headers
Aaron W. LaFramboise | 6 Sep 2004 02:45

fixincludes causing GCC mainline bootstrap to fail.

make[1]: Entering directory
`/aaronwl/cs/compilers/gcc/build/mingw-head/build-i686-pc-mingw32/fixincludes'
gcc -c -g -O2   -DHAVE_CONFIG_H -I.
-I../../../../src/cvs/head/gcc/fixincludes -I../include
-I../../../../src/cvs/head/gcc/fixincludes/../include
../../../../src/cvs/head/gcc/fixincludes/fixincl.c
../../../../src/cvs/head/gcc/fixincludes/fixincl.c: In function
`initialize':
../../../../src/cvs/head/gcc/fixincludes/fixincl.c:316: error: `SIGQUIT'
undeclared (first use in this function)
../../../../src/cvs/head/gcc/fixincludes/fixincl.c:316: error: (Each
undeclared identifier is reported only once
../../../../src/cvs/head/gcc/fixincludes/fixincl.c:316: error: for each
function it appears in.)
../../../../src/cvs/head/gcc/fixincludes/fixincl.c:323: error: `SIGALRM'
undeclared (first use in this function)
../../../../src/cvs/head/gcc/fixincludes/fixincl.c: In function
`create_file':
../../../../src/cvs/head/gcc/fixincludes/fixincl.c:578: error: too many
arguments to function `mkdir'
../../../../src/cvs/head/gcc/fixincludes/fixincl.c: In function
`internal_fix':
../../../../src/cvs/head/gcc/fixincludes/fixincl.c:861: error: `F_DUPFD'
undeclared (first use in this function)
make[1]: *** [fixincl.o] Error 1
make[1]: Leaving directory
`/aaronwl/cs/compilers/gcc/build/mingw-head/build-i686-pc-mingw32/fixincludes'

Was fixincludes ever being built for i?86-pc-mingw before it the recent
changes to put it in a toplevel directory?
(Continue reading)

Permalink | Reply | 0 comments |
headers
Danny Smith | 6 Sep 2004 08:51
Picon

Fw: [Mingw-users] Released: w32api-3.1.tar.gz

From: "Earnie Boyd"
|
| I've released the 3.1 version of the MinGW Windows API.

Earnie,
Are you planning on doing a full Mingw-?.?  release, soon?  I'm not
pushing for one, just asking since you mentioned something about that
earlier.

If so, there are a few bugs in gcc-3.4.1.  I can either do a 3.4.1
bugfix upload or or a 3.4.2 bugfix upload (I have time to do either on
10 Sept, pm NZ).

Thanks for fixing my typo in mingwex makefile.

Danny

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click
Permalink | Reply | 1 comment |
headers
Earnie Boyd | 6 Sep 2004 15:28
Picon

Re: Fw: [Mingw-users] Released: w32api-3.1.tar.gz

Danny Smith wrote:

>From: "Earnie Boyd"
>|
>| I've released the 3.1 version of the MinGW Windows API.
>
>Earnie,
>Are you planning on doing a full Mingw-?.?  release, soon?  I'm not
>pushing for one, just asking since you mentioned something about that
>earlier.
>
>  
>
Yes, but I would like to upgrade the previous binutils Candidate to Current.

>If so, there are a few bugs in gcc-3.4.1.  I can either do a 3.4.1
>bugfix upload or or a 3.4.2 bugfix upload (I have time to do either on
>10 Sept, pm NZ).
>
>  
>
I would like 3.4.1 or 3.4.2 to Current, you decide which.

>Thanks for fixing my typo in mingwex makefile.
>  
>
You're welcome.

Earnie.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Aaron W. LaFramboise | 10 Sep 2004 20:56

Marking mingwrt headers with system_header

I've noticed that only one of the mingwrt/include headers is marked as a
system header (math.h), yet many of them are generally unclean against
misc. warning types (such as -Wmissing-prototype) and possible could
never be completely clean (such as with -Winline).  Should the headers
be marked with #pragma GCC system_header?

Cleaning them up against things like -Wmissing-prototype might be a good
idea anyway, though.

Aaron W. LaFramboise

-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM. 
Deadline: Sept. 13. Go here: http://sf.net/ppc_contest.php
Permalink | Reply | 5 comments |
headers
Danny Smith | 11 Sep 2004 00:13
Picon

Re: Marking mingwrt headers with system_header


----- Original Message -----
From: "Aaron W. LaFramboise

| I've noticed that only one of the mingwrt/include headers is marked as
a
| system header (math.h), yet many of them are generally unclean against
| misc. warning types (such as -Wmissing-prototype) and possible could
| never be completely clean (such as with -Winline).  Should the headers
| be marked with #pragma GCC system_header?

If you wish.  Nornally they are considered system headers anyway by
virtue of their location in a system include dir, but of course they may
be installed elsewhere.

The math.h header was marked specifically for the __builtin_scalbn
conflict

|
| Cleaning them up against things like -Wmissing-prototype might be a
good
| idea anyway, though.

One of the biggies is cleaning up the __CRT_INLINE definitions in
ctype.h.  These should really be static inline for ISO C99, but marking
as such would cause warnings with __declspec(dllimport)  prototypes.

Danny
|
| Aaron W. LaFramboise
(Continue reading)

Permalink | Reply | 4 comments |
headers
Aaron W. LaFramboise | 11 Sep 2004 01:35

Re: Marking mingwrt headers with system_header

Danny Smith wrote:
> | Cleaning them up against things like -Wmissing-prototype might be a
> good
> | idea anyway, though.
> 
> One of the biggies is cleaning up the __CRT_INLINE definitions in
> ctype.h.  These should really be static inline for ISO C99, but marking
> as such would cause warnings with __declspec(dllimport)  prototypes.

Can you elaborate on this problem?  How are inline and dllimport
supposed to work together?

Aaron W. LaFramboise

-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM. 
Deadline: Sept. 13. Go here: http://sf.net/ppc_contest.php
Permalink | Reply | 3 comments |
headers
Danny Smith | 11 Sep 2004 02:13
Picon

Re: Marking mingwrt headers with system_header

From: "Aaron W. LaFramboise
| Danny Smith wrote:
| > | Cleaning them up against things like -Wmissing-prototype might be
a
| > good
| > | idea anyway, though.
| >
| > One of the biggies is cleaning up the __CRT_INLINE definitions in
| > ctype.h.  These should really be static inline for ISO C99, but
marking
| > as such would cause warnings with __declspec(dllimport)  prototypes.
|
| Can you elaborate on this problem?  How are inline and dllimport
| supposed to work together?

Its not so much the dllimport but the extern that it implies that causes
warnings (harmless in this case because there are no intervening refs to
the extern symbol)

__declspec(dllimport) void foo(void);
static inline void foo(void) {}

> gcc -Wall  -c inline.c
> inline.c:2: warning: static declaration of 'foo' follows non-static
declaration
> inline.c:1: warning: previous declaration of 'foo' was here

You'll get the same with any extern declaration

Danny
(Continue reading)

Permalink | Reply | 2 comments |
headers
Aaron W. LaFramboise | 11 Sep 2004 03:10

Re: Marking mingwrt headers with system_header

Danny Smith wrote:
> Its not so much the dllimport but the extern that it implies that causes
> warnings (harmless in this case because there are no intervening refs to
> the extern symbol)
> 
> __declspec(dllimport) void foo(void);
> static inline void foo(void) {}

Well, I've noticed that MSVC 7.1 does not support inline in C: only
__inline, which has C++ semantics, not C semantics, so they aren't
setting a useful example.

However, would it make sense for GCC to treat dllimport inline functions
as having implicit internal linkage?

What is the 'right way' to make this apparently legitimate usage work?

Aaron W. LaFramboise

-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM. 
Deadline: Sept. 13. Go here: http://sf.net/ppc_contest.php
Permalink | Reply | 1 comment |
headers
Danny Smith | 11 Sep 2004 03:27
Picon

Re: Marking mingwrt headers with system_header

From: "Aaron W. LaFramboise"

| However, would it make sense for GCC to treat dllimport inline
functions
| as having implicit internal linkage?

Not to me.

|
| What is the 'right way' to make this apparently legitimate usage work?
|

It works now with the "extern inline" GNU extension.  If we change it to
static inline, we get legitimate warnings, just as you do when _CRTIMP
is defined to nothing.  We could always  turn them into macros if not
__cplusplus

Danny

| Aaron W. LaFramboise
|
|
|
| -------------------------------------------------------
| This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
| Project Admins to receive an Apple iPod Mini FREE for your judgement
on
| who ports your project to Linux PPC the best. Sponsored by IBM.
| Deadline: Sept. 13. Go here: http://sf.net/ppc_contest.php
| _______________________________________________
(Continue reading)

Permalink | Reply | 0 comments |
headers
Earnie Boyd | 8 Sep 2004 16:14
Picon

GCC Compiled Headers

Should we support an upload file of w32api and mingw-runtime .gch files?

What is the ramification of:

$ gcc windows.h
windows.h:16:13: warning: #pragma system_header ignored outside include file

$ gcc -v
Reading specs from C:/msys/unified/bin/../lib/gcc/mingw32/3.4.1/specs
Configured with: ../gcc/configure --with-gcc --with-gnu-ld --with-gnu-as 
--host=mingw32 --target=mingw32 --prefix=/mingw --enable-threads 
--disable-nls --enable-languages=c,c++,f77,ada,objc,java 
--disable-win32-registry --disable-shared --enable-sjlj-exceptions 
--enable-libgcj --disable-java-awt --without-x --enable-java-gc=boehm 
--disable-libgcj-debug --enable-interpreter 
--enable-hash-synchronization --enable-libstdcxx-debug
Thread model: win32
gcc version 3.4.1 (mingw special)

Earnie

--

-- 
http://www.mingw.org
http://sourceforge.net/projects/mingw
https://sourceforge.net/donate/index.php?user_id=15438

-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM.
(Continue reading)

Permalink | Reply | 0 comments |

Gmane