mingw-users Mailing List for MinGW - Minimalist GNU for Windows

"Andres Rand" <lo...@ho...> writes:
> My old Borland Builder used to make simple application (a window and
> a menu) around 140kb, I couldn't get anything less from that.

Apples and oranges here, of course.  Current BC++ (the freely
available command-line version) compiles your code to 6,656 bytes.
That is using the DLL version of the RTL similar to Mingw's use of
MSCRTL.DLL.  I'm sure previous versions didn't do very much worse.

> > Hello,
> > We have developed a superset of C interpreter called Ch to run Win32
> > programs. The header file from Mingw is distributed with
> > our Free Ch Windows toolkits.
> > 
> > Ch and Windows toolkit is availabe at
> > http://www.softintegration.com. Ch also supports
> > POSIX, X11/Motif, OpenGL, ODBC, GTK+, and CGI. You are welcome to
> > have a try. Any comments or suggestions are appreciated.
> > 
> > Based on Readme file, we need to notify the author about it.
> > However, the email address ano...@he... doesn't seem
> > to work. Sorry if it is the right place to post such a news.
> > 
> 
> I noticed.  I will be changing the license for a future release to
> remove that "restriction" of use.  Consider your obligation of
> notification fulfilled.
> 
> Earnie.

Hi Earnie,

Thanks for your information. If possible, I would like to check
if you can add a link to our company in MinGW related website
at http://www.mingw.org/links.shtml.

Your consideration is appreciated.

Best regards,

Xiaodong Zhou
http://www.softintegration.com

You'll need to supply -lwsock32.

Earnie.

David Reiss wrote:
> 
> I'm trying to build gaim (gaim.sourceforge.net) to be run on windows.  I've got the latest mingw
> and msys, but running the configure script gives me the following error:
> 
> checking for inet_aton... no
> checking for inet_aton in -lresolv... no
> configure: error: inet_aton not found
> 
> It would seem to me that inet_aton would be a simple enough function to write.  Does anyone know
> how I can fix this problem?  Thanks.
> 
> __________________________________________________
> Do You Yahoo!?
> Yahoo! - Official partner of 2002 FIFA World Cup
> http://fifaworldcup.yahoo.com
> 
> _______________________________________________________________
> 
> Don't miss the 2002 Sprint PCS Application Developer's Conference
> August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm
> 
> _______________________________________________
> MinGW-users mailing list
> Min...@li...
> 
> You may change your MinGW Account Options or unsubscribe at:
> https://lists.sourceforge.net/lists/listinfo/mingw-users

I'm trying to build gaim (gaim.sourceforge.net) to be run on windows.  I've got the latest mingw
and msys, but running the configure script gives me the following error:

checking for inet_aton... no
checking for inet_aton in -lresolv... no
configure: error: inet_aton not found

It would seem to me that inet_aton would be a simple enough function to write.  Does anyone know
how I can fix this problem?  Thanks.

__________________________________________________
Do You Yahoo!?
Yahoo! - Official partner of 2002 FIFA World Cup
http://fifaworldcup.yahoo.com

Hi,

I looked into it a bit further. I'm sure most of the time there's no good
reason not to link to a runtime, but occasionally it can handy. Anyway, the
'-nostdlib' option lets you do it. The gcc docs say the compiler can insert
calls to memcmp, memset & memcpy, so you'd have to write these as well as a
startup.

~$ cat test.c
#include <windows.h>

void __cdecl WinMainCRTStartup(void) {
    ExitProcess(WinMain(GetModuleHandle(NULL), NULL, "", SW_SHOWDEFAULT));
}

int WINAPI WinMain( HINSTANCE hInstance,
                    HINSTANCE hPrevInstance,
                    LPSTR lpCmdLine,
                    int nCmdShow    )
{
    MessageBox(NULL, "Test", "Test", MB_OK);
    return 0;
}

~$ gcc -o test test.c -nostdlib -e_WinMainCRTStartup -lkernel32 -luser32
~$

On Fri, 31 May 2002, John Brown wrote:

> If a few people get together here and there and provide free stuff as their 
> contribution towards making the world a better place, that's OK. But if 
> everybody did that, it simply would not work.
> 
> The programmer would have to get free groceries from the grocer so tht he 
> can develop free software. The grocer would have to get free produce from 
> the farmer so that the programmer can get free groceries. The farmer would 
> have to get free tools from the hardware store. But if the produce is free, 
> what is to stop the farmer from taking away several truckloads of 
> vegetables? Trucks are free. He can take ownership of as many trucks as he 
> wants, if he can find them. What would stop the programmer from 
> appropriating all the potato chips for himself? It would get ugly.
> 

among the wealth of amusing stories this week, this one takes
the cake (or is it potato chips?).

M Joshua Ryan
Unfortunatly people forget that the first wheel was actually square until Ogg reinvented it, this time making it round.

José Fonseca wrote:

>And I'm not a capitalist pig nor do I work for one: I'm a Ph.D student >of 
>mechanical engineering and I get my money from my research. I program >on 
>my free-time, as I've been doing since I was 10 year old.
...


This is my point exactly. By day, Bruce Wayne runs the Wayne Foundation. By 
night, he is Batman, the scourge of evildoers everywhere. If he gave up his 
day job to fight crime full-time (which is noble, but unpaid work for him), 
he would not have the funds to buy gas for the Batmobile, much less the 
instruments and the weapons.

Similarly, by day, you are a meek, mild-mannered researcher. By night you 
can be Captain Coder, fearless defender of the Free Software Movement, but 
only because the research pays your bills.

In  short:

1) Money is not evil. It is not even the root of all evil.

>(And I must >say I found your comment rather insulting).

2) We all have to get money from somewhere. That is what I meant. No insult 
was intended.

3) If Danny, Earnie, and company quit their day jobs, they would be able to 
do more. They might even tell fewer people to "submit a patch", because they 
would now be in a position to do it themselves. That is, until they start to 
get weak and finally die of starvation. We cannot support any cause, noble 
or otherwise, if we work full-time for free.

4) There is a limit to what you can achieve working part-time, especially 
since, ,if you are like me, your greedy capitalist pig of an employer gets 
the best hours of your day. It is generally speaking, impossible for amateur 
atheletes to compete with professionals, because of the demands of their day 
jobs, lack of funding, etc.

5) I still maintain that the main reason for interest in Linux, Java, mingw, 
etc. is that they are free (and offer some hope of breaking the dominance of 
the evil M$). If Linux cost the same as UNIX, it would not look nearly as 
sexy as it does now. Likewise for mingw and M$ VC++. A price increase 
reduces demand. Simple economics.

If a few people get together here and there and provide free stuff as their 
contribution towards making the world a better place, that's OK. But if 
everybody did that, it simply would not work.

The programmer would have to get free groceries from the grocer so tht he 
can develop free software. The grocer would have to get free produce from 
the farmer so that the programmer can get free groceries. The farmer would 
have to get free tools from the hardware store. But if the produce is free, 
what is to stop the farmer from taking away several truckloads of 
vegetables? Trucks are free. He can take ownership of as many trucks as he 
wants, if he can find them. What would stop the programmer from 
appropriating all the potato chips for himself? It would get ugly.


_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com

> I hope I could contribute ideas and opinions to the free community. I hope
> people (esp. developers) would appreciate these ideas and opinions. If
they
> could not (or have not time) do it, tell me (or put it off to a later
date).
> Don't use "submit a patch" to frustrate people contributing ideas and
> opinions. Use it only when you know the other person can do it, or
probably
> can do. Maybe "would you ...". Maybe I am just too sentimental.

<soapbox>
I do not think you are being too sentimental. Rather I believe that is a
real world view. If one is capable of submitting a patch and he/she has the
time, the individual should do so. This may be a view of the world through
rose colored glasses, but I firmly believe that developers who have the
ability and the time will submit patches. What needs to be avoided by the
key players or "overseers" of any freeware project is the use of the phrase,
"Well submit a patch" as a device to end a discussion. Judging from the
comments posted, here that while there may be a significant number of people
who post messages here who are capable of making patches to the mingw suite,
they are very much in the minority. I believe that there are three major
determinants which impact an individual's ability to create a usable patch.

First, the person has to have the time to dig into the source code and
become damn near as familiar with the internal functioning of the program as
the developers who may have spent years writing the code.

Second, he/she must possess a high level of proficiency in the source
language, the target language, and assembly code for the target processor.

Third, one must also understand compilers and how they work.

When you apply those three criteria, the number of people who are truly
qualified to comply with the directive, "Well just submit a patch" shrinks
almost to the point of invisibility. Recently someone pointed out that more
than 100,000 downloads of the mingw system had been counted on the
SourceForge site. That probably relates to some were between 50,000 and
75,000 different people. That is a fairly large installed base for a highly
specialized product. The key question, however, is what is the level of
proficiency of these users? In what percentile would they rank in the three
factors above? Since mingw is free, many who download it, like me for
example, are trying to learn the C++ language. How many? I suspect the
number is at least 50%.

What is all this leading up to? Well, I would like to suggest two things.
First, the use of the phrase "submit a patch" either needs to end or needs
some modification such as:  "If you have the ability to write a patch,
please do so and submit it. If not please let me know by email." Second,
those who are capable of making the patch should do so and submit it.

One last comment unrelated to the discussion above I was the person who made
the chiding comment about freeware changing the world. Let's take the
premise to its logical conclusion, i.e., all software is free. Let's further
assume that all software has been free for 20 years. What are the changes?
Is the United States now a third world country? I doubt it. Has the free
software brought wealth to India and China? Again, I doubt it. What has
happened? Contrary to the populist view I see computers and software
development stagnating and becoming as fragmented as it was during the early
phases of the PC revolution. Why? Because the profit motive has been removed
from the equation. When that happens, capital will move to places where it
can earn a return on investment and all that will be left are hobbyists to
develop software.

That said, I strongly, strongly, strongly believe in the importance of
freeware projects such as mingw, dev-c++, GNU, etc. Just like amateur or ham
radio, innovation bubbles through the freeware movement and, as a result,
the leading edge of computer science or software development is often found
in various freeware projects.
</soapbox>

Regards,

Bob Wilson

> You should create a bug report to binutils for this.

Sorry, I may have misunderstood you, if you meant "you should file
a bug at sourceforge in the mingw section", then I'm all for it
and will do so.

Regards,
Wolfram.

> You should create a bug report to binutils for this.

Eh, no, the non-portable stricmp() is only introduced in the
binutils-2_12_90-20020518-1.diff patch for mingw, AFAICS.

Regards,
Wolfram.

Wolfram Gloger wrote:
> 
> > i assume, it is already know, but while trying to build a linux->mingw cross
> > compiler i came across this:
> > To compile binutils, i had to change
> > stricmp into strcmp in ld/pe-dll.c:516
> 
> For Linux, it's probably more correct to replace stricmp() with
> strcasecmp().  It's at least SUS-compliant..
> 

You should create a bug report to binutils for this.

Earnie.

> i assume, it is already know, but while trying to build a linux->mingw cross
> compiler i came across this:
> To compile binutils, i had to change
> stricmp into strcmp in ld/pe-dll.c:516 

For Linux, it's probably more correct to replace stricmp() with
strcasecmp().  It's at least SUS-compliant..

Regards,
Wolfram.

xz...@so... wrote:
> 
> Hello,
> 
> We have developed a superset of C interpreter called Ch to run Win32
> programs. The header file from Mingw is distributed with
> our Free Ch Windows toolkits.
> 
> Ch and Windows toolkit is availabe at
> http://www.softintegration.com. Ch also supports
> POSIX, X11/Motif, OpenGL, ODBC, GTK+, and CGI. You are welcome to
> have a try. Any comments or suggestions are appreciated.
> 
> Based on Readme file, we need to notify the author about it.
> However, the email address ano...@he... doesn't seem
> to work. Sorry if it is the right place to post such a news.
> 

I noticed.  I will be changing the license for a future release to
remove that "restriction" of use.  Consider your obligation of
notification fulfilled.

Earnie.

Hiya,

once I began w/ the MingW32 stuff I found docu written by Collin 
Peters there's a html-page saying which fn is in which lib.

One may query google:
http://www.google.com/search?q=%22Find+the+Library+for+a+Win32+API+Function%22&sourceid=opera&num=0

Unfortunately this page is no longer maintained and Collin's 
address is no longer in order :-(

Kind regards,
Jan
-- 
 Jan Wegner                          Software Engineer 
 Zuken GmbH                          Tel: +49 5251 150627
 Vattmannstrasse 3                   Fax: +49 5251 150700
 D-33100 Paderborn                   http://www.zuken.com
----*... and may all your GUIs be fast and friendly.*----

Hi,
i assume, it is already know, but while trying to build a linux->mingw cross
compiler i came across this:
To compile binutils, i had to change
stricmp into strcmp in ld/pe-dll.c:516 

Bye,
Martin.

-- 
The early bird catches the worm. If you want something else for       
breakfast, get up later.

greets. :)

>> seem to get the feeling that mingw might end up trying to be another
>> unixish emulation layer for windows. hope it doesn't.
>> 
>
>Nope, this thread is concerning alternatives.  It is a separate
>project.  Opinions where being asked.

ahh.. maybe i wasn't paying attention.. :)
good to know this.
hoorah mingw! :P


rgds,

kh

Hello,

We have developed a superset of C interpreter called Ch to run Win32
programs. The header file from Mingw is distributed with
our Free Ch Windows toolkits. 

Ch and Windows toolkit is availabe at
http://www.softintegration.com. Ch also supports
POSIX, X11/Motif, OpenGL, ODBC, GTK+, and CGI. You are welcome to 
have a try. Any comments or suggestions are appreciated.

Based on Readme file, we need to notify the author about it.
However, the email address ano...@he... doesn't seem
to work. Sorry if it is the right place to post such a news.

Best regards, 

Xiaodong Zhou
http://www.softintegration.com

Hi Wu et al.

I think I know whare you are coming from.

> Surely there are very different kinds of people working for open software.

Sure are.

My perspective on the "submit a patch" discussion is as a user of free
software dev tools who works intermittently on free software applications.
Those are Grass, a geographic information system (Cygwin), Gnu Common Lisp
(Mingw32), and Mosmlgl, a binding to Open GL for Moscow ML and others.

I use the word intermittently because these projects spread my time thinly
in the cracks between wife, children and paid work and except for Gnu Common
Lisp, I work on these projects because I use them at home, and because I
believe that free software is a Good Thing for Humanity.

Jose's excellent Dr Mingw played a pivotal role in keeping the current
Mingw32 Windows port of Gnu Common Lisp running, which in turn allowed the
current Maxima version (computer algebra, my real interest in that regard)
to continue to work on the Windows platform.

This is a typical example of a chain of interdependencies between projects,
and although the words "submit a patch" are probably not meant to piss
people off, they can have a greater effect than intended.

This greater effect is caused in my case by the internal conflict between
the desire to contribute and the need for a "life".  There is no way I can
possibly spend more time on improving free software, (I sometimes find
myself working at three in the morning) yet I am fairly devoted to the
"cause" on a moral and practical level.  I also recognise my own areas of
expertise, which definitely don't include the guts of C compilers, so I am
dependent on those very competent experts in the Cygwin and Mingw32 worlds.

Cheers, and thanks for all the great work

Mike Thomas.

> The major open-source contributors I know in China contributed their work
> when they were post-graduate students.
>
> Free software HAS changed the world. Though, as I compared it to
communism,
> I don't feel it mature enough to enter people's normal life. The first
> lesson Eric Raymond learnt from Linux in his famous The Cathedral and the
> Bazaar is:
>
>   Every good work of software starts by scratching a developer's personal
>   itch.
>
> Since developers write software out of his own interest, the software is
> generally high-quality. Since developers write software out of his own
> interest, it is possible he will not care much what the users want; which
> great software companies generally do very well. These are my worries.
>
> I hope I could contribute ideas and opinions to the free community. I hope
> people (esp. developers) would appreciate these ideas and opinions. If
they
> could not (or have not time) do it, tell me (or put it off to a later
date).
> Don't use "submit a patch" to frustrate people contributing ideas and
> opinions. Use it only when you know the other person can do it, or
probably
> can do. Maybe "would you ...". Maybe I am just too sentimental.
>
> Too bad I am posting in THIS list, which is the best open-source community
I
> know, though I don't know many.
>
> So much for this thread.
>
> Best regards for all, esp. for Jose, Danny, and Earnie,
>
> Wu Yongwei
>
>
> _______________________________________________________________
>
> Don't miss the 2002 Sprint PCS Application Developer's Conference
> August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm
>
> _______________________________________________
> MinGW-users mailing list
> Min...@li...
>
> You may change your MinGW Account Options or unsubscribe at:
> https://lists.sourceforge.net/lists/listinfo/mingw-users
>

Curious.

On 31 May 2002 at 10:08, Wu Yongwei wrote:

> Since I started this off-topic thread, I think I need to say some more
> words.
> 
> Firstly, I was just complaining. Complaint from a somehow old corporate guy
> find not enough time to study the auto* tools and to dig into the free
> software.
> 
> Surely there are very different kinds of people working for open software.
> The major open-source contributors I know in China contributed their work
> when they were post-graduate students.
> 
> Free software HAS changed the world. Though, as I compared it to communism,
> I don't feel it mature enough to enter people's normal life. The first
> lesson Eric Raymond learnt from Linux in his famous The Cathedral and the
> Bazaar is:
> 
>   Every good work of software starts by scratching a developer's personal
>   itch.
> 
> Since developers write software out of his own interest, the software is
> generally high-quality. Since developers write software out of his own
> interest, it is possible he will not care much what the users want; which
> great software companies generally do very well. These are my worries.
> 
> I hope I could contribute ideas and opinions to the free community. I hope
> people (esp. developers) would appreciate these ideas and opinions. If they
> could not (or have not time) do it, tell me (or put it off to a later date).
> Don't use "submit a patch" to frustrate people contributing ideas and
> opinions. Use it only when you know the other person can do it, or probably
> can do. Maybe "would you ...". Maybe I am just too sentimental.

	Not sure what you are trying to say.

> 
> Too bad I am posting in THIS list, which is the best open-source community I
> know, though I don't know many.

	Again, not sure what you are saying.  I hear you are complimenting the community, and yet it sounds like you 
are dissatisfied with something about the community.  Again, I am not sure what it is you are concerned with or 
attempting to say here.

	Is it possible that maybe others are just as confused as I am?

> 
> So much for this thread.
> 
> Best regards for all, esp. for Jose, Danny, and Earnie,
> 
> Wu Yongwei

	Paul G.

Since I started this off-topic thread, I think I need to say some more
words.

Firstly, I was just complaining. Complaint from a somehow old corporate guy
find not enough time to study the auto* tools and to dig into the free
software.

Surely there are very different kinds of people working for open software.
The major open-source contributors I know in China contributed their work
when they were post-graduate students.

Free software HAS changed the world. Though, as I compared it to communism,
I don't feel it mature enough to enter people's normal life. The first
lesson Eric Raymond learnt from Linux in his famous The Cathedral and the
Bazaar is:

  Every good work of software starts by scratching a developer's personal
  itch.

Since developers write software out of his own interest, the software is
generally high-quality. Since developers write software out of his own
interest, it is possible he will not care much what the users want; which
great software companies generally do very well. These are my worries.

I hope I could contribute ideas and opinions to the free community. I hope
people (esp. developers) would appreciate these ideas and opinions. If they
could not (or have not time) do it, tell me (or put it off to a later date).
Don't use "submit a patch" to frustrate people contributing ideas and
opinions. Use it only when you know the other person can do it, or probably
can do. Maybe "would you ...". Maybe I am just too sentimental.

Too bad I am posting in THIS list, which is the best open-source community I
know, though I don't know many.

So much for this thread.

Best regards for all, esp. for Jose, Danny, and Earnie,

Wu Yongwei

I think both versions of gcc generate stabs debugging information by default
on the Mingw platform. At least, I am able to use gdb 5.1.1 to debug test
programs compiled with gcc 3.1, so would it be possible to provide a small
test case that crashes gdb? Versions of binutils and other relevant tools
would be useful too.

Luke Dunstan

----- Original Message -----
From: "Holger Vogt" <hol...@un...>
To: <min...@li...>
Sent: Friday, May 31, 2002 4:37 AM
Subject: [Mingw-users] gcc-3.1 and gdb-5.1.1


> Hi,
>
> gdb (gdb-5.1.1-1-mingw.tar.gz) crashes when I try to set a breakpoint to
> my application (SystemC). The same happens when I use Insight (gdb-5.0)
> from within CYGWIN.
>
> A sample application compiled with gcc-2.95 runs o.k. under both gdb
> versions.
>
> http://gcc.gnu.org/gcc-3.1/changes.html states that at least gdb-5.1.1
> is required for DWARF2 debugging format made by gcc-3.1.
>
> Is there a switch to force gcc to use stabs as debugging format? Or is
> there any other idea or experience to run gdb correctly?
>
> By the way gcc-3.1 allowed me to successfully port SystemC
> (www.systemc.org) to MINGW, which failed with runtime errors due to
> incorrect stack after switching fibers in earlier versions.
>
> Regards
>
> Holger Vogt
>
>
> _______________________________________________________________
>
> Don't miss the 2002 Sprint PCS Application Developer's Conference
> August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm
>
> _______________________________________________
> MinGW-users mailing list
> Min...@li...
>
> You may change your MinGW Account Options or unsubscribe at:
> https://lists.sourceforge.net/lists/listinfo/mingw-users
>

The Mingw startup code (in crt2.o, etc.) needs to call some functions in
MSVCRT.DLL to initialise various things like the stdio handles (stdout,
stdin, stderr). If you don't want your application to depend on MSVCRT at
all, you would need to modify the startup code in mingw-runtime and build
your own version. However, code generated by GCC (and libgcc) may call a few
standard C functions, so alternative implemenations for these would also
need to be supplied. With these modifications, it would be possible to add
an option for this to GCC, but I don't think this is a goal of the Mingw
project so somebody would have to do this separately. This is my opinion so
other developers should correct me if I am wrong.

Luke Dunstan

----- Original Message -----
From: "Chuck Higgins" <chu...@nt...>
To: <min...@li...>
Sent: Friday, May 31, 2002 12:49 AM
Subject: [Mingw-users] Not linking to the runtime


>
> Is it possible at the moment to create executables that don't link to any
> runtime library, just the Win API ? Just wondering.
>
> _______________________________________________________________
>
> Don't miss the 2002 Sprint PCS Application Developer's Conference
> August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm
>
> _______________________________________________
> MinGW-users mailing list
> Min...@li...
>
> You may change your MinGW Account Options or unsubscribe at:
> https://lists.sourceforge.net/lists/listinfo/mingw-users
>

Reading Wu's comments i have clearly come to the insight
that the possibility to (satisfyingly) Raise The Fist In
Anger against that Corporate Ignorancy that doesn't fix
This Or That is one brilliant feature commercial
software has that no software will ever have that is
free in both aspects.

In the free world there are only those nice people saying
"submit a patch". And they are so right about it.



-- 


Ulf Schreiber

"Diese Email wurde elektronisch erstellt und
    ist deshalb ohne Unterschrift gültig."

Hi,

gdb (gdb-5.1.1-1-mingw.tar.gz) crashes when I try to set a breakpoint to 
my application (SystemC). The same happens when I use Insight (gdb-5.0) 
from within CYGWIN.

A sample application compiled with gcc-2.95 runs o.k. under both gdb 
versions.

http://gcc.gnu.org/gcc-3.1/changes.html states that at least gdb-5.1.1 
is required for DWARF2 debugging format made by gcc-3.1.

Is there a switch to force gcc to use stabs as debugging format? Or is 
there any other idea or experience to run gdb correctly?

By the way gcc-3.1 allowed me to successfully port SystemC 
(www.systemc.org) to MINGW, which failed with runtime errors due to 
incorrect stack after switching fibers in earlier versions.

Regards

Holger Vogt

On 2002.05.30 19:32 John Barnard wrote:
> I'm struggling to get the readline library version 4.2 usable on a
> Windows
> 2000 system. I initially tried the readline package from
> mingwrep.sourceforge.net, which provides a readline dll, import library,
> headers, and some test programs. When I ran the test program rl I got a
> memory access fault. I got a similar fault when trying to use the
> provided
> dll in any program (compiled with Mingw gcc). So I got the readline-4.2
> source, applied the patch file provided in the mingwrep package, and
> rebuilt the dll and test programs. Same result -- memory access error.
> I fiddled with the dll build command in the provided makefile but no
> success. I get the same problem whether I use gcc 2.95.3-7 or 3.1 (using
> the latest released binutils).

Try to compile the readline library again with debug info "-g" and then 
use Dr.Mingw 
(http://mefriss1.swan.ac.uk/~jfonseca/gnu-win32/software/drmingw/index.html) 
or gdb when you run your program so that we can pin point where the fault 
is, and get some insight of the cause.

> ...

José Fonseca