Home / Browse / MinGW - Minimalist GNU for Windows / Mail / Archive

MinGW - Minimalist GNU for Windows

Email Archive: mingw-users (read-only)

2000:	_Jan	_Feb	_Mar	_Apr	_May	_Jun	_Jul	_Aug	_Sep (371)	_Oct (167)	_Nov (412)	_Dec (208)
2001:	_Jan (378)	_Feb (302)	_Mar (269)	_Apr (296)	_May (306)	_Jun (381)	_Jul (346)	_Aug (315)	_Sep (195)	_Oct (216)	_Nov (280)	_Dec (227)
2002:	_Jan (309)	_Feb (333)	_Mar (328)	_Apr (407)	_May (517)	_Jun (519)	_Jul (400)	_Aug (580)	_Sep (1273)	_Oct (984)	_Nov (683)	_Dec (538)
2003:	_Jan (578)	_Feb (454)	_Mar (312)	_Apr (366)	_May (505)	_Jun (431)	_Jul (415)	_Aug (374)	_Sep (470)	_Oct (578)	_Nov (372)	_Dec (309)
2004:	_Jan (308)	_Feb (247)	_Mar (372)	_Apr (413)	_May (333)	_Jun (323)	_Jul (269)	_Aug (239)	_Sep (469)	_Oct (383)	_Nov (400)	_Dec (332)
2005:	_Jan (411)	_Feb (363)	_Mar (346)	_Apr (316)	_May (275)	_Jun (248)	_Jul (396)	_Aug (396)	_Sep (279)	_Oct (340)	_Nov (319)	_Dec (218)
2006:	_Jan (317)	_Feb (263)	_Mar (304)	_Apr (296)	_May (209)	_Jun (349)	_Jul (246)	_Aug (198)	_Sep (174)	_Oct (138)	_Nov (201)	_Dec (270)
2007:	_Jan (223)	_Feb (182)	_Mar (350)	_Apr (350)	_May (259)	_Jun (221)	_Jul (299)	_Aug (465)	_Sep (356)	_Oct (265)	_Nov (417)	_Dec (225)
2008:	_Jan (421)	_Feb (327)	_Mar (219)	_Apr (389)	_May (375)	_Jun (262)	_Jul (215)	_Aug (289)	_Sep (257)	_Oct (383)	_Nov (237)	_Dec (209)
2009:	_Jan (232)	_Feb (327)	_Mar (306)	_Apr (251)	_May (146)	_Jun (247)	_Jul (302)	_Aug (252)	_Sep (263)	_Oct (376)	_Nov (270)	_Dec (244)
2010:	_Jan (225)	_Feb (184)	_Mar (300)	_Apr (290)	_May (275)	_Jun (535)	_Jul (192)	_Aug (237)	_Sep (304)	_Oct (142)	_Nov (384)	_Dec (186)
2011:	_Jan (305)	_Feb (337)	_Mar (331)	_Apr (318)	_May (306)	_Jun (299)	_Jul (205)	_Aug (271)	_Sep (232)	_Oct (179)	_Nov (252)	_Dec (216)
2012:	_Jan (195)	_Feb (268)	_Mar (142)	_Apr (226)	_May (203)	_Jun (132)	_Jul (211)	_Aug (429)	_Sep (289)	_Oct (291)	_Nov (182)	_Dec (188)
2013:	_Jan (205)	_Feb (259)	_Mar (224)	_Apr (125)	_May (163)	_Jun	_Jul	_Aug	_Sep	_Oct	_Nov	_Dec


S	M	T	W	T	F	S
1	2	3	4	5	6	7
(5)	(11)	(12)	(9)	(16)	(7)	(10)
8	9	10	11	12	13	14
(17)	(28)	(8)	(15)	(10)	(19)
15	16	17	18	19	20	21
(2)	(19)	(27)	(6)	(26)	(15)	(17)
22	23	24	25	26	27	28
(13)	(22)	(19)	(27)	(18)	(11)	(16)
29	30
(11)	(15)

[Mingw-users] Re: static compilation of library/dll in an executable

From: Andrew Neitsch <yoheythere@ho...> - 2003-06-19 23:59

"Marc Boscher" <marc_boscher@...> wrote:
| ... Is it possible to
| compile myapp statically to include the library and so be able to run on
| its own?

Try adding -static to your command line, e.g.

gcc -static myapp.c CORE_RL_magick_.lib -o myapp.exe

This works with things like -lpdcurses, -lssl, etc.

--Andrew Neitsch

[Mingw-users] Re: Released: mingw32-make-3.80.0-3.exe

From: Parinya Thipchart <thipchart@ya...> - 2003-06-19 22:23

Hi Earnie,

> File: http://prdownloads.sf.net/mingw/mingw32-make-3.80.0-3.exe?download
>
> Fixes the win32 file names being considered targets problem.  I actually
> used this one with dev-c++ to create the helloworld sample.

I have tested this make version with my Makefile (genrerated from MinGW
Studio)
It works fine now.

Thanks very much.

Regards,
Parinya

[Mingw-users] DLL Creating Problem

From: RomikB <romikb@ma...> - 2003-06-19 21:06

Hello All,

I have a problem when creating a dll.
I wrote:

---file.c---
static long ram_memory[0x00200000];
------------

I compile this file:

gcc -c file.c -o file.o
gcc -shared -o file.dll file.o

After this I get a DLL that don't loading.
LoadLibrary returns 0, and GetLastError returns 0.

If I '0x00200000' replace with '0x00002000' for exmpl than a DLL is
working...

Please give me a hint?

PS: It's probably a 'ld' bug.

[Mingw-users] coding practices

From: Henning, Brian <Brian.Henning@na...> - 2003-06-19 16:17

Greetings:
I am new to using cpp and I was wondering someone could
make a comment or two on my coding practices. Is it good
practice to pass args the same as c? how about exiting the
program?=20


thanks,
brian



#include <iostream>

using namespace std;

int main( int argc, char *argv[] ) {
    if( argc !=3D 2 ) {
        cerr << "Usage: " << argv[0] << " <arg>" << endl;
        exit(1);
    }

    cout << arv[1] << endl;
    return 0;
}

Re: [Mingw-users] coding practices

From: <mingw@ca...> - 2003-06-19 17:20

Good coding practices that I can come up with right now:

Create objects when you can.

Many times, switch blocks can be avoided by having classes that inherits virtually from another class.
(Sorry, I dont have a good example).

Pass objects as references. Use the stack in favor of the free storage. (Then, you eliminate the risk of memory leaks).

Use as many private class members as possible.

If you do the same thing with instances of a certain class, consider creating a method for it.

Using cin, cerr and cout in favour of their C equivalents.

Use #include <iostream> instead of #include <iostream.h>, add the following line:
using namespace std;

before you use cin, cerr or cout, because else, you need to call them std::cin, std::cout or std::cerr.
(Many people thinks its the best to call them that and not "pollute the namespace". They would not
think its a good coding practice to use the using statement).

Code practices that also C has:
return from the main function using "return". Always return a value.

Many people thinks its always a good thing to define the main function with the arguments, even if they do not intend to use the
command line arguments.

Always prefer local variables. Many people thinks it is a "crime" to have globals at all.

Use many shorter functions/methods instead of longer. Divide functionality in atomic pieces. Only do one thing in a function/method.

Use descriptive variables names/function names.


I use command line arguments like this:
#include <iostream>
using namespace std;
int main(int argc, char **argv)
{
  if (argc!=2)
  {
     cerr << "Syntax: test <foo>\n";
     return 1;
  }
  cout << "The argument was " << argv[1]);
  return 0;
}



I think many people have different views on this, and much has been written. I usually have a lazy attitude to the demand of clean
code. I prefer to do things instead of planning them ;)

/Anders Lindén


----- Original Message -----
From: "Henning, Brian" <Brian.Henning@...>
To: <MinGW-users@...>
Sent: Thursday, June 19, 2003 6:10 PM
Subject: [Mingw-users] coding practices


Greetings:
I am new to using cpp and I was wondering someone could
make a comment or two on my coding practices. Is it good
practice to pass args the same as c? how about exiting the
program?


thanks,
brian



#include <iostream>

using namespace std;

int main( int argc, char *argv[] ) {
    if( argc != 2 ) {
        cerr << "Usage: " << argv[0] << " <arg>" << endl;
        exit(1);
    }

    cout << arv[1] << endl;
    return 0;
}


-------------------------------------------------------
This SF.Net email is sponsored by: INetU
Attention Web Developers & Consultants: Become An INetU Hosting Partner.
Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission!
INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php
_______________________________________________
MinGW-users mailing list
MinGW-users@...

You may change your MinGW Account Options or unsubscribe at:
https://lists.sourceforge.net/lists/listinfo/mingw-users

Re: [Mingw-users] coding practices

From: Greg Chicares <chicares@mi...> - 2003-06-19 20:07

"Henning, Brian" wrote:
> 
> I am new to using cpp and I was wondering someone could
> make a comment or two on my coding practices. Is it good
> practice to pass args the same as c? how about exiting the
> program?
> 
> #include <iostream>
> 
> using namespace std;
> 
> int main( int argc, char *argv[] ) {
>     if( argc != 2 ) {
>         cerr << "Usage: " << argv[0] << " <arg>" << endl;
>         exit(1);
>     }
> 
>     cout << arv[1] << endl;
>     return 0;
> }

Best C++ coding guidelines anywhere:

http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/*checkout*/boost/boost/more/Attic/coding_guidelines.html?rev=1.1.2.5

Also see

  http://www.comeaucomputing.com/techtalk/

which discusses technical issues (e.g. see its sixth
and eight points) but doesn't really give a stylistic
recommendation.

Best place for this question is really a newsgroup like:

  alt.comp.lang.learn.c-c++

where it's more on topic.

Re: [Mingw-users] fseek with SEEK_SET

From: Oscar Fuentes <ofv@wa...> - 2003-06-19 15:29

mls@... writes:

> when fseek is called, for example, like this:
>
> file_pointer=fopen(filename,"ab");
> fseek(file_pointer,4,SEEK_SET);
>
> it actually seeks to the end of the file instead of the 4th byte.
> Is this a known problem under Windows ?

I guess the problem is this: the 'a' mode value is for *appending*
data. The C RTL just *keeps* the file pointer at the end. 

Have you checked the fseek return value?

> Because compiled with DJGPP it works as expected.

Please correct me if I'm wrong, but DJGPP is the one that seems wrong
here.

A look at the C standard should clarify things, but it's too late
here.

> I'm using MinGW 2.0.0.3, Runtime 3.0 and GCC 3.2.3-20030504-1

-- 
Oscar

[Mingw-users] Re: fseek with SEEK_SET

From: Benjamin Riefenstahl <Benjamin.Riefenstahl@ep...> - 2003-06-19 18:29

Hi Bernd, Oscar,

> mls@... writes:
>> when fseek is called, for example, like this:
>>
>> file_pointer=fopen(filename,"ab");
>> fseek(file_pointer,4,SEEK_SET);
>>
>> it actually seeks to the end of the file instead of the 4th byte.
>> Is this a known problem under Windows ?

Oscar Fuentes <ofv@...> writes:
> A look at the C standard should clarify things, but it's too late
> here.

If I may, C89 says:

  Opening a file with append mode ('a' [...]) causes all subsequent
  writes to the file to be forced to the then current end-of-file,
  regardless of intervening calls to the fseek function.

It doesn't seem to mention how ftell() is supposed to behave, but I
guess that just ignoring fseek() on such a stream is reasonable.

Things get more interesting with "a+b", because that should allow
reading at any place in the file, so with such a stream fseek() can
not be a no-op.

benny

Re: [Mingw-users] Re: fseek with SEEK_SET

From: Oscar Fuentes <ofv@wa...> - 2003-06-19 19:57

Benjamin Riefenstahl <Benjamin.Riefenstahl@...> writes:

>> mls@... writes:
>>> when fseek is called, for example, like this:
>>>
>>> file_pointer=fopen(filename,"ab");
>>> fseek(file_pointer,4,SEEK_SET);
>>>
>>> it actually seeks to the end of the file instead of the 4th byte.
>>> Is this a known problem under Windows ?
>
> Oscar Fuentes <ofv@...> writes:
>> A look at the C standard should clarify things, but it's too late
>> here.
>
> If I may, C89 says:
>
>   Opening a file with append mode ('a' [...]) causes all subsequent
>   writes to the file to be forced to the then current end-of-file,
>   regardless of intervening calls to the fseek function.

Thanks for looking up this, Benny.

> It doesn't seem to mention how ftell() is supposed to behave, but I
> guess that just ignoring fseek() on such a stream is reasonable.
>
> Things get more interesting with "a+b", because that should allow
> reading at any place in the file, so with such a stream fseek() can
> not be a no-op.

Indeed, a fseek, fsetpos or rewind is *required* when you want to read
after a write, or vice-versa. For the specific case of output, you can
use fflush instead, unless you encounter end-of-file. On that last
case no function needs to be called before the output. For every write
operation the implementation positions the file pointer at the end of
the file.

-- 
Oscar

Re: [Mingw-users] _M_IX86

From: Luke Dunstan <coder_infidel@ho...> - 2003-06-19 02:14

----- Original Message ----- 
From: "Danny Smith" <danny_r_smith_2001@...>
To: "Mingw" <mingw-users@...>
Sent: Thursday, June 19, 2003 8:22 AM
Subject: [Mingw-users] _M_IX86

> Hi,
>
> I don't have a MS compiler handy, so I need some help here before
> submitting a patch to GCC to add a builtin define for _M_IX86
>
> My understanding is that (for example) the /G6 switch to MSVC compiler
> optimises the  code for PentiumPro (and PII, PIII). It also sets _M_IX86
> to 600 as an intrinsic define.

Correct.

>
> Question:  Will MSVC compiled  /G6 exes also run on i386, i486, i586?
> Or, in other words, does it add i[456]86-specific asm instructions or
> just tune the scheduling of basic i386 instructions for an i686?
> Or does it do both?

It probably does both, but I don't have any more documentation that what is
available on MSDN. It says "Use this option for programs meant only for the
Pentium Pro", but I don't have proof that it generates Pentium Pro-specific
instructions. I think /G5 would work the same way, but /G4 doesn't because
"The compiler ignores this option".

>
> Secondly, is there another MSVC switch/predefine that indicates SSE/SSE2
support?
>
> Danny

I don't think so. There is a "Processor Pack" for MSVC that adds support for
MMX/SSE/etc. compiler intrinsics but there don't seem to be any compiler
options to enable this, and since the compiler won't generate any
SSE-specific code unless you use the intrinsics, it is the responsibility of
the developer whether to use them. I haven't found any documentation for how
to detect whether the Processor Pack is installed using the preprocessor,
and I haven't found an option equivalent to GCC's "-dM".

Luke

Re: [Mingw-users] _M_IX86

From: Waldo Alvarez <wac@gh...> - 2003-06-19 07:04

Hi:

I am new to mingw (I used to be a msvc user ;P ) but now I use mingw to do everything. After the change there is something I miss from VS. I'm pretty sure it can be done easily with mingw but I don't know how. The thing is that I want to set in my executable the entry point function and bypass the standart libs. I mean strip them completely.

Thanks in advance.
Waldo



-- 
Looking for an Open Source Windows? Then check this out: http://www.reactos.com
Buscando un Windows Open Source? Entonces no te pierdas esto: http://www.reactos.com

Re: [Mingw-users] _M_IX86

From: Luke Dunstan <coder_infidel@ho...> - 2003-06-19 09:07

----- Original Message ----- 
From: "Waldo Alvarez" <wac@...>
To: "Mingw" <mingw-users@...>
Sent: Thursday, June 19, 2003 11:06 AM
Subject: Re: [Mingw-users] _M_IX86

> Hi:
>
> I am new to mingw (I used to be a msvc user ;P ) but now I use mingw to do
everything. After the change there is something I miss from VS. I'm pretty
sure it can be done easily with mingw but I don't know how. The thing is
that I want to set in my executable the entry point function and bypass the
standart libs. I mean strip them completely.
>
> Thanks in advance.
> Waldo

Do you want to use pure Win32 API with no C runtime? If not, it may be
difficult to get the C library to work properly without any startup code.
This probably won't work at with C++ either, because I'm sure the language
requires certain C runtime functionality to use exceptions, etc. I haven't
tried it, but if you use the flags "-Wl,--entry,my_main_func" and
"-nostdlib" then it should be a good start. See the GCC manual though
because it has more information, and it suggests that you also pass -lgcc to
link with libgcc.a.

Luke

Re: [Mingw-users] _M_IX86

From: Oscar Fuentes <ofv@wa...> - 2003-06-19 12:05

Hi, Danni.

Danny Smith <danny_r_smith_2001@...> writes:

> Hi,
>
> I don't have a MS compiler handy, so I need some help here before
> submitting a patch to GCC to add a builtin define for _M_IX86
>
> My understanding is that (for example) the /G6 switch to MSVC compiler 
> optimises the  code for PentiumPro (and PII, PIII). It also sets _M_IX86
> to 600 as an intrinsic define.
>
> Question:  Will MSVC compiled  /G6 exes also run on i386, i486, i586?
> Or, in other words, does it add i[456]86-specific asm instructions or
> just tune the scheduling of basic i386 instructions for an i686?
> Or does it do both?

Well, as you know, the MSDN documentation is somewhat ambiguous. On
the MSVC++ v6 IDE, the tooltil help for "Project" -> "Settings" ->
"C/C++" -> "Code Generation" -> "Processor" says:

"This setting specifies the target platform for which your project is
compiled. This setting changes the value of the /Q compiler switch."

Note that it mentions "/Q" when actually it changes "/G".

Let's continue with the confussion. The Windows Intel compiler is
advertised as a drop-in replacement for MS compiler. My experience
confirms this. They go to the point of being purposely "bug
compatible" and document such "bugs". It's documentation says about
/G:

"The -G5,-G6, and -G7 options optimize your application's performance
for a specific Intel processor. The resulting binary will also run on
the processors listed in the table below:"

The table starts with the Pentium and goes up to the P4. The i386 and
i486 are ignored all along the Intel compiler documentation.

Intel provides a switch for allowing the use of model-specific
instructions. The switch is (surprise!) /Qx. For instance, /QxK is the
Pentium III. The documentation states:

" Caution
If a program compiled with -Qx{M|i|K|W} is executed on a processor
that lacks the specified set of instructions, it can fail with an
illegal instruction exception, or display other unexpected behavior."

My suspicion is that /G is the same for Intel and MS, and that MS v6
lacks /Qx.

> Secondly, is there another MSVC switch/predefine that indicates
> SSE/SSE2 support?

My information about this topic coincides with what Luke and Wu
mentioned.

Regards,

-- 
Oscar

Re: [Mingw-users] _M_IX86

From: <wac@gh...> - 2003-06-19 18:05

On Thu, 19 Jun 2003, Luke Dunstan wrote:

> 
> ----- Original Message ----- 
> From: "Waldo Alvarez" <wac@...>
> To: "Mingw" <mingw-users@...>
> Sent: Thursday, June 19, 2003 11:06 AM
> Subject: Re: [Mingw-users] _M_IX86
> 
> 
> > Hi:
> >
> > I am new to mingw (I used to be a msvc user ;P ) but now I use mingw to do
> everything. After the change there is something I miss from VS. I'm pretty
> sure it can be done easily with mingw but I don't know how. The thing is
> that I want to set in my executable the entry point function and bypass the
> standart libs. I mean strip them completely.
> >
> > Thanks in advance.
> > Waldo
> 

> Do you want to use pure Win32 API with no C runtime? If not, it may be
> difficult to get the C library to work properly without any startup code.
> This probably won't work at with C++ either, because I'm sure the language
> requires certain C runtime functionality to use exceptions, etc. I haven't
> tried it, but if you use the flags "-Wl,--entry,my_main_func" and
> "-nostdlib" then it should be a good start. See the GCC manual though
> because it has more information, and it suggests that you also pass -lgcc to
> link with libgcc.a.
> 
> Luke
> 

Thanks for your help Luke. I'll try those switches today at home. I am  
not using 
stdlib functions, directly the api even to read files or allocate 
and free memory. 

The program used to be only 40 kb compiling it with VC  but with mingw it 
is 96kb, also I'm getting crashes when freeing memory(using the process 
heap) and I guess it is because the stdlib is 
tampering with it.

I also was taking a look at gcc documentation. Pretty extensive and very 
usefull (I did that before asking here) but it does not 
explains very well how to bypass the linking with stdlib. It says 
for the switch -nostdlib something like  "you have to do some things to 
bypass the linking", is very little information  and yes it also suggest 
to link with libgcc.a .

Let's see what I get with --entry. Also using the switch -notdlib I was 
getting a link error (something with exceptions) but 
I commented the protected blocks I was using (try catch RaiseException) 
and I was getting the error anyway. Searching on the Internet I found 
lots of people asking for the same error (some of them using gcc in Linux 
and compiling programs known to work with gcc) and others answered that it 
was a problem with Java exceptions !?!? 

Again, thanks a lot

Best Regards
Waldo

Re: [Mingw-users] Re: [ANNOUCEMENT] binutils 2.14.90-20030612-1 snapshot available

From: Luke Dunstan <coder_infidel@ho...> - 2003-06-19 01:28

Could you provide a small self-contained test case that demonstrates the
problem?

Luke

----- Original Message ----- 
From: "Chris Hansen" <popeofpop@...>
To: <mingw-users@...>
Sent: Tuesday, June 17, 2003 8:27 PM
Subject: RE: [Mingw-users] Re: [ANNOUCEMENT] binutils 2.14.90-20030612-1
snapshot available


> For some reason my EDIT controls doesn?t show up in (resource) dialogs.
> Has something happended to windres?
>
> Regards
> Chris

RE: [Mingw-users] Re: [ANNOUCEMENT] binutils 2.14.90-20030612-1 snapshot available

From: Chris Hansen <popeofpop@so...> - 2003-06-19 17:51

Attachments: test.zip

> -----Original Message-----
> From: Luke Dunstan [mailto:coder_infidel@...]
> Sent: Thursday, June 19, 2003 3:28 AM
> To: Chris Hansen; mingw-users@...
> Subject: Re: [Mingw-users] Re: [ANNOUCEMENT] binutils 2.14.90-20030612-1
> snapshot available
>
> Could you provide a small self-contained test case that demonstrates the
> problem?

Here is a simple project that creates a (resource) dialog with an EDIT
control.

I have compiled it with the latest snapshot of binutils, and no EDIT control
shows up.
I have also tested it with the previous snapshot of binutils (I cannot
remember the version number), and it worked as it should.

Regards
Chris

[Mingw-users] GDB & asynchronous program execution

From: Francois Ferrand <francois.ferrand@en...> - 2003-06-19 17:21

Hi,

I am trying to debug my program with GDB asynchronously, so that I can
interrupt execution at any time... However, it does not seem to run.
Has anybody been able to do this ? Or is it not possible at all on
Windows/Mingw ?

Regards,

-- 
Francois Ferrand

[Mingw-users] Released: mingw32-make-3.80.0-3.exe

From: Earnie Boyd <earnie_boyd@ya...> - 2003-06-19 17:14

File: http://prdownloads.sf.net/mingw/mingw32-make-3.80.0-3.exe?download

Fixes the win32 file names being considered targets problem.  I actually 
used this one with dev-c++ to create the helloworld sample.

Earnie.

This version of make is installed as mingw32-make.exe and is meant to be 
installed into your mingw/ directory.

   GNU make port to mingw32
   Version: 3.80.0
   Production Release-3

1) Manually renamed make.exe to mingw32-make.exe.
2) Added autoconfiscation for w32/ to build an uninstalled libw32.a.
3) Modified configury to add --enable-case-insensitive-file-system.
4) Modified configury to enable WINDOWS32.
5) Modified configury to enable HAVE_DOS_PATHS.
6) This version is compiled with case insensitivity enabled.
7) Modified doc/make.text to change all ifinfo to ifnottex to allow
     use of makeinfo --html.
8) make/configure --prefix-/mingw \
    --enable-case-insensitive-file-system &&
    make CFLAGS='-O3 -s -mms-bitfields -mcpu=pentiumpro'
9) Work around a MinGW-runtime bug.

Re: [Mingw-users] fseek with SEEK_SET - correction

From: WELLS, ROGER K. <ROGER.WELLS@sa...> - 2003-06-19 12:11

<br><font size=3D2 face=3D"sans-serif">Hi,</font>
<br>
<br><font size=3D2 face=3D"sans-serif">The standard is not particularly cle=
ar on this but here goes:</font>
<br>
<br><font size=3D2 face=3D"sans-serif">If you open a file with &quot;ab&quo=
t; it is available only for writing at the end of the file, no reading.</fo=
nt>
<br><font size=3D2 face=3D"sans-serif">If you open it with &quot;a+b&quot; =
you can read as well and the results are as you expect.</font>
<br>
<br><font size=3D2 face=3D"sans-serif">MS documentation is more clear here:=
</font>
<br>
<br><font size=3D3 face=3D"Times New Roman">The character string mode speci=
fies the type of access requested for the file, as follows: </font>
<br><font size=3D3 face=3D"Times New Roman">&quot;r&quot; </font>
<br><font size=3D3 face=3D"Times New Roman">Opens for reading. If the file =
does not exist or cannot be found, the <b>fopen</b> call fails. </font>
<br><font size=3D3 face=3D"Times New Roman">&quot;w&quot; </font>
<br><font size=3D3 face=3D"Times New Roman">Opens an empty file for writing=
. If the given file exists, its contents are destroyed. </font>
<br><font size=3D3 face=3D"Times New Roman">&quot;a&quot; </font>
<br><font size=3D3 face=3D"Times New Roman">Opens for writing at the end of=
 the file (appending) without removing the EOF marker before writing new da=
ta to the file; creates the file first if it doesn't exist. </font>
<br><font size=3D3 face=3D"Times New Roman">&quot;r+&quot; </font>
<br><font size=3D3 face=3D"Times New Roman">Opens for both reading and writ=
ing. (The file must exist.) </font>
<br><font size=3D3 face=3D"Times New Roman">&quot;w+&quot; </font>
<br><font size=3D3 face=3D"Times New Roman">Opens an empty file for both re=
ading and writing. If the given file exists, its contents are destroyed. </=
font>
<br><font size=3D3 face=3D"Times New Roman">&quot;a+&quot; </font>
<br><font size=3D3 face=3D"Times New Roman">Opens for reading and appending=
; the appending operation includes the removal of the EOF marker before new=
 data is written to the file and the EOF marker is restored after writing i=
s complete; creates the file first if it doesn't exist. </font>
<p><font size=3D3 face=3D"Times New Roman">When a file is opened with the &=
quot;a&quot; or &quot;a+&quot; access type, all write operations occur at t=
he end of the file. The file pointer can be repositioned using <b>fseek</b>=
, but is always moved back to the end of the file before any write operatio=
n is carried out. Thus, existing data cannot be overwritten. </font>
<p><font size=3D3 face=3D"Times New Roman">The &quot;a&quot; mode does not =
remove the EOF marker before appending to the file. After appending has occ=
urred, the MS-DOS TYPE command only shows data up to the original EOF marke=
r and not any data appended to the file. The &quot;a+&quot; mode does remov=
e the EOF marker before appending to the file. After appending, the MS-DOS =
TYPE command shows all data in the file. The &quot;a+&quot; mode is require=
d for appending to a stream file that is terminated with the CTRL+Z EOF mar=
ker. </font>
<p><font size=3D3 face=3D"Times New Roman">When the &quot;r+&quot;, &quot;w=
+&quot;, or &quot;a+&quot; access type is specified, both reading and writi=
ng are allowed (the file is said to be open for "update"). However, when yo=
u switch between reading and writing, there must be an intervening <b>fflus=
h</b>, <b>fsetpos</b>, or <b>fseek</b> operation. The current position can =
be specified for the <b>fsetpos</b> or <b>fseek</b> operation, if desired. =
</font>
<p><font size=3D3 face=3D"Times New Roman">In addition to the above values,=
 the following characters can be included in mode to specify the translatio=
n mode for newline characters: </font>
<p><font size=3D3 face=3D"Times New Roman">t </font>
<br><font size=3D3 face=3D"Times New Roman">Open in text (translated) mode.=
 In this mode, CTRL+Z is interpreted as an end-of-file character on input. =
In files opened for reading/writing with &quot;a+&quot;, <b>fopen</b> check=
s for a CTRL+Z at the end of the file and removes it, if possible. This is =
done because using <b>fseek</b> and <b>ftell</b> to move within a file that=
 ends with a CTRL+Z, may cause <b>fseek</b> to behave improperly near the e=
nd of the file. </font>
<p><font size=3D3 face=3D"Times New Roman">Also, in text mode, carriage ret=
urn&#8211;linefeed combinations are translated into single linefeeds on inp=
ut, and linefeed characters are translated to carriage return&#8211;linefee=
d combinations on output. When a Unicode stream-I/O function operates in te=
xt mode (the default), the source or destination stream is assumed to be a =
sequence of multibyte characters. Therefore, the Unicode stream-input funct=
ions convert multibyte characters to wide characters. For the same reason, =
the Unicode stream-output functions convert wide characters to multibyte ch=
aracters. </font>
<p><font size=3D3 face=3D"Times New Roman">b </font>
<br><font size=3D3 face=3D"Times New Roman">Open in binary (untranslated) m=
ode; translations involving carriage-return and linefeed characters are sup=
pressed. </font>
<p><font size=3D3 face=3D"Times New Roman">If t or b is not given in mode, =
the default translation mode is defined by the global variable =5F<b>fmode<=
/b>. If t or b is prefixed to the argument, the function fails and returns =
NULL. </font>
<p><font size=3D3 face=3D"Times New Roman">For more information about using=
 text and binary modes in Unicode and multibyte stream-I/O, see Text and Bi=
nary Mode File I/O and Unicode Stream I/O in Text and Binary Modes. </font>
<p><font size=3D3 face=3D"Times New Roman">c </font>
<br><font size=3D3 face=3D"Times New Roman">Enable the commit flag for the =
associated filename so that the contents of the file buffer are written dir=
ectly to disk if either <b>fflush</b> or <b>=5Fflushall</b> is called. </fo=
nt>
<br><font size=3D3 face=3D"Times New Roman">n </font>
<br><font size=3D3 face=3D"Times New Roman">Reset the commit flag for the a=
ssociated filename to "no-commit." This is the default. </font>
<br><font size=3D2 face=3D"sans-serif"><br>
I hope this helps</font>
<br>
<br><font size=3D2 face=3D"sans-serif">-<br>
Roger Wells, P.E.<br>
SAIC<br>
221 Third St<br>
Newport, RI 02840<br>
401-847-4210 (voice)<br>
401-849-1585 (fax)<br>
roger@...>
</font>
<br>
<br>
<br>
<table width=3D100%>
<tr valign=3Dtop>
<td>
<td><font size=3D1 face=3D"sans-serif"><b>mls@...>
<br><font size=3D1 face=3D"sans-serif">Sent by: mingw-users-admin@...=
rceforge.net</font>
<p><font size=3D1 face=3D"sans-serif">06/18/2003 08:41 PM</font>
<br><font size=3D1 face=3D"sans-serif">Please respond to Bernd Becker</font>
<br>
<td><font size=3D1 face=3D"Arial">&nbsp; &nbsp; &nbsp; &nbsp; </font>
<br><font size=3D1 face=3D"sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; To: &nbs=
p; &nbsp; &nbsp; &nbsp;mingw-user &lt;mingw-users@...;=
</font>
<br><font size=3D1 face=3D"sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; cc: &nbs=
p; &nbsp; &nbsp; &nbsp;</font>
<br><font size=3D1 face=3D"sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; Subject:=
 &nbsp; &nbsp; &nbsp; &nbsp;[Mingw-users] fseek with SEEK=5FSET - correctio=
n</font></table>
<br>
<br>
<br><font size=3D2 face=3D"Courier New">Hi,<br>
<br>
sorry for the first e-mail, it wasn't meant to be sent :-P<br>
Here's now the real problem:<br>
after fseek is called, for example, like this:<br>
<br>
file=5Fpointer=3Dfopen(filename,&quot;ab&quot;);<br>
fseek(file=5Fpointer,4,SEEK=5FSET);<br>
<br>
bytes written with<br>
<br>
putc(byte=5Fto=5Fbe=5Fwritten,file=5Fpointer);<br>
<br>
are appended instead of written at the current file position.<br>
Is this a known problem under Windows ?<br>
Because compiled with DJGPP it works as expected.<br>
<br>
I'm using MinGW 2.0.0.3, Runtime 3.0 and GCC 3.2.3-20030504-1<br>
<br>
<br>
cu,<br>
-- <br>
Bernd<br>
<br>
<br>
<br>
-------------------------------------------------------<br>
This SF.Net email is sponsored by: INetU<br>
Attention Web Developers &amp; Consultants: Become An INetU Hosting Partner=
.<br>
Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission!<br>
INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php<br>
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F<br>
MinGW-users mailing list<br>
MinGW-users@...>
<br>
You may change your MinGW Account Options or unsubscribe at:<br>
https://lists.sourceforge.net/lists/listinfo/mingw-users<br>
<br>
</font>
<br>
<br>

[Mingw-users] fseek with SEEK_SET - correction

From: <mls@mu...> - 2003-06-19 00:45

Hi,

sorry for the first e-mail, it wasn't meant to be sent :-P
Here's now the real problem:
after fseek is called, for example, like this:

file_pointer=fopen(filename,"ab");
fseek(file_pointer,4,SEEK_SET);

bytes written with

putc(byte_to_be_written,file_pointer);

are appended instead of written at the current file position.
Is this a known problem under Windows ?
Because compiled with DJGPP it works as expected.

I'm using MinGW 2.0.0.3, Runtime 3.0 and GCC 3.2.3-20030504-1


cu,
-- 
Bernd

Re: [Mingw-users] fseek with SEEK_SET - correction

From: Oscar Fuentes <ofv@wa...> - 2003-06-19 01:39

mls@... writes:

> sorry for the first e-mail, it wasn't meant to be sent :-P
> Here's now the real problem:
> after fseek is called, for example, like this:
>
> file_pointer=fopen(filename,"ab");
> fseek(file_pointer,4,SEEK_SET);
>
> bytes written with
>
> putc(byte_to_be_written,file_pointer);
>
> are appended instead of written at the current file position.
> Is this a known problem under Windows ?
> Because compiled with DJGPP it works as expected.

This confirms my suspicions. The C RTL used by DJGPP is buggy. MSVCRT
is right. You used the 'a' mode, so data is always appended.

> I'm using MinGW 2.0.0.3, Runtime 3.0 and GCC 3.2.3-20030504-1

-- 
Oscar

Re: [Mingw-users] fseek with SEEK_SET - correction

From: Luke Dunstan <coder_infidel@ho...> - 2003-06-19 02:25

----- Original Message ----- 
From: <mls@...>
To: "mingw-user" <mingw-users@...>
Sent: Thursday, June 19, 2003 8:41 AM
Subject: [Mingw-users] fseek with SEEK_SET - correction

> Hi,
>
> sorry for the first e-mail, it wasn't meant to be sent :-P
> Here's now the real problem:
> after fseek is called, for example, like this:
>
> file_pointer=fopen(filename,"ab");
> fseek(file_pointer,4,SEEK_SET);
>
> bytes written with
>
> putc(byte_to_be_written,file_pointer);
>
> are appended instead of written at the current file position.
> Is this a known problem under Windows ?

This is not a problem at all because it is documented under the fopen()
function and is required by the ANSI C standard. The solution is simply to
not open a file in append mode if you want to do more than append to it, and
you can always fseek() to the end of the file when necessary.

> Because compiled with DJGPP it works as expected.

Then DJGPP is not ANSI C-compliant.

Luke

>
> I'm using MinGW 2.0.0.3, Runtime 3.0 and GCC 3.2.3-20030504-1
>
>
> cu,
> -- 
> Bernd

Re: [Mingw-users] _M_IX86

From: Wu Yongwei <adah@ne...> - 2003-06-19 02:23

Danny Smith wrote:

  > Question:  Will MSVC compiled  /G6 exes also run on i386, i486, i586?
  > Or, in other words, does it add i[456]86-specific asm instructions or
  > just tune the scheduling of basic i386 instructions for an i686?
  > Or does it do both?

I suppose both.  Here is what the old (not the online) MSDN (of Visual
Studio 98) says:

/G6 Pentium Pro Optimizes the code created to favor the Pentium Pro. Use
this option for programs meant only for the Pentium Pro. This option
forces a value of 600 for the _M_IX86 preprocessor macro.

Note its use of "only".  And unlike GCC, MSVC does not have two
different switches to control the CPU-related optimization.

  > Secondly, is there another MSVC switch/predefine that indicates
  > SSE/SSE2 support?

The online MSDN shows that the new Visual Studio .NET also has /G7
(optimizes code for the Intel Pentium 4 or AMD Athlon) and
/arch:[SSE|SSE2] (supports generation of code using the Streaming SIMD
Extensions (SSE) and Streaming SIMD Extensions 2 (SSE2) instructions).

I have not found any information of Predefined Macros regarding SSE(2)

http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vclang/html/_predir_predefined_macros.asp

Best regards,

Wu Yongwei

Re: [Mingw-users] static compilation of library/dll in an executable

From: Luke Dunstan <coder_infidel@ho...> - 2003-06-19 01:20

Attachments: Message as HTML

If CORE_RL_magick_.lib is a static library (or if one is provided by the =
developers), then you can link with it to avoid dependency on a DLL. If =
the file is an import library however, then you need the DLL, and there =
is no way to convert between these types of libraries. You'll have to =
ask the imagemagick developers for further information, but couldn't you =
just put the DLL in the same directory as the .exe?

Luke

  ----- Original Message -----=20
  From: Marc Boscher=20
  To: mingw-users@...
  Sent: Thursday, June 19, 2003 2:37 AM
  Subject: [Mingw-users] static compilation of library/dll in an =
executable


  Hi,

  =20

  I have an application which uses Imagemagick (www.imagemagick.net ) =
libraries and so I must compile it using

  gcc myapp.c CORE_RL_magick_.lib -o myapp.exe

  in order to link the api (I also include api.h the header file for the =
imagemagick api).

  I need to run this app on a remote server that doesn't have =
imagemagick installed on it and so I get missing dll errors. Is it =
possible to compile myapp statically to include the library and so be =
able to run on its own?

  =20

  Thanks

  Marc

Re: [Mingw-users] postgres

From: Oscar Fuentes <ofv@wa...> - 2003-06-19 01:09

"Henning, Brian" <Brian.Henning@...> writes:

> does a c library for postgres exist for mingw?

Google is your friend:

http://www.google.com/search?q=postgres+mingw&num=100

-- 
Oscar

Next Messages