- here's the missing attachment.
ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ht(a)inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
9:51 p.m.
>>>> Henry S Thompson <ht(a)inf.ed.ac.uk> writes:
OOps -- here's the missing attachment.
ht
Thanks. I have come up with a similar patch and it builds OK. The
binary src/xemacs can be used but the "make install" step
fails. Getting some error about
[...]
While compiling toplevel forms in file
/cygdrive/w/src/xemacs/work/windows-a/lisp/cus-edit.el:
!! Symbol's value as variable is void ((ment))
backtrace(nil t)
[...]
I could try your patch but that will take some time. Tomorrow
maybe. So before that I just want to check if you can get the make
install target to build. Does it build?
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
8:45 a.m.
Mats Lidell writes:
Thanks. I have come up with a similar patch and it builds OK. The
binary src/xemacs can be used but the "make install" step
fails. Getting some error about
[...]
While compiling toplevel forms in file
/cygdrive/w/src/xemacs/work/windows-a/lisp/cus-edit.el:
!! Symbol's value as variable is void ((ment))
backtrace(nil t)
[...]
Nope, nothing like that. But note for me the elc files were built
(successfully) by the vanilla 'make' step. . .
I could try your patch but that will take some time. Tomorrow
maybe. So before that I just want to check if you can get the make
install target to build. Does it build?
Yes -- I have no problem with
make install
and the result (emacs-version = "21.5 (beta33) \"horseradish\"
68f8d295be49 XEmacs Lucid") seems to work just fine.
ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ht(a)inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4309
days inactive
4312
days old
16 comments
5 participants
participants (5)
-
Henry S. Thompson -
Jerry James -
Stephen J. Turnbull -
The XEmacs Package Smoketest -
Vin Shelton