date:20021125

On Mon, 25 Nov 2002, Andrew DeFaria wrote:

> Christopher Faylor wrote:
> 
> > [one more for the archives]
> > On Mon, Nov 25, 2002 at 04:51:38PM -0800, Andrew DeFaria wrote:
> >
> >> If this were really so then why, if I don't specify -I/usr/include I 
> >> get  getopt.h not found?!? I should be found in /usr/include/getopt.h no?
> >
> > /usr/include is for cygwin apps. If you add -I/usr/include you subvert 
> > the normal operation of -mno-cygwin by including cygwin-only header 
> > files which is *specifically* what -mno-cygwin is avoiding. The system 
> > include area for -mno-cygwin is /usr/include/mingw.
> 
> Understood, however Peter did say:
> 
> If you add the "-v" flag to you compile flags, gcc will dump the list of directories 
>it searches for include files.  In this case it would be:
> 
>  ../../include
>  /usr/include/mingw
>  /usr/include
>  /usr/lib/gcc-lib/i686-pc-cygwin/2.95.3-5/include
>  /usr/include/mingw
>  /usr/include/w32api

No, the example above is from the command line in which *you* specified
"-I../../include -I/usr/include -I/usr/include/mingw".  If you don't
specify any include directories the default list is:

  /usr/lib/gcc-lib/i686-pc-cygwin/2.95.3-5/include
  /usr/include/mingw
  /usr/include/w32api

I'm afraid the real solution is to file a bug report with the mingw list,
just like Christopher said.  I'm betting there are other functions which
aren't prototyped in headers too.  For your purposes, it's sufficient to
copy the function prototype and the external variable declarations from
getopt.h (or make a symlink in your private ../../include directory).
Developing for a minimal environment has it's price :)

> So mine was more of a question: If as Peter says /usr/include is the 3rd 
> library searched, why isn't getopt.h being found? (Another side question 
> would be why is /usr/include/mingw listed twice?).
> 
> >> In any event, after reading up on this problem in the MingW mailing 
> >> list  I find that I am not the only one with this problem. Seems that 
> >> -liberty  does indeed have the code for getopt but there is no mingw 
> >> header file  that defines the interface. My solution ended up being 
> >> define the getopt  API by hand and put it in my own mksf.h header 
> >> file. Ugly but it works.
> >
> > Did you send a bug report or patch to the mingw mailing list?
> 
> Didn't really want to have to sign up for yet another account on another 
> mailing list, yadda, yadda. It seemed to me that they were aware of the 
> problem already so why should I bother to report a problem that is 
> already known?

But, since that *is* the environment you are building in, it makes sense
to subscribe to that list.  It's a very low traffic list anyways, and it
only takes a few minutes to subscribe.  C'mon, it's for a good cause!

-- 
Peter A. Castro <[EMAIL PROTECTED]> or <[EMAIL PROTECTED]>
"Cats are just autistic Dogs" -- Dr. Tony Attwood


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [Patch] skipping import libraries for performance reasons - direct auto-import of dll's

On Mon, Nov 25, 2002 at 11:43:42PM -0500, Charles Wilson wrote:
>Christopher Faylor wrote:
>
>
>>
>>Another thing that "would be nice" would be to speed up the handling of
>>import libraries.  It might not be necessary for ld to be as slow as
>>it is.
>
>
>What would probably speed that up dramatically is to construct the 
>importlib with more than a single symbol per bfd.  But I'm not sure 
>exactly how that should be done: all-in-one-massive-bfd (which might 
>cause problems with very large libs), or somewhere in between all-in-one 
>and one-per (which is much more complicated)?  If between, where? 
>5-per?  30-per? 100?

Yeah, this is what I was thinking.  I thought there was some code added
to ld recently which allowed selective linking of functions from a
single object file.  I remember this being a desirable feature of
Microsoft's linker.  If we don't have this, we stand the chance of
increasing the size of an executable for now good reason as it links
in the import library jmp *foo magic for unreferenced functions.

cgf

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [Patch] skipping import libraries for performance reasons -direct auto-import of dll's

On Tue, 2002-11-26 at 15:43, Charles Wilson wrote:

  If between, where? 
> 5-per?  30-per? 100?

When in doubt, set a backwards compatible default (1), and add a knob
:}.

Rob
-- 
---
GPG key available at: http://users.bigpond.net.au/robertc/keys.txt.
---



signature.asc
Description: This is a digitally signed message part

Re: [Patch] skipping import libraries for performance reasons - direct auto-import of dll's

Christopher Faylor wrote:




Another thing that "would be nice" would be to speed up the handling of
import libraries.  It might not be necessary for ld to be as slow as
it is.



What would probably speed that up dramatically is to construct the 
importlib with more than a single symbol per bfd.  But I'm not sure 
exactly how that should be done: all-in-one-massive-bfd (which might 
cause problems with very large libs), or somewhere in between all-in-one 
and one-per (which is much more complicated)?  If between, where? 
5-per?  30-per? 100?

--Chuck





--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [Patch] skipping import libraries for performance reasons - direct auto-import of dll's

On Mon, Nov 25, 2002 at 10:20:13PM -0500, Charles Wilson wrote:
>Christopher Faylor wrote:
>>On Mon, Nov 25, 2002 at 01:46:50PM +0100, Ralf Habacker wrote:
>>>3.  ld works more like the linux version.  There are only static
>>>archives and shared libraries which could be linked directly without
>>>the indirection of using import libraries.  This simplifies for example
>>>libtool handling.
>>>
>>
>>I don't see how.  If anything it would complicate libtool handling
>>since libtool would have to know about both import libraries and dlls.
>>You can't just give up on import libraries, if for no other reason than
>>some libraries (like cygwin's for instance) contain a combination of
>>import data and static data.
>
>
>Yep.  I think this is better handled as follows:

Another thing that "would be nice" would be to speed up the handling of
import libraries.  It might not be necessary for ld to be as slow as
it is.

cgf

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [Patch] skipping import libraries for performance reasons - direct auto-import of dll's

Christopher Faylor wrote:


On Mon, Nov 25, 2002 at 01:46:50PM +0100, Ralf Habacker wrote:


3. ld works more like the linux version. There are only static archives and
shared libraries which could be linked directly without the indirection of using
import libraries. This simplifies for example libtool handling.



I don't see how.  If anything it would complicate libtool handling since libtool
would have to know about both import libraries and dlls.  You can't just give
up on import libraries, if for no other reason than some libraries (like
cygwin's for instance) contain a combination of import data and static data.



Yep.  I think this is better handled as follows:

1) leave the basic libtool behavior alone; let it go ahead and build 
import libs (and try to link to them).

2) perhaps a new libtool switch could be created, that you would put in 
to Makefile.am:  libfoo_la_LDFLAGS = -no-undefined -fast-import

when -fast-import is specified, libtool (for cygwin target) "installs" 
the "import lib" into $libdir as a symlink:
  $libdir/libfoo.dll -> ../bin/cygfoo-X.dll
And $libdir/libfoo.la includes stuff like

dlname='cygcxxdll-0.dll'
library_names='libcxxdll.dll'
old_library='libcxxdll.a'

(the only difference from "current" (e.g. CVS HEAD) behavior is that 
right now, library_names='libcxxdll.dll.a')

I'm not sure how what the uninstalled file layout should be, or how the 
uninstalled .la file should refer to the dll/"implib" prior to 
installation (e.g. for linking further libs/exes in the same buildtree). 
 Probably even the link command should be different (don't specify 
-Wl,--out-implib=... at all; create the "implib" as a simple 'cp 
$dllfilename $implibname')

But all of this is way premature.  First, get the feature tested and 
merged into CVS.  Then into the [curr] binutils for cygwin.  Somewhere 
along the way, the runtime pseudo-reloc code goes into cygwin1.dll CVS 
or crt0.o CVS or wherever the heck it ends up.  And then wait for THAT 
to make it into the appropriate [curr] cygwin package.

And then we can worry about mucking with libtool.  It'll probably be 
post-libtool-1.5 at that time, but I doubt there will be another huge 
delay between libtool-1.5 and -1.5.1/2/3/..., as there was/is between 
libtool-1.3.x/1.4.x and 1.5release.  So don't worry about missing the 
1.5 window.


However, as Chuck mentions that doesn't detract from the merits of the patch.
I'm sure that it would still be very useful to a number of people.



pending testing, of course...

--Chuck



--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: nice really nice?

2002-11-25 Thread Randall R Schulz

Robert,

Well, I guess it's a good thing I sent that to the list (given that I 
stated inaccurate information), but I thought I was replying to Thomas 
privately. (He used that "thomas <[EMAIL PROTECTED]>" address even though 
the message to which I replied was sent to me only--I just hit reply 
without looking at the return address, since we've been working on the 
problem with "mkisofs" piped to "cdrecord.")

Anyway, thanks for clearing up the Windows priority misinformation I sent 
out. I guess if I would have read the MSDN tech not Thomas referred me to 
first, I wouldn't have said that...

Randall Schulz
Mountain View, CA USA

At 18:06 2002-11-25, Robert Collins wrote:
On Tue, 2002-11-26 at 14:00, Randall R Schulz wrote:
> Thomas,
>
> One thing to keep in mind is that while Unix (and work-alikes) has a -20
> (best scheduling priority) ... +20 (worst priority) range, Windows has 
only
> the six distinct levels. I don't know how Cygwin maps the Unix nice values
> to the Windows priorities, offhand. Probably it's a linear mapping.
>
> I haven't had a chance to read the information about scheduling in 
Windows,
> but I will. Thanks for referring me to it.

Windows has (offhand) ~ 30 scheduling levels. It has priority classes, 
which 'group' processes, and then relative priorities within each 
class.IIRC you can check sched,cc via CVS to see the actual mapping I 
used, it's not linear as such, but nearly so.

Thomas,

Those tests show nothing other than the time it takes to push the iso 
through to a bitbucket. Unless there is serious other load on the CPU, the 
time *should* be constant.

Rob

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: nice really nice?

On Tue, 2002-11-26 at 14:00, Randall R Schulz wrote:
> Thomas,
> 
> One thing to keep in mind is that while Unix (and work-alikes) has a -20 
> (best scheduling priority) ... +20 (worst priority) range, Windows has only 
> the six distinct levels. I don't know how Cygwin maps the Unix nice values 
> to the Windows priorities, offhand. Probably it's a linear mapping.
> 
> I haven't had a chance to read the information about scheduling in Windows, 
> but I will. Thanks for referring me to it.

Windows has (offhand) ~ 30 scheduling levels. It has priority classes,
which 'group' processes, and then relative priorities within each
class.IIRC you can check sched,cc via CVS to see the actual mapping I
used, it's not linear as such, but nearly so.

Thomas,
 those tests show nothing other than the time it takes to push the iso
through to a bitbucket. Unless there is serious other load on the CPU,
the time *should* be constant.
Rob
-- 
---
GPG key available at: http://users.bigpond.net.au/robertc/keys.txt.
---



signature.asc
Description: This is a digitally signed message part

Re: nice really nice?

2002-11-25 Thread Randall R Schulz

Thomas,

One thing to keep in mind is that while Unix (and work-alikes) has a -20 
(best scheduling priority) ... +20 (worst priority) range, Windows has only 
the six distinct levels. I don't know how Cygwin maps the Unix nice values 
to the Windows priorities, offhand. Probably it's a linear mapping.

I haven't had a chance to read the information about scheduling in Windows, 
but I will. Thanks for referring me to it.

Randy


At 10:14 2002-11-25, you wrote:

I'm also wondering what nice really does in cygwin. Look at the
following test:

$ time mkisofs -J -R -l * 2>/dev/null | nice -19 dd of=/dev/null
real0m7.482s

$ time mkisofs -J -R -l * 2>/dev/null | nice -1 dd of=/dev/null
real0m7.384s

$ time mkisofs -J -R -l * 2>/dev/null | nice -0 dd of=/dev/null
real0m7.419s

$ time mkisofs -J -R -l * 2>/dev/null | nice --1 dd of=/dev/null
real1m51.516s

$ time mkisofs -J -R -l * 2>/dev/null | nice --19 dd of=/dev/null
real1m51.760s



--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: impure_ptr/Mingw and cygwin

On Mon, Nov 25, 2002 at 05:42:43PM -0800, Andrew DeFaria wrote:
>Christopher Faylor wrote:
>
>>[one more for the archives]
>>On Mon, Nov 25, 2002 at 04:51:38PM -0800, Andrew DeFaria wrote:
>>
>>>If this were really so then why, if I don't specify -I/usr/include I 
>>>get  getopt.h not found?!? I should be found in /usr/include/getopt.h no?
>>
>>/usr/include is for cygwin apps. If you add -I/usr/include you subvert 
>>the normal operation of -mno-cygwin by including cygwin-only header 
>>files which is *specifically* what -mno-cygwin is avoiding. The system 
>>include area for -mno-cygwin is /usr/include/mingw.
>
>Understood, however Peter did say:
 ^^
Apparently not.

>If you add the "-v" flag to you compile flags, gcc will dump the list of 
>directories it searches for include files.  In this case it would be:
>
>../../include
>/usr/include/mingw
>/usr/include
>/usr/lib/gcc-lib/i686-pc-cygwin/2.95.3-5/include
>/usr/include/mingw
>/usr/include/w32api

Because the compiler options specified included -I/usr/include.  He was
responding to your question about "why am i getting this warning".  Your
command line looked like this:

> /bin/gcc -c -g -mno-cygwin -I. -I../../include -I/usr/include/mingw -I/usr/include 
>-DCYGWIN mksf.c
>

>So mine was more of a question: If as Peter says /usr/include is the 3rd 
>library searched, why isn't getopt.h being found?

In the example that you specified there was no hint that it wasn't
found.  In fact, you said "this works" (sic).  You were asking
about a warning.

>(Another side question would be why is /usr/include/mingw listed
>twice?).

You specified -I/usr/include/mingw on the command line and it is already
included by default when you use -mno-cygwin.

Let me say it again.  If you want to compile a windows-only app, you
just use -mno-cygwin.  If you find that some function is missing when
you do this, that means that the function is likely missing from the
mingw distribution.  The solution in this case is not to link in the
bits that you need from cygwin directories.  -mno-cygwin is supposed to
be a world by itself.  You usually can't mix it with cygwin bits and get
anything useful.

cgf

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: impure_ptr/Mingw and cygwin

2002-11-25 Thread Andrew DeFaria

Christopher Faylor wrote:

[one more for the archives]
On Mon, Nov 25, 2002 at 04:51:38PM -0800, Andrew DeFaria wrote:

If this were really so then why, if I don't specify -I/usr/include I
get getopt.h not found?!? I should be found in /usr/include/getopt.h no?

/usr/include is for cygwin apps. If you add -I/usr/include you subvert
the normal operation of -mno-cygwin by including cygwin-only header
files which is *specifically* what -mno-cygwin is avoiding. The system
include area for -mno-cygwin is /usr/include/mingw.

Understood, however Peter did say:

If you add the "-v" flag to you compile flags, gcc will dump the list of directories it searches for include files. In this case it would be:

../../include
/usr/include/mingw
/usr/include
/usr/lib/gcc-lib/i686-pc-cygwin/2.95.3-5/include
/usr/include/mingw
/usr/include/w32api

So mine was more of a question: If as Peter says /usr/include is the 3rd
library searched, why isn't getopt.h being found? (Another side question
would be why is /usr/include/mingw listed twice?).

In any event, after reading up on this problem in the MingW mailing
list I find that I am not the only one with this problem. Seems that
-liberty does indeed have the code for getopt but there is no mingw
header file that defines the interface. My solution ended up being
define the getopt API by hand and put it in my own mksf.h header
file. Ugly but it works.

Did you send a bug report or patch to the mingw mailing list?

Didn't really want to have to sign up for yet another account on another
mailing list, yadda, yadda. It seemed to me that they were aware of the
problem already so why should I bother to report a problem that is
already known?

--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/

Re: [Patch] skipping import libraries for performance reasons - direct auto-import of dll's

On Mon, Nov 25, 2002 at 11:50:13PM +0100, Ralf Habacker wrote:
>Chuck Wilson wrote:
>>Minor nit about patch format: watch out for your tab/indentation.  It
>>doesn't match the surrounding text in many cases.
>
>Okay.  How are the tab width ?

It's the standard 8 characters per tab.  You also aren't adhering to GNU
coding standards and a ChangeLog would be nice.

cgf

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [Patch] skipping import libraries for performance reasons - direct auto-import of dll's

On Mon, Nov 25, 2002 at 01:46:50PM +0100, Ralf Habacker wrote:
>3. ld works more like the linux version. There are only static archives and
>shared libraries which could be linked directly without the indirection of using
>import libraries. This simplifies for example libtool handling.

I don't see how.  If anything it would complicate libtool handling since libtool
would have to know about both import libraries and dlls.  You can't just give
up on import libraries, if for no other reason than some libraries (like
cygwin's for instance) contain a combination of import data and static data.

However, as Chuck mentions that doesn't detract from the merits of the patch.
I'm sure that it would still be very useful to a number of people.

cgf

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: impure_ptr/Mingw and cygwin

[one more for the archives]
On Mon, Nov 25, 2002 at 04:51:38PM -0800, Andrew DeFaria wrote:
>If this were really so then why, if I don't specify -I/usr/include I get 
>getopt.h not found?!? I should be found in /usr/include/getopt.h no?

/usr/include is for cygwin apps.  If you add -I/usr/include you subvert
the normal operation of -mno-cygwin by including cygwin-only header
files which is *specifically* what -mno-cygwin is avoiding.  The system
include area for -mno-cygwin is /usr/include/mingw.

The same applies for library files.  The system area for libraries when
-mno-cygwin is active is /usr/lib/mingw.

(This is a simplification because it isn't exactly like this for gcc 3.2
but it is close enough)

In any event adding a library path to /usr/lib or an include path search
to /usr/include is *asking for trouble*.

>In any event, after reading up on this problem in the MingW mailing list 
>I find that I am not the only one with this problem. Seems that -liberty 
>does indeed have the code for getopt but there is no mingw header file 
>that defines the interface. My solution ended up being define the getopt 
>API by hand and put it in my own mksf.h header file. Ugly but it works.

Did you send a bug report or patch to the mingw mailing list?

cgf

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: impure_ptr/Mingw and Cygwin

2002-11-25 Thread Andrew DeFaria

Peter A. Castro wrote:


On Mon, 25 Nov 2002, Andrew DeFaria wrote:


Peter A. Castro wrote:
   

On Sun, 24 Nov 2002, Andrew DeFaria wrote:
 

Peter A. Castro wrote:

   

What you show below is only linking. I believe you need to 
re-compile all of your source with -mno-cygwin -mwindows as well to 
make the _impure_ptr references go away.
 

But I did re-compile all my sources with -mno-cygwin -mwindows! Still 
have the error.
   

Suggest you run nm against all your objects and libraries and find which
has the _impure_ptr references, then recompile just those.
 

But I did recompile everything already! The _impure_ptr keeps getting 
generated:

$ nm mksf.o | grep impure
U __impure_ptr
$ rm mksf.o
$ make mksf.o
/bin/gcc -c -g -mno-cygwin -mwindows -I. -I../../include -I/usr/include 
-DCYGWIN mksf.c
$ nm mksf.o | grep impure
U __impure_ptr
$
   


You're getting stdio.h from /usr/include, that's why.

 

As you can see the _impure_ptr reference gets regenerated. But wait! 
There's more!

I can cause these _impure_ptr even with the little foo.c merely by 
including -I/usr/include on the command line. Why am I referencing 
/usr/include? Because in my larger application I need it for getopt.h. 
The mere inclusion of -I/usr/include causes_impure_ptr's to be emitted. 
So the question now is: How do I satisfy my need for getopt and still 
produce objects without _impure_ptr's?
   


libiberty.a in /usr/lib/mingw has getopt(), so link with -liberty

 

Ah ha!

$ grep impure /usr/include/*.h
/usr/include/stdio.h:#definestdin   (_impure_ptr->_stdin)
/usr/include/stdio.h:#definestdout  (_impure_ptr->_stdout)
/usr/include/stdio.h:#definestderr  (_impure_ptr->_stderr)

Hmmm... Changed options to:

/bin/gcc -c -g -mno-cygwin -I. -I../../include -I/usr/include/mingw 
-I/usr/include -DCYGWIN mksf.c
cc1: warning: changing search order for system directory 
"/usr/include/mingw"
cc1: warning:   as it has already been specified as a non-system directory

This works but why am I getting this warning?
   


Because -mno-cygwin specifies /usr/include/mingw as a system dir and
you're specifying it again on the command line with -I, which makes it a
User dir too.  gcc is a bit puzzled about why you're doing that. 

If you add the "-v" flag to you compile flags, gcc will dump the list of
directories it searches for include files.  In this case it would be:

../../include
/usr/include/mingw
/usr/include
/usr/lib/gcc-lib/i686-pc-cygwin/2.95.3-5/include
/usr/include/mingw
/usr/include/w32api

 

If this were really so then why, if I don't specify -I/usr/include I get 
getopt.h not found?!? I should be found in /usr/include/getopt.h no?

In any event, after reading up on this problem in the MingW mailing list 
I find that I am not the only one with this problem. Seems that -liberty 
does indeed have the code for getopt but there is no mingw header file 
that defines the interface. My solution ended up being define the getopt 
API by hand and put it in my own mksf.h header file. Ugly but it works.

Thanks for working this out with me.
--

Salira 
Ethernet Simple, Fiber Fast

5451 Patrick Henry Drive
Santa Clara, CA 95054
Phone: (408)-845-5321
Fax: (408)-845-5205
Email: [EMAIL PROTECTED] 

Web: http://www.salira.com

Instant Messaging
AIM:
defaria
MSN:
[EMAIL PROTECTED]
Yahoo:
andrew_defaria
ICQ #:
23552673


Andrew DeFaria 
Clearcase Administrator
Email: [EMAIL PROTECTED] 
Web: http://DeFaria.com







--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: impure_ptr/Mingw and Cygwin

On Mon, 25 Nov 2002, Andrew DeFaria wrote:

> Peter A. Castro wrote:
> > On Sun, 24 Nov 2002, Andrew DeFaria wrote:
> >> Peter A. Castro wrote:
> >>
> >>> What you show below is only linking. I believe you need to 
> >>> re-compile all of your source with -mno-cygwin -mwindows as well to 
> >>> make the _impure_ptr references go away.
> >>
> >> But I did re-compile all my sources with -mno-cygwin -mwindows! Still 
> >>  have the error.
> >
> > Suggest you run nm against all your objects and libraries and find which
> > has the _impure_ptr references, then recompile just those.
> 
> But I did recompile everything already! The _impure_ptr keeps getting 
> generated:
>
> $ nm mksf.o | grep impure
>  U __impure_ptr
> $ rm mksf.o
> $ make mksf.o
> /bin/gcc -c -g -mno-cygwin -mwindows -I. -I../../include -I/usr/include 
> -DCYGWIN mksf.c
> $ nm mksf.o | grep impure
>  U __impure_ptr
> $

You're getting stdio.h from /usr/include, that's why.

> As you can see the _impure_ptr reference gets regenerated. But wait! 
> There's more!
> 
> I can cause these _impure_ptr even with the little foo.c merely by 
> including -I/usr/include on the command line. Why am I referencing 
> /usr/include? Because in my larger application I need it for getopt.h. 
> The mere inclusion of -I/usr/include causes_impure_ptr's to be emitted. 
> So the question now is: How do I satisfy my need for getopt and still 
> produce objects without _impure_ptr's?

libiberty.a in /usr/lib/mingw has getopt(), so link with -liberty

> Ah ha!
> 
> $ grep impure /usr/include/*.h
> /usr/include/stdio.h:#definestdin   (_impure_ptr->_stdin)
> /usr/include/stdio.h:#definestdout  (_impure_ptr->_stdout)
> /usr/include/stdio.h:#definestderr  (_impure_ptr->_stderr)
> 
> Hmmm... Changed options to:
> 
> /bin/gcc -c -g -mno-cygwin -I. -I../../include -I/usr/include/mingw 
> -I/usr/include -DCYGWIN mksf.c
> cc1: warning: changing search order for system directory 
> "/usr/include/mingw"
> cc1: warning:   as it has already been specified as a non-system directory
> 
> This works but why am I getting this warning?

Because -mno-cygwin specifies /usr/include/mingw as a system dir and
you're specifying it again on the command line with -I, which makes it a
User dir too.  gcc is a bit puzzled about why you're doing that. 

If you add the "-v" flag to you compile flags, gcc will dump the list of
directories it searches for include files.  In this case it would be:

 ../../include
 /usr/include/mingw
 /usr/include
 /usr/lib/gcc-lib/i686-pc-cygwin/2.95.3-5/include
 /usr/include/mingw
 /usr/include/w32api
 
-- 
Peter A. Castro <[EMAIL PROTECTED]> or <[EMAIL PROTECTED]>
"Cats are just autistic Dogs" -- Dr. Tony Attwood


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: The Black Art of DLL Creation (revisited)

2002-11-25 Thread Soren A

A brief update. Partial replies to Gerritt. Self-corrections.

Soren A <[EMAIL PROTECTED]> wrote
in [EMAIL PROTECTED]:">news:[EMAIL PROTECTED]: 
> The scenario is that I am trying to re-create the process of building 
> bleadperl [...] but one thing that
> you can get from using dlltool, that you don't (I think?) get from
> using 'gcc -shared' is a .DEF file that is basically correct. If you
> ask 'gcc -shared' for a .DEF output using the appropriate switches,
> you get included stuff like a .bss[data?] that isn't supposed to be
> exported. 

OK, I see that this had been fixed in the interim. No longer any need at
all to use dlltool explicitly unless we need a .exp file (and why would
we? I do not in fact know). 


> I would like (for reasons that may not seem important to others) to
> have a valid (by which I mean 'containing all symbols [functions and
> global data] that need to be exported [visible to external exes] by
> the DLL, and none that do not') .DEF file to work with. (OK, one
> reason is that I am interested in the internals of Perl, and examining
> the contents of the .DEF file gives a glimpse into those internals). 


>   2. Only a handful of symbols were listed under EXPORTS. These
>   symbols were all prefixed "XS_[foo]" and it is possible, based on my
>  examination of the source, that these functions were declared in
>  some way that doesn't involve "__declspec(dllexport)". The
>  dozens of other symbols that one expects and needs to see weren't
>  exported. 

  [...]
> I made sure that the Perl macros "EXT" and "cEXT" (constant) etc. were
> being properly substituted with "__declspec(dllexport)". The header in
> the cygwin/ subdir of the Perl src tree causes it NOT to be set, --
> Gerrit and others who are knowledgeable about CygwinPerl may note this
> -- but I found a way to override that. 

Actually I didn't. I hoped that I had and believed that I had, but it
was NOT working. I had to learn more about cpp through trial-and-error
experimentation and using some tricks, in order to detect this and
control it. 

> Backing up: in theory TTBOMU, if I've thus marked a symbol for export
> this way, the Cygwin port of GNU ld(1) *should* export it
> unconditionally. Furthermore, TTBOMU, this SHOULD only be necessary
> for global variables (data) in the first place -- all *functions*
> should be exported anyway(?). In cases where the src package has no
> provisions for this, we've used the "--export-all" switch (for either
> dlltool OR 'gcc -shared') to cause eveything but the special Cygwin
> excluded symbols relating to entry points, etc. to be marked for
> export in the final DLL. This is what the CygwinPerl build setup does
> at present, and I don't understand why. The mechanism for marking
> symbols with "__declspec(dll[ex|in]port" is already part of Perl src.
> Why can't we use it? Clearly *somebody* does or did -- probably MSVC
> does. 

I forced the symbols to be marked for export.

So, anyway, I was wrong about this behavior on the part of dlltool and ld.

I have built the DLL and linked in the Perl executable with some
manipulation of which symbols get exported from the DLL, by using `nm'
and friends to examine the exports closely. The current setup I have is
that I can either choose to generate a .DEF file, or not, and link using
the file, or not. There is some difference between what gets exported
between those two cases. I'd be happy to discuss that if anyone who is
knowledgeable, is interested. 

I can also control (with approaches that i haven't written about in this 
thread) the names of the generated libraries (all flavors: DLL [which is 
really an executable instea dof a classic library], interface lib, and 
static archive lib. I'v got a massively strange and interesting (to me 
alone, probably) experiment going here.

  Best,
   Soren A


-- 
What do Internet Mailing Lists and crowded neighborhoods have in
common? Both will either drive you out or teach you how to ignore
barking dogs.



--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

Robert Collins <[EMAIL PROTECTED]> wrote:

> On Tue, 2002-11-26 at 08:56, Max Bowsher wrote:
>> Robert Collins <[EMAIL PROTECTED]> wrote:
> 
> 
>>> If you don't have McAfee installed, then I'll roll an Event Log
>>> based setup for you.
>> 
>> Why not just flush the logs to disc after each line? I've got a
>> patch lying around here somewhere to do that. I'll update it to the
>> latest version. 
> 
> As long as:
> A) It's on NTFS
> B) The must-not-cache flag is set in CreateFile.
> 
> Otherwise, the last entries will be lost when the crash occurs.

Oh. I think I was just closing the file all the time.

Max.


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Changing user ID.

2002-11-25 Thread Igor Pechtchanski

On Mon, 25 Nov 2002, Kurt Roeckx wrote:

> I've been trying to log in as an other user under NT.
>
> From the archives I seem to get that su was removed, and that you
> should be using login for it instead, so I've tried to use that.
>
> It comes asking for the password, but always says it failed.
> Anybody got any idea why it's not working?
>
> The user doesn't exist local, but does exist in the domain.
>
> I don't want to try to use sshd/ssh, since I really don't need
> them, but would that work instead?
>
> Kurt

Switching user context on NT-based systems requires special privileges,
which normally belong only to the LocalSystem user.  Unless your user has
those privileges, login and others will fail.  sshd/ssh will work, since
sshd runs as LocalSystem, and thus has enough privileges to switch the
user.
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

"Water molecules expand as they grow warmer" (C) Popular Science, Oct'02, p.51


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

P.S. can't set $USER under 1.3.16-1

2002-11-25 Thread Lester Ingber

I see some similar problem has been reported.  To be clear, I have
in my /etc/profile
--8<-- top cut 1/2 -> bottom 1/2 --->8--
PATH=/usr/X11R6/bin:/usr/local/bin:/usr/bin:/bin:$PATH

export CYGWIN="check_case:strict nontsec"

USER="ingber"

# Set up USER's home directory
if [ -z "$HOME" ]; then
  HOME="/home/$USER"
fi

if [ ! -d "$HOME" ]; then
  mkdir -p "$HOME"
fi

export HOME USER

for i in /etc/profile.d/*.sh ; do
  if [ -f $i ]; then
. $i
  fi
done

export MAKE_MODE=unix
export PS1='\[\033]0;\w\007
\033[32m\]\u@\h \[\033[33m\w\033[0m\]
$ '

cd "$HOME"

test -f ./.bashrc && . ./.bashrc
--8<-- bottom cut 1/2 <- top 1/2 --->8--

The fix suggested for a similar problem was to add HOME to /cygwin.bat.
I tried in my /cygwin.bat
--8<-- top cut 2/2 -> bottom 2/2 --->8--
set USER=ingber
set HOME=/home/ingber
--8<-- bottom cut 2/2 <- top 2/2 --->8--
My login is smoother, but I still get logged on as the user in `id -un`,
not "ingber" .

Lester


: From [EMAIL PROTECTED] Mon Nov 25 16:17:58 2002
: Date: Mon, 25 Nov 2002 16:17:58 -0500
: From: Lester Ingber <[EMAIL PROTECTED]>
: To: [EMAIL PROTECTED]
: Subject: can't set $USER under 1.3.16-1
: 
: Up until now, through cygwin 1.3.15-2, I have set my $USER in /etc/profile
: to a username different from `id -un`, with the same name set in /etc/passwd.
: I have had no problems logging in.
: 
: Now, under 1.3.16-1, an attempt is made to login in under `id -un` and
: I cannot even telent to my perferred username?  I reinstalled 1.3.15-2,
: and everything works as before installing  1.3.16-1?
: 
: What else must I do under 1.3.16-1 to set my own username?
: 
: Lester

-- 
 Prof. Lester Ingber  [EMAIL PROTECTED]  [EMAIL PROTECTED]
 www.ingber.com www.alumni.caltech.edu/~ingber

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Enquiry

2002-11-25 Thread a peters


Can you provide details re, or a link to a how-to on the subject of, 
configuring cygwin inetd for use on Windows 2000? 

I'm trying to set up an rlogind service to permit access via rlogin from a 
host running VMS 7.1-2 (not that this is relevant) to resources on my Win 
2000 server.

Thus far am getting the message

" The service name is invalid "

When I attempt to net start inetd. 

/etc/inetd.conf includes entries for telnet and rlogind.

Cheers,

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

On Tue, 2002-11-26 at 08:56, Max Bowsher wrote:
> Robert Collins <[EMAIL PROTECTED]> wrote:


> > If you don't have McAfee installed, then I'll roll an Event Log based
> > setup for you.
> 
> Why not just flush the logs to disc after each line? I've got a patch lying
> around here somewhere to do that. I'll update it to the latest version.

As long as:
A) It's on NTFS
B) The must-not-cache flag is set in CreateFile.

Otherwise, the last entries will be lost when the crash occurs.

Rob

> 
-- 
---
GPG key available at: http://users.bigpond.net.au/robertc/keys.txt.
---



signature.asc
Description: This is a digitally signed message part

Re: impure_ptr/Mingw and cygwin

On Mon, Nov 25, 2002 at 01:16:04PM -0800, Andrew DeFaria wrote:
>So the question now is: How do I satisfy my need for getopt and still 
>produce objects without _impure_ptr's?

You find some native windows getopt, of course.

>Ah ha!

Yes.  Don't include the cygwin headers when you're compiling with
-mno-cygwin.

>$ grep impure /usr/include/*.h
>/usr/include/stdio.h:#definestdin   (_impure_ptr->_stdin)
>/usr/include/stdio.h:#definestdout  (_impure_ptr->_stdout)
>/usr/include/stdio.h:#definestderr  (_impure_ptr->_stderr)
>
>Hmmm... Changed options to:
>
>/bin/gcc -c -g -mno-cygwin -I. -I../../include -I/usr/include/mingw 
>-I/usr/include -DCYGWIN mksf.c
>cc1: warning: changing search order for system directory 
>"/usr/include/mingw"
>cc1: warning:   as it has already been specified as a non-system directory
>
>This works but why am I getting this warning?

Because you're using /usr/include/mingw and it is already included on
the command line by default.  That's what -mno-cygwin does.

cgf
--
Please do not send me personal email with cygwin questions or observations.
Use the resources at http://cygwin.com/ .

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

Robert Collins <[EMAIL PROTECTED]> wrote:

> On Tue, 2002-11-26 at 03:53, Daniel Armbrust wrote:
>> Where can I get this?
>>
>>>
>>>
>>> If the NT 4 machine crashes every time, would you be willing to run
>>> a special debug version of setup, that will write it's log entries
>>> to the event log? They should be visible post install then, and you
>>> can tell us what the last ones recorded are.
>
>
> Well, I was going to roll it for you. However, having been reminded of
> the McAfee problem - do you have McAfee installed? - I'd rather have
> you try the current setup snapshot, which will give you the option for
> disabling McAfee during the download and install process.
>
> If you don't have McAfee installed, then I'll roll an Event Log based
> setup for you.

Why not just flush the logs to disc after each line? I've got a patch lying
around here somewhere to do that. I'll update it to the latest version.

Max.


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Changing user ID.

2002-11-25 Thread Kurt Roeckx

I've been trying to log in as an other user under NT.

>From the archives I seem to get that su was removed, and that you
should be using login for it instead, so I've tried to use that.

It comes asking for the password, but always says it failed.
Anybody got any idea why it's not working?

The user doesn't exist local, but does exist in the domain.

I don't want to try to use sshd/ssh, since I really don't need
them, but would that work instead?


Kurt


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

On Tue, 2002-11-26 at 02:09, Daniel Armbrust wrote:
> One further clarification.  It happens well before the
> 
>   Installation complete
>   [OK]
> 
> step.  It actually happens before any of the post download scripts run. 
>   It happens when it is at (or nearly at) the very end of downloading 
> the packages.  After the blue screen, when I come back in, and just 
> click through the install steps again, it usually redownloads or 
> finishes one package, and then moves on to the install step (which works 
> fine).
> 
> 
> Do the MS debugging tools ship with NT 4? 

Some basic ones do.

 Or do I have to pick them up 
> somewhere on the MS site?

This is the best approach. Uhm, I don't have the download handy,
something like www.microsoft.com/ddk/debugging.


Rob
-- 
---
GPG key available at: http://users.bigpond.net.au/robertc/keys.txt.
---



signature.asc
Description: This is a digitally signed message part

Re: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

On Tue, 2002-11-26 at 03:53, Daniel Armbrust wrote:
> Where can I get this?
> 
> >
> >
> > If the NT 4 machine crashes every time, would you be willing to run a
> > special debug version of setup, that will write it's log entries to the
> > event log? They should be visible post install then, and you can tell us
> > what the last ones recorded are.

Well, I was going to roll it for you. However, having been reminded of
the McAfee problem - do you have McAfee installed? - I'd rather have you
try the current setup snapshot, which will give you the option for
disabling McAfee during the download and install process.

If you don't have McAfee installed, then I'll roll an Event Log based
setup for you.

Rob

-- 
---
GPG key available at: http://users.bigpond.net.au/robertc/keys.txt.
---

signature.asc
Description: This is a digitally signed message part

Re: Installing w/o setup.ini

[EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:

> I am attempting to install Cygwin from a CD-ROM containing downloaded
> Cygwin files (src and bin) but which does not have a
> setup.ini/setup.hint. I do not have a Cygwin/Unix enabled computer or
> access to the net and can't use, e.g., Polley Christopher W. (8 Nov
> 2002) script to generate a .ini file. Is there any way to do an
> installation? 
> 
> Can I retrieve .ini/.hint files from ftp sites? How?

First: setup.hint files are for packagers only *not* for end users.

Why not just run setup.exe, and let it download a setup.ini for you?

Max.


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [Performance Regression] cygwin 1.3.15-2 -> 1.3.16-1

Christopher Faylor <[EMAIL PROTECTED]> wrote:

> On Mon, Nov 25, 2002 at 05:56:12PM -, Max Bowsher wrote:
>> [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>>
>>> On Sun, Nov 24, 2002 at 11:59:00PM -, Max Bowsher wrote:
 Max Bowsher <[EMAIL PROTECTED]> wrote:

> With 16-1, I'm seeing *extreme* latency in keypresses appearing in
> bash, when I have a compile running in another Cygwin window (in
> my case, compiling setup). With 15-2, there are no such delays. I
> will try to localize the problem to between 2 snapshots.

 Localized to between 1.3.15-2 and 20021115, which is the earliest
 snapshot still available.
>>>
>>> I'll bet it works ok on B20, too.
>>
>> Putting flippancy aside, I think this is a real problem.
>
> Hmm.  A real problem would have real details, wouldn't it?  Beyond
> the fact that you think it is slow, there really weren't too many of
> those.

Fair point. I should have emphasised that I flipped between several dll
versions, and the slowness was not just a subconcious feel, it was very
obvious and easily detectable.

>> If necessary, I will start building some DLLs from cvs so I can roll
>> forward commit by commit.  But I was hoping someone might go "Aha - I
>> know the change that might have caused that."
>
> Yeah, that's always the magical thinking that goes into cygwin bug
> reports.

OK, time to start building myself some dlls, then.


Max.


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Jonathan FREESON/dassault-systemes is out of the office.

2002-11-25 Thread Jonathan FREESON

I will be out of the office starting  11/25/2002 and will not return until
11/28/2002.

I will respond to your message when I return.


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

[ANNOUNCEMENT] Updated: w32api-2.1-1

2002-11-25 Thread Earnie Boyd

This is a multi-part message in MIME format.
--060809000909020701010503
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

I've made a new version of the w32api headers and libraries available
for download.  A list of what has changed is attached.

Earnie.

-Installation Instructions-
To update your installation, click on the "Install Cygwin now" link on
the http://cygwin.com/ web page.  This downloads setup.exe to your
system.  Then, run setup and answer all of the questions.

To update your installation, click on the "Install Cygwin now" link on
the http://cygwin.com/ web page.  This downloads setup.exe to your
system.  Then, run setup and answer all of the questions.

Note that we do not allow downloads from sources.redhat.com (aka
cygwin.com) due to bandwidth limitations.  This means that you will need

to find a mirror which has this update.

In the US,
ftp://mirrors.rcn.net/mirrors/sources.redhat.com/cygwin/
is a reliable high bandwidth connection.

In Germany,
ftp://ftp.uni-erlangen.de/pub/pc/gnuwin32/cygwin/mirrors/cygnus/ is
usually pretty good.

In the UK,
http://programming.ccp14.ac.uk/ftp-mirror/programming/cygwin/pub/cygwin/

is usually up-to-date within 48 hours.

If one of the above doesn't have the latest version of this package then

you can either wait for the site to be updated or find another mirror.

If you have questions or comments, please send them to the Cygwin
mailing list at: [EMAIL PROTECTED] .  I would appreciate it if you would

use this mailing list rather than emailing me directly.  This includes
ideas and comments about the setup utility or Cygwin in general.
Really.
No kidding.  Email cygwin stuff to [EMAIL PROTECTED]

If you want to make a point or ask a question, the Cygwin mailing list
is the appropriate place.

Did I mention that I'd prefer that all cygwin questions should go to
[EMAIL PROTECTED]?  I can't remember...

   *** CYGWIN-ANNOUNCE UNSUBSCRIBE INFO ***

If you want to unsubscribe from the cygwin-announce mailing list, look
at the "List-Unsubscribe: " tag in the email header of this message.
Send email to the address specified there.  It will be in the format:

[EMAIL PROTECTED]

If you need more information on unsubscribing, start reading here:

http://sources.redhat.com/lists.html#unsubscribe-simple

Please read *all* of the information on unsubscribing that is available
starting at this URL.

I implore you to READ this information before sending email about how
you "tried everything" to unsubscribe.  In 100% of the cases where
people were unable to unsubscribe, the problem was that they hadn't
actually read and comprehended the unsubscribe instructions.

If you need to unsubscribe from cygwin-announce or any other mailing
list, reading the instructions at the above URL is guaranteed to
provide you with the info that you need.

Christopher Faylor
Red Hat, Inc.



--060809000909020701010503
Content-Type: text/plain;
 name="Changes.w32api"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline;
 filename="Changes.w32api"

2002-11-25  Earnie Boyd  <[EMAIL PROTECTED]>

* lib/Makefile.in (dist, install): Correct the install destinations.
* lib/ddk/Makefile.in (dist, install): Ditto.

2002-11-16  Danny Smith  <[EMAIL PROTECTED]>

* lib/uuid.c (CATID_InternetAware,CLSID_HttpSProtocol,
CLSID_PSUrlMonProxy,CLSID_StdURLMoniker,IID_IObjectWithSite,
IID_IPersistMoniker,IID_IProgressNotify): Correct off-by-one bug
in definitions.
(CLSID_CFontPropPage00, GUID_FONTUNDERSCORE11,
IID_IOleControlSite89): Remove.

2002-11-15  Christopher Faylor  <[EMAIL PROTECTED]>

* lib/uuid.c (IID_IRunnableObject): Remove duplicate definition.
(IID_IStdMarshalInfo): Ditto.

2002-11-15  Christopher Faylor  <[EMAIL PROTECTED]>

* lib/uuid.c (IID_IOleInPlaceSite): Remove duplicate definition.
(IID_IPersistStorage): Ditto.

2002-11-15  Weiqi Gao  <[EMAIL PROTECTED]>

* lib/uuid.c (CLSID_PSClassObject, CLSID_PSEnumerators,
IID_IClassActivator, IID_IClientSecurity, IID_ICreateTypeInfo,
IID_ICreateTypeLib2, IID_IOleInPlaceSite, IID_IPersistStorage,
IID_IRunnableObject, IID_IServerSecurity, IID_IStdMarshalInfo):
Correct definitions.

2002-11-15  Danny Smith  <[EMAIL PROTECTED]>

* include/ddk/(atm.h, batclass.h, cfg.h, cfgmgr32.h, d4drvif.h,
d4iface.h, ddkmapi.h, hidclass.h, hidpi.h, hidusage.h, mcd.h,
miniport.h, minitape.h, mountdev.h, mountmgr.h, ndis.h,
ndisquid.h, ndistapi.h, ndisvan.h, netevent.h, netpnp.h,
netdev.h, ntapi.h, ntdd8042.h, ntddbeep.h, ntddcdrm.h,
ntddcdvd.h, ntddchgr.h, ntdddisk.h, ntddk.h, ntddkbd.h,
ntddmou.h, ntddndis.h, ntddpar.h, ntddpcm.h, ntddscsi.h,
ntddser.h, ntddstor.h, ntddtape.h, ntddtdi.h, ntddvdeo.h,
ntddvol.h, ntifs.h, ntpoapi.h, ntstatus.h, parallel.h, pfhook

can't set $USER under 1.3.16-1

2002-11-25 Thread Lester Ingber

Up until now, through cygwin 1.3.15-2, I have set my $USER in /etc/profile
to a username different from `id -un`, with the same name set in /etc/passwd.
I have had no problems logging in.

Now, under 1.3.16-1, an attempt is made to login in under `id -un` and
I cannot even telent to my perferred username?  I reinstalled 1.3.15-2,
and everything works as before installing  1.3.16-1?

What else must I do under 1.3.16-1 to set my own username?

Lester

-- 
 Prof. Lester Ingber  [EMAIL PROTECTED]  [EMAIL PROTECTED]
 www.ingber.com www.alumni.caltech.edu/~ingber

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: impure_ptr/Mingw and Cygwin

2002-11-25 Thread Andrew DeFaria

Peter A. Castro wrote:


On Sun, 24 Nov 2002, Andrew DeFaria wrote:


Peter A. Castro wrote:


What you show below is only linking. I believe you need to 
re-compile all of your source with -mno-cygwin -mwindows as well to 
make the _impure_ptr references go away.

But I did re-compile all my sources with -mno-cygwin -mwindows! Still 
 have the error.

Suggest you run nm against all your objects and libraries and find which
has the _impure_ptr references, then recompile just those.


But I did recompile everything already! The _impure_ptr keeps getting 
generated:

$ nm mksf.o | grep impure
U __impure_ptr
$ rm mksf.o
$ make mksf.o
/bin/gcc -c -g -mno-cygwin -mwindows -I. -I../../include -I/usr/include 
-DCYGWIN mksf.c
$ nm mksf.o | grep impure
U __impure_ptr
$

As you can see the _impure_ptr reference gets regenerated. But wait! 
There's more!

I can cause these _impure_ptr even with the little foo.c merely by 
including -I/usr/include on the command line. Why am I referencing 
/usr/include? Because in my larger application I need it for getopt.h. 
The mere inclusion of -I/usr/include causes_impure_ptr's to be emitted. 
So the question now is: How do I satisfy my need for getopt and still 
produce objects without _impure_ptr's?

Ah ha!

$ grep impure /usr/include/*.h
/usr/include/stdio.h:#definestdin   (_impure_ptr->_stdin)
/usr/include/stdio.h:#definestdout  (_impure_ptr->_stdout)
/usr/include/stdio.h:#definestderr  (_impure_ptr->_stderr)

Hmmm... Changed options to:

/bin/gcc -c -g -mno-cygwin -I. -I../../include -I/usr/include/mingw 
-I/usr/include -DCYGWIN mksf.c
cc1: warning: changing search order for system directory 
"/usr/include/mingw"
cc1: warning:   as it has already been specified as a non-system directory

This works but why am I getting this warning?




--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

undefined reference to impesd_audio_fromat

2002-11-25 Thread Gadget

I cant't find any refine to any fonction starting with __imp__

If anyone has ever had this problem, i would enjoy some help.
In fact I try compile Gnome.

bye
Gadget


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Problem with crontab -e: tabs aren't getting installed

2002-11-25 Thread trevin

I've recently installed Cygwin on a Windows 2000 Pro machine.  Installation 
was done under a user account.  I tried to install a cron job by running:

EDITOR=emacs crontab -e 

After reading through the mailing lists on a similar problem, I remembered 
to set the variable 'make-backup-files' to nil, and verified that the inum 
for the temporary file remained unchanged after saving the file.  However, 
when I exit out of emacs, nothing happens.  It returns to the bash prompt 
with no messages.  According to crontab -l, the table was not installed. 

Any other ideas what could be wrong? 

-- Trevin Beattie 

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

Peter A. Castro wrote:



And, just to provide a counter example: I always run Norton AntiVirus
with automatic / background scanning enabled.  I generally have to,
because of infected machines at work which probe the network whenever we
get hit with the latest rash of viruses :(.  I've done all my Cygwin
installs/updates with AntiVirus enabled and never had a BSOD (I'm running
NT4 on fairly stable hardware and up to date drivers).  So, it is
possible and if something does fail, it more like a driver fault than an
application fault.



Yes, it HAS to be a "driver" fault (more accurately, it CANNOT be an
application fault).  Applications run in ring 3 (call it "user mode" --
but it's a protected execution mode enforced by the x86 processor
hardware).  drivers and such (like virus scanners) run in ring 0 (call
it kernel mode) along with the ntos kernel.  If you get a BSOD, it is by
definition a ring 0 fault == kernel mode == drivers/virus-scanner/kernel 
problem.  [AFAIK, ring 1 and ring 2 are not used by MSWin]

Now, as far as Norton vs. McAfee, the setup.exe BSODs have most often 
been reported wrt McAfee.  I only recall one report from a year back or 
more where Norton was the culprit.

--Chuck


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: emacs 100% cpu usage busy

2002-11-25 Thread Hans Larsen

Hi,

>OK, I'll release a new emacs 

with the new cygwin.dll the *emacs problem* disappeared. In fact there never 
was an  *emacs problem*, but an cygwin.dll problem. As I already mentioned 
rxvt also shown a strange behaviour - which -of course- also *disappeared*. 
Your .9 works pretty fine with the new cygwin.dll. Please notify me in case 
there are differences between the official release and the one I downloaded.

Cheers
-Hans


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

RE:'less' can't determine terminal size?

2002-11-25 Thread Rodrigo Medina


Dear Sirs,

I think that this bug is related to the one I reported on 2002/11/25

RXVT geometry bug.

Rodrigo Medina
[EMAIL PROTECTED]



--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Setup is not installing everything!!!

On Mon, Nov 25, 2002 at 01:42:56PM -0700, Bradley Hultine wrote:
>It appears that I had a corrupt registry.  Once I went through the registry 
>and deleted all occurances of "cygwin", I could finally install a new 
>version of Cygwin.

More likely, you installed another package, like "NetworkSimplicity ssh" and
that changed your registry from the standard cygwin.

Just typing the "mount" command would probably have shown what the problem
was.  And, the mount command would have been able to fix it, too, just for
future edification.

cgf
--
Please do not send me personal email with cygwin questions or observations.
Use the resources at http://cygwin.com/ .

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: dos2unix/d2u does nothing

2002-11-25 Thread Shing-Fat Fred Ma

>
> Subject:
> RE: dos2unix/d2u does nothing
> From:
> David Kilroy <[EMAIL PROTECTED]>
> Date:
> Mon, 25 Nov 2002 13:33:42 -
> To:
> [EMAIL PROTECTED]
>
>
>I have seen similar behaviour to what Fred sees, but I can't remember if
>that was u2d or d2u. In that case the files had mixed use of \n and \r\n
>line endings. I presumed d2u/u2d detected a single \n (or \r\n) in the first
>X bytes, and assumed the file was already in the appropriate format.
>
>I 'fixed' this by running u2d then d2u (or vice versa).
>
>Dave.
>
>
>
>>-Original Message-
>>From: Randall R Schulz [mailto:[EMAIL PROTECTED]]
>>Sent: 23 November 2002 05:12
>>To: [EMAIL PROTECTED]
>>Subject: Re: dos2unix/d2u does nothing
>>
>>
>>Fred,
>>
>>It works OK for me. You may be experiencing an interaction
>>with a text mode
>>mount (though from the looks of it, "conv.c" was ported for
>>cygwin to open
>>files in binary mode, so this shouldn't happen).
>>
>>As to the mod time, perhaps you wrote the file and then
>>converted it within
>>the same minute, so "ls -l" doesn't show a change in the
>>modification time
>>(even though the difference is there at the finer time
>>resolution that the
>>OS and / or file system uses to record file modification times).
>>
>>By the way, dos2unix and d2u are identical (byte-for-byte).
>>
>>The other thing I can think of is that you're not running the
>>dos2unix from
>>the "cygutils" package, that the version you're running was
>>not ported to
>>Cygwin to be immune to the mount type and (conceivably) that
>>it resets the
>>file's modification time after reformatting it.
>>
>>Randall Schulz
>>Mountain View, CA USA
>>

Hi, All,

It's in the cygwin file tree, /usr/bin/dos2unix version 0.1.2.
The file was a few minutes old when I tried in-place conversion.
Anything is possible, regarding preserving the file timestamp
even after the conversion (afterall, I think some gzip's do that).
My "fix" is to use it as a filter.

Fred

--
Fred Ma, [EMAIL PROTECTED]
Carleton University, Dept. of Electronics
1125 Colonel By Drive, Ottawa, Ontario
Canada, K1S 5B6





--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Setup is not installing everything!!!

2002-11-25 Thread Bradley Hultine

It appears that I had a corrupt registry.  Once I went through the registry 
and deleted all occurances of "cygwin", I could finally install a new 
version of Cygwin.




_
Add photos to your e-mail with MSN 8. Get 2 months FREE*. 
http://join.msn.com/?page=features/featuredemail


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Setup is not installing everything!!!

2002-11-25 Thread Bradley Hultine

I am having a problem installing cygwin on my Win98 Laptop.  I run the setup 
program to install the base cygwin apps.  Everything goes fine, but when I 
try to run cygwin, the cygwin.bat file fails.  After further investigation, 
I found that a lot of directories are missing.  The only directories I have 
are c:\cygwin\lib, c:\cygwin\tmp, and c:\cygwin\usr.  The \bin, \etc, and 
many other directories are nowhere to be found!!

Does anyone know what is happening here?  I have installed Cygwin before, 
and have never had any problems.  I am using the current Setup.exe version 
2.249.2.5.

Thanks in advance,
Brad.





_
MSN 8 with e-mail virus protection service: 2 months FREE* 
http://join.msn.com/?page=features/virus


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

HOME set to / [Was: cygwin-1.3.16-1]

2002-11-25 Thread Eric Hanchrow

For what it's worth, I too had this problem on Windows 2000, but I was
able to work around it by putting

set HOME=/home/Administrator

into my cygwin.bat.
-- 
PGP Fingerprint: 3E7B A3F3 96CA 8958 ACC5  C8BD 6337 0041 C01C 5276



--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Unable to get Cygwin SSHD to work on Win2000 Pro SP3

2002-11-25 Thread Emilio A Icaza

Since the original upgrade to OpenSSH3.4p1-5 I have been unable to make the
Cygwin SSHD server work on my Windows 2000 Pro machine SP3.

After several weeks of testing and poking and re-installing, the problem is
the same: I have no clues as to what is wrong!

I can install the package OK.
I can start the SSHD server OK.
BUT if I try to connect to the server from localhost, I get:

(See attached file: localhost.log)

If I try to connect from a RedHat 7.2 machine, I get:

(See attached file: remote.log)

My cygcheck is:

(See attached file: cygcheck.log)

The server /var/log/ssh.log file is empty.

I would appreciate if you all can look at these text files and let me know
what I should do or if you need to look at anything else.

Thank you in advance,

[EMAIL PROTECTED]


localhost.log
Description: Binary data


remote.log
Description: Binary data


cygcheck.log
Description: Binary data
--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

select question

2002-11-25 Thread Neal D. Becker

I'm trying to use select to see if there is any user input (that would
be stdin, file descriptor 0).  If not I want to run my background
computation.  I want to execute this loop at a pretty high rate (on
the order of 100 usec).

I tried the following:

fd_set set;
FD_ZERO(&set);
FD_SET (0, &set);

timeval timeout;
timeout.tv_usec = 200;
timeout.tv_sec = 0;

int res = select (1, &set, 0, 0, &timeout);
if (res == 0) {
  do computation
else
  read user input

The problem is it seems to be ignoring the user input, even though the
load average (as reported by top) is still quite low.

Is there a problem with select being used this way?  Is there a better
way to implement this with cygwin?


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: RXVT geometry bug

2002-11-25 Thread Francis Litterio

[EMAIL PROTECTED] wrote:

> But if I run:
> rxvt -g 80x25 -fn 8X16
>
> the window that appears is BIGGER and does not properly function.

I can confirm that it's bigger -- I count 26 lines and 80 columns when
rxvt 2.7.2 is invoked with "-geometry 80x25".  The font does not seem to
matter.
--
Francis Litterio
[EMAIL PROTECTED]
http://world.std.com/~franl/
GPG and PGP public keys available on keyservers.


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Installing w/o setup.ini

2002-11-25 Thread aschwarz1309

I am attempting to install Cygwin from a CD-ROM containing downloaded Cygwin 
files (src and bin) but which does not have a setup.ini/setup.hint. I do not 
have a Cygwin/Unix enabled computer or access to the net and can't use, e.g., 
Polley Christopher W. (8 Nov 2002) script to generate a .ini file. Is there any 
way to do an installation?

Can I retrieve .ini/.hint files from ftp sites? How?

Do I have to hand generate a setup.ini file (gruesome)? And do I have to 
generate dependent package references by hand - or can I just assume that there 
are no dependencies?

Any suggestions?

art

[EMAIL PROTECTED]

Sorry if this is a duplicate post. I keep getting error messages from the e-
mail S/W.

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [Performance Regression] cygwin 1.3.15-2 -> 1.3.16-1

On Mon, Nov 25, 2002 at 05:56:12PM -, Max Bowsher wrote:
>[EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>
>> On Sun, Nov 24, 2002 at 11:59:00PM -, Max Bowsher wrote:
>>> Max Bowsher <[EMAIL PROTECTED]> wrote:
>>>
 With 16-1, I'm seeing *extreme* latency in keypresses appearing in
 bash, when I have a compile running in another Cygwin window (in my
 case, compiling setup). With 15-2, there are no such delays. I will
 try to localize the problem to between 2 snapshots.
>>>
>>> Localized to between 1.3.15-2 and 20021115, which is the earliest
>>> snapshot still available.
>>
>> I'll bet it works ok on B20, too.
>
>Putting flippancy aside, I think this is a real problem.

Hmm.  A real problem would have real details, wouldn't it?  Beyond the fact
that you think it is slow, there really weren't too many of those.

>If necessary, I will start building some DLLs from cvs so I can roll
>forward commit by commit.  But I was hoping someone might go "Aha - I
>know the change that might have caused that."

Yeah, that's always the magical thinking that goes into cygwin bug reports.

cgf

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

RE: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

2002-11-25 Thread Robb, Sam

> Anyone know of a tool (besides Purify) which could track all
> the resource usages of a given program?

Sysinternals has a great collection of tools for this sort
of thing:

Handle v2.01
http://www.sysinternals.com/ntw2k/freeware/handle.shtml
"This handy command-line utility will show you what files
are open by which processes, and much more."

Or, for the GUI minded:

Process Explorer v5.25
http://www.sysinternals.com/ntw2k/freeware/procexp.shtml
"Find out what files, registry keys and other objects processes
have open, which DLLs they have loaded, and more. This uniquely
powerful utility will even show you who owns each process."

There are other tools (regmon, filemon, diskmon, portmon)
that can be used to get blow-by-blow accounts of how different
resources are being used.

-Samrobb

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

On Sun, 24 Nov 2002, Randall R Schulz wrote:

> Hi, Chuck,

Hi All!

> Do I have to say it?
>  D'Oh!

I feel this is really just masking the real problem.  In my experiences
doing development and testing under Windows, I feel that fault drivers
are mostly what causes problems (someone else mentioned this earlier and
I tend to agree).  Occasionally hardware does go bad.  I once had a
machine that the specs claimed it could take 128Mb of memory using 32Mb
SIMMs, but actually could only handle 64Mb using 16Mb SIMMs.  With 128Mb
it became very unstable and BSOD'ed often.  In that specific case, the
problem was the memory controller.  If you are using fairly new hardware,
it's less likely that hardware is the problem, but more likely that OS
drivers are faulty.  This is *Windows* we're talking about, here, so what
did you expect :)

> At 14:35 2002-11-24, Charles Wilson wrote:
> >Guys, I'm really surprised nobody has mentioned the obvious culprit (or 
> >asked the obvious question):
> >
> >Anti virus software?  Was McAfee or Norton (but esp. McAfee) running while 
> >setup.exe was executed?
> >
> >I *always* disable McAfee before running setup.  Every time I forget, I 
> >get a bluescreen.  Minidump analysis shows that the fault is in fact 
> >McAfee -- which runs in kernel mode -- and not setup.
> 
> I use Norton AntiVirus (2002) and I *never* enable automatic / background 
> scanning. I always stick to manual scanning. Mostly I apply it to email 
> attachments and occasionally to file downloads when I don't know or trust 
> the source.

And, just to provide a counter example: I always run Norton AntiVirus
with automatic / background scanning enabled.  I generally have to,
because of infected machines at work which probe the network whenever we
get hit with the latest rash of viruses :(.  I've done all my Cygwin
installs/updates with AntiVirus enabled and never had a BSOD (I'm running
NT4 on fairly stable hardware and up to date drivers).  So, it is
possible and if something does fail, it more like a driver fault than an
application fault.

The KB article (195857) is interesting in that the system needs
additional resources to be able to free up handles.  I kinda find that
funny (and prefectly inline with the way MS does things :).  Also, I'd
think that unless the machine in question is suffering from a shortage of
resources to begin with, that it's not likely setup.exe's fault.  Anyone
know of a tool (besides Purify) which could track all the resource usages
of a given program?

> >(I also turn off McAfee when compiling; it really slows down disk IO)
> >--Chuck
> Randall

-- 
Peter A. Castro <[EMAIL PROTECTED]> or <[EMAIL PROTECTED]>
"Cats are just autistic Dogs" -- Dr. Tony Attwood

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Standard out from remsh/rsh

2002-11-25 Thread Mike Dame

I haven't tried doing a remsh from my HPUX box to cygwin 
recently, but I thought I remembered that I could get 
standard and error output from the command that I issued to 
the cygwin client. The cygwin client executes the command, but 
does not return anything to standard out.

Is this a Bug/Config/or just what?

Mike


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Help with LTrace

2002-11-25 Thread Hotmail Saturn

Here I go again...

There is no difference setting the perms via the Windows GUI... :_(

To: Pavel,

thanks too.

My  programm is in the same folder I execute strace.

I think that the problemm could be the "permission denied" to execute .
I try to execute:
./programm.exe
with the permissions: -rwxr-xr-x
and it fails me by "permission denied"

Could be this problemm due to bad downloading file?

Have anyone a list of errors reported by strace? Perhaps it helps me...

Thanks,

Saturn
S;-D

- Original Message - 
From: "Max Bowsher" <[EMAIL PROTECTED]>
To: "Hotmail Saturn" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Monday, November 25, 2002 12:15 AM
Subject: Re: Help with LTrace


> Hotmail Saturn <[EMAIL PROTECTED]> wrote:
> 
> > Thanks, Max.
> >
> > I Know strace, but it fails with:
> >
> > strace.exe: error creating process , (error 193)
> >
> > I think it fails because the execution permissions. I try to change
> > its permissions with:
> >
> > chmod +rw 
> >
> > but the result is still the same, and its permissions are:
> >
> > -rwxr-xr-x
> >
> > Is there any way to change the execution permission that let me
> > execute correctly my programm ?
> > My programm  doesn't execute because "Permission denied" error.
> 
> Hmm. Try setting the perms via the Windows GUI, just to see if that makes
> any difference.
> Does it run without strace?
> 
> > Another question: Do you know if exist an utility that let me trace
> > the libraries (like LTrace) that runs with CygWIN?
> 
> No, sorry.
> 
> --
> Max.
> 
> 
> --
> Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
> Bug reporting: http://cygwin.com/bugs.html
> Documentation: http://cygwin.com/docs.html
> FAQ:   http://cygwin.com/faq/
> 

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [Performance Regression] cygwin 1.3.15-2 -> 1.3.16-1

[EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:

> On Sun, Nov 24, 2002 at 11:59:00PM -, Max Bowsher wrote:
>> Max Bowsher <[EMAIL PROTECTED]> wrote:
>>
>>> With 16-1, I'm seeing *extreme* latency in keypresses appearing in
>>> bash, when I have a compile running in another Cygwin window (in my
>>> case, compiling setup). With 15-2, there are no such delays. I will
>>> try to localize the problem to between 2 snapshots.
>>
>> Localized to between 1.3.15-2 and 20021115, which is the earliest
>> snapshot still available.
>
> I'll bet it works ok on B20, too.

Putting flippancy aside, I think this is a real problem.
If necessary, I will start building some DLLs from cvs so I can roll forward
commit by commit.
But I was hoping someone might go "Aha - I know the change that might have
caused that."

Max.


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: impure_ptr/Mingw and Cygwin

On Sun, 24 Nov 2002, Andrew DeFaria wrote:

> Peter A. Castro wrote:
> 
> > What you show below is only linking. I believe you need to re-compile
> > all of your source with -mno-cygwin -mwindows as well to make the
> > _impure_ptr references go away.
> 
> But I did re-compile all my sources with -mno-cygwin -mwindows! Still 
> have the error.

Suggest you run nm against all your objects and libraries and find which
has the _impure_ptr references, then recompile just those. 

> >> /bin/gcc mksf.o mksfOpts.o mksfIO.o flmCrc.o -mno-cygwin -mwindows
> >> -liberty -lmingw32 -o mksf
> >> mksf.o(.text+0x60f): In function `main':
> >> /dview/defaria_2.0/salira/neopon/build/maketools/mksf.c:29: undefined
> >> reference to `_impure_ptr'
> >> mksf.o(.text+0x874):/dview/defaria_2.0/salira/neopon/build/maketools/mksf.c:66:
> >> undefined reference to `_impure_ptr'
> >> mksfOpts.o(.text+0x309): In function `usage':
> >> /dview/defaria_2.0/salira/neopon/build/maketools/mksfOpts.c:13: undefined
> >> reference to `_impure_ptr'
> >> mksfOpts.o(.text+0x331):/dview/defaria_2.0/salira/neopon/build/maketools/mksfOp
> >> s.c:14: undefined reference to `_impure_ptr'
> >> mksfOpts.o(.text+0x352):/dview/defaria_2.0/salira/neopon/build/maketools/mksfOp
> >> s.c:15: undefined reference to `_impure_ptr'
> >> mksfOpts.o(.text+0x36a):/dview/defaria_2.0/salira/neopon/build/maketools/mksfOp
> >> s.c:16: more undefined references to `_impure_ptr' follow
> >> make: *** [mksf] Error 1
> >>
> >> Again the _impure_ptr things come up. Where is this supposed to be
> >> resolved from?

-- 
Peter A. Castro <[EMAIL PROTECTED]> or <[EMAIL PROTECTED]>
"Cats are just autistic Dogs" -- Dr. Tony Attwood


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: impure_ptr/Mingw and Cygwin

On Sun, 24 Nov 2002, Max Bowsher wrote:

> Peter A. Castro <[EMAIL PROTECTED]> wrote:
> 
> > On Sat, 23 Nov 2002, Max Bowsher wrote:
> >> $ gcc -g foo.c -mno-cygwin -mwindows -o foo -liberty -lmingw32
> >> $ ./foo.exe > x
> >> Hello World 2
> >> $ cat x
> >> Hello World
> >>
> >> How odd. I get the stderr output just fine.
> >
> > Yes, very interesting.  I'll have to see if perhaps stderr output is
> > appearing and the command prompt is simple overlaying it after the
> > command returns.  BTW, I was running the program from a regular DOS
> > Prompt, not a bash shell (not that that should make a big difference
> > for console I/O).
> >
> >> Why the -mwindows -liberty -lmingw32 switches? They are unnecessary.
> >
> > Because that's how setup.exe is built (that's where I got the above
> > switches from).
> 
> -liberty ? Not on my computer.

iberty just provides extra functions which might be missing from a given
environment.  It's strictly not needed, but doesn't hurt either. 

> -lmingw32 ? I think that's just one of the redundant things libtool like to
> add.

This really has nothing to do with libtool.  Try adding that '-v' option
to the command line and look at the list of libraries being pulled in by
the compiler.  Since the compiler pulls in mingw32 implicitly from the
-mno-cygwin flag you can leave this library off. 

> -mwindows ? Sparsely documented, if at all. I *think* it is to link as a
> Windows GUI exe rather than a Windows console exe.

Yep.  Adds flags to the linker to pull in windows GUI libraries and
possibly determine which startup object to use (console vs GUI vs ...)

> So, I think you can lose all 3 of these.

Depends on what he's coding for.  I was under the impression from other
email that he might need more than the bare minimum that foo.c represents.
If it's likely to have graphics in it, then I'd leave the -mwindows and
he can always add -liberty back in if things don't resolve.

However, his current problem is that he still gets references to
_impure_ptr which doesn't make sense, since that would really only be
generated if he compiled without -mno-cygwin, and he claims he has
recompiled with it.

> --
> Max.

-- 
Peter A. Castro <[EMAIL PROTECTED]> or <[EMAIL PROTECTED]>
"Cats are just autistic Dogs" -- Dr. Tony Attwood

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [Performance Regression] cygwin 1.3.15-2 -> 1.3.16-1

2002-11-25 Thread cygwin

On Sun, Nov 24, 2002 at 11:59:00PM -, Max Bowsher wrote:
>Max Bowsher <[EMAIL PROTECTED]> wrote:
>
>> With 16-1, I'm seeing *extreme* latency in keypresses appearing in
>> bash, when I have a compile running in another Cygwin window (in my
>> case, compiling setup). With 15-2, there are no such delays. I will
>> try to localize the problem to between 2 snapshots.
>
>Localized to between 1.3.15-2 and 20021115, which is the earliest snapshot
>still available.

I'll bet it works ok on B20, too.

cgf

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: cygwin with opengl

2002-11-25 Thread Andre Bleau

Nigel wrote:

James, Thanks for such an informative and detailed response. It turns out 
to be a FAQ for our OpenGL students who like to compile on Linux and 
Cygwin. http://goanna.cs.rmit.edu.au/~gl/teaching/cs549/index.html Do you 
mind if I recycle your posting? And yes, we hit the same "but why doesn't 
gcc just _know_ what I want, like MSVC _seems_ to..." mentality... Cheers, 
Nigel

It would be simpler and better if you directed your students to the 
relevant README file:

/usr/doc/opengl-1.1.0/README.txt



André Bleau, Cygwin's OpenGL package maintainer.

email: bleau at igb dot umontreal dot ca
(Fight SPAM: encode your email-address)

Please address all questions and problem reports about Cygwin's OpenGL 
package to [EMAIL PROTECTED] .


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [Patch] skipping import libraries for performance reasons - direct auto-import of dll's

Please separate policy arguments from functional arguments.  Whether 
cygwin/mingw should drop import libs is policy.  Whether libtool should 
use/create import libs is policy.  Whether ld should support auto-import 
when doing "link-directly-to-dll" is functional.

[Note: I haven't tested your patch, but it looks mostly ok]

I think the ability to do auto-import when using ld's virtual, 
on-the-fly import lib feature (e.g. link-directly-to-dll) is valuable, 
and complete support (once achieved) should be accepted.

However...

There does not seem to be any support in your patch for the most recent 
auto-import capabilities -- the runtime-pseudo-reloc stuff added by 
Egor.  So, if your patch is accepted as-is, then we have:

(1) if linking to implib, then "full" auto-import support is available; 
--enable-pseudo-reloc allows auto-import of "complex" data (see ld.info).

(2) if linking directly to the dll without an implib, then "partial" 
auto-import is available; --enable-pseudo-reloc has no effect.  And 
doesn't warn that it fails.  I *think* you'd end up with a runtime 
error, if attempting to import "complex" data.

Bad.

Two solutions:

(1) add support for runtime-pseudo-relocs to your patch, and resubmit. 
You might want to work with Egor on this.

(2) When doing "link-directly-to-dll" AND a symbol has a non-zero addend 
(e.g. "complex" data import), then behave as if 
link_info.pei386_runtime_pseudo_reloc = 0, even if --enable-pseudo-reloc 
has been specified.  (Of course, IF --enable-pseudo-reloc has been 
specified and the current link obj is an import lib, then psuedo-reloc 
should occur).  This complex behavior may actually be more difficult -- 
and more confusing -- than simply implementing solution (1).

Minor nit about patch format: watch out for your tab/indentation.  It 
doesn't match the surrounding text in many cases.

Note: yes, I realize that the runtime-pseudo-reloc stuff requires actual 
runtime support (extra code in crt0.o for mingw, extra code in 
cygwin1.dll for cygwin) which does not yet exist.  But it will, 
eventually -- and that's outside ld's purview.  If "normal" 
import-lib-based linking includes support for --enable-pseudo-relocs, 
then "virtual on-the-fly direct" linking should support it too.  Let 
mingw/crt0.o and cygwin/cygwin1.dll take care of themselves.  Downside: 
it's hard for you, Ralf, test test whether you have properly implemented 
the runtime-pseudo-reloc support for virtual on-the-fly direct linking, 
unless you build your own cygwin kernel with the appropriate 
support...talk to Egor...

--
Chuck


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

RE: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

2002-11-25 Thread John Vincent


Hi,

I don't want to be accused of advertising, but we use SOPHOS
in my office, and it doesn't interfere with the download/install
processes or setup.exe at all. I've used it on NT, 2000, and XP.

Just for information.

/John.



_
Help STOP SPAM with the new MSN 8 and get 2 months FREE*  
http://join.msn.com/?page=features/junkmail


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

RE: Help with LTrace

2002-11-25 Thread Pavel Rozenboim



> -Original Message-
> From: Randall R Schulz [mailto:[EMAIL PROTECTED]]
> Sent: Mon, November 25, 2002 5:22 PM
> To: [EMAIL PROTECTED]
> Subject: RE: Help with LTrace
> 
> 
> Pavel,
> 
> Perhaps that's because "strace" isn't a Cygwin program:
> 
> % cygcheck strace
> Found: D:\cygwin\bin\strace.exe
> D:\cygwin\bin\strace.exe
>D:\WINNT\System32\msvcrt.dll
>  D:\WINNT\System32\KERNEL32.dll
>D:\WINNT\System32\NTDLL.DLL
> 
> 
> Cygcheck itself exhibits the same symptom (it also is not 
> linked against 
> libcygwin.a):
> 
> % cygcheck /bin/strace
> /bin/strace - Cannot open
> 
> % cygcheck /bin/strace.exe
> /bin/strace.exe - Cannot open
> 
> 
> Randall Schulz
> Mountain View, CA USA
> 
> 
> At 23:35 2002-11-24, Pavel Rozenboim wrote:
> >It seems that sometimes, full path confuses strace.
> >For example in my system:
> >
> >~> strace /bin/ls
> >strace.exe: error creating process /bin/ls , (error 3)
> >
> >~> strace ls
> >[Works, lots of output]
> 
> 
> --
> Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
> Bug reporting: http://cygwin.com/bugs.html
> Documentation: http://cygwin.com/docs.html
> FAQ:   http://cygwin.com/faq/
> 

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

RE: Help with LTrace

2002-11-25 Thread Randall R Schulz

Pavel,

Perhaps that's because "strace" isn't a Cygwin program:

% cygcheck strace
Found: D:\cygwin\bin\strace.exe
D:\cygwin\bin\strace.exe
  D:\WINNT\System32\msvcrt.dll
D:\WINNT\System32\KERNEL32.dll
  D:\WINNT\System32\NTDLL.DLL


Cygcheck itself exhibits the same symptom (it also is not linked against 
libcygwin.a):

% cygcheck /bin/strace
/bin/strace - Cannot open

% cygcheck /bin/strace.exe
/bin/strace.exe - Cannot open


Randall Schulz
Mountain View, CA USA


At 23:35 2002-11-24, Pavel Rozenboim wrote:
It seems that sometimes, full path confuses strace.
For example in my system:

~> strace /bin/ls
strace.exe: error creating process /bin/ls , (error 3)

~> strace ls
[Works, lots of output]



--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

2002-11-25 Thread Daniel Armbrust

One further clarification.  It happens well before the

	Installation complete
		[OK]

step.  It actually happens before any of the post download scripts run. 
 It happens when it is at (or nearly at) the very end of downloading 
the packages.  After the blue screen, when I come back in, and just 
click through the install steps again, it usually redownloads or 
finishes one package, and then moves on to the install step (which works 
fine).


Do the MS debugging tools ship with NT 4?  Or do I have to pick them up 
somewhere on the MS site?  I don't do much windows programming, and I 
haven't yet been able to make java cause a blue screen...  : )

But I'm pretty sure I can recreate the crash later today.

Dan


--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: emacs 100% cpu usage busy

2002-11-25 Thread Joe Buehler

Hans Larsen wrote:


With regards to emacs, which worked properly until some weeks ago, I 
definitely fixed the problem by replacing the binary with the binary I 
downloaded from http://68.98.180.124:3000/cygwin/emacs/ without changing 
anything else.

OK, I'll release a new emacs.

Joe Buehler




--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

RXVT geometry bug

2002-11-25 Thread [EMAIL PROTECTED]

Dear Sirs,

I have observed that rxvt v2.7.2 (installed on 2002/10/27) does not
 correctly processes the geometry option.
If I run:
rxvt -fn 8X16

a 80x25 window appears that functions OK.

But if I run:
rxvt -g 80x25 -fn 8X16

the window that appears is BIGGER and does not properly function.
For example, inside "less" the down-arrow does not work.

If I resize this broken window, making it smaller and then bigger,
so that at the end it has the same size as before, it starts working
normaly.

This behavior is the same for any dimensions you give.

I hope that this observation could be usefull.

Rodrigo Medina
[EMAIL PROTECTED]



--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: [ANNOUNCEMENT] Updated Cygwin Package: python-2.2.2-1

2002-11-25 Thread Jason Tishler

Rolf,

On Sun, Nov 24, 2002 at 01:33:31AM -0500, Rolf Campbell wrote:
> > -Original Message-
> > From: Jason Tishler [mailto:[EMAIL PROTECTED]] 
> > Sent: Tuesday, November 19, 2002 7:18 AM
> > To: [EMAIL PROTECTED]
> > Subject: Re: [ANNOUNCEMENT] Updated Cygwin Package: python-2.2.2-1
> > 
> > On Mon, Nov 18, 2002 at 03:49:32PM -0500, Rolf Campbell wrote:
> > > I have patched pyserial (locally) and it seems to work fine now.
> > 
> > Please submit your patch to pyserial's patch collector on SF 
> > for consideration.
> > 
> 
> The CVS version of pyserial now works with Cygwin python-2.2.2-1.

Thanks for following up -- it is much appreciated.

Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

RE: dos2unix/d2u does nothing

2002-11-25 Thread David Kilroy

I have seen similar behaviour to what Fred sees, but I can't remember if
that was u2d or d2u. In that case the files had mixed use of \n and \r\n
line endings. I presumed d2u/u2d detected a single \n (or \r\n) in the first
X bytes, and assumed the file was already in the appropriate format.

I 'fixed' this by running u2d then d2u (or vice versa).

Dave.

> -Original Message-
> From: Randall R Schulz [mailto:[EMAIL PROTECTED]]
> Sent: 23 November 2002 05:12
> To: [EMAIL PROTECTED]
> Subject: Re: dos2unix/d2u does nothing
> 
> 
> Fred,
> 
> It works OK for me. You may be experiencing an interaction 
> with a text mode 
> mount (though from the looks of it, "conv.c" was ported for 
> cygwin to open 
> files in binary mode, so this shouldn't happen).
> 
> As to the mod time, perhaps you wrote the file and then 
> converted it within 
> the same minute, so "ls -l" doesn't show a change in the 
> modification time 
> (even though the difference is there at the finer time 
> resolution that the 
> OS and / or file system uses to record file modification times).
> 
> By the way, dos2unix and d2u are identical (byte-for-byte).
> 
> The other thing I can think of is that you're not running the 
> dos2unix from 
> the "cygutils" package, that the version you're running was 
> not ported to 
> Cygwin to be immune to the mount type and (conceivably) that 
> it resets the 
> file's modification time after reformatting it.
> 
> Randall Schulz
> Mountain View, CA USA
> 
> 
> At 20:36 2002-11-22, Shing-Fat Fred Ma wrote:
> 
> >Hello,
> >
> >I'm finding that dos2unix and d2u doesn't change a file 
> in-place, even 
> >with -U (the timestamp doesn't even change).  It works find for 
> >stdin-to-stdout, though.   Just thought I'd share my 
> feelings on that 
> >(that is, I feel it doesn't work in-place, but thank 
> goodness it works).
> >
> >Fred
> 
> 
> --
> Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
> Bug reporting: http://cygwin.com/bugs.html
> Documentation: http://cygwin.com/docs.html
> FAQ:   http://cygwin.com/faq/
> 

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Text Problems with Cygwin > 1.3.14 and TinyFugue WAS Re: Text problems with Cygwin >1.1.14 and TinyFugue

2002-11-25 Thread Jonathan Fosburgh

On Wednesday 20 November 2002 02:37 pm, you wrote:

>
> Curiouser and curiouser.  With 1.3.15 and the 11/20 snapshot I can run tf
> from within strace. That is, I can enter text into it and even connect to
> worlds.  That is not to say that I can *use* it, the output doesn't work
> quite correctly.  HTH.

SImply noting that the problem persists in 1.3.16.  I'll continue trying to 
look into it.

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: gcc problem?

2002-11-25 Thread CHAUVIERE Jean-Raymond

I didn't program in C for a long time but as far as I remember,
scanf read 1 string using white space as separator.

to enter aaa bbb ccc
you need scanf("%s %s %s", ...) or gets

use gets
or scanf("%s", ..)
followed by
while('\n'!=getchar());
but you loose the extra typed words

hope this helps.


On Thu, 21 Nov 2002 10:46:26 +0800, "Carlo Florendo" <[EMAIL PROTECTED]>
wrote:

>Hello,
>
>Ever since I installed a newer cygwin, I've encountered problems which I
>didn't encounter before.  First, there was the
>"ls -l"problem which has not yet been resolved (and which is threaded as "ls
>problem" in this list.).  Just today, i discovered something wrong while
>using gcc.  I  compiled the snippet below and it's supposed to prompt me for
>input twice.  However, I only get prompted once.  (Using the visual c++
>compiler, the borland 5.5 compiler gives the correct results)

The reverse is true for the example below: ie you were encountering
problems before, but not realising it.

>My gcc version is 2.95.3-5.
>Cygwin version is The cygwin1.dll version I am using is
>1.3.15-cygwin-1-3-15-1.
>
>---begin snippet-
>#include 
>int main()
>{
>   int n;
>   char string[80];
>   for ( n=0 ; n<2 ; n++ )
>   {
> printf( "Enter some words: " );
> scanf( "%s", string);
> printf( "The first word you entered is : %s\n", string );
> fflush ( stdin );
^^^hint: "while('\n'!=getchar());"
>   }
>   return 0;
>}

You can't flush(stdin). Nothing wrong with Cygwin here, simply the other
compilers implementing stuff that isn't part of the language. The FAQ
for comp.lang.c or alt.comp.lang.learn.c-c++ will no doubt give you
ample insight.





--
CHAUVIERE Jean-Raymond
06 80 38 01 14

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Problem with Cygwin DLL: 1003.15.0.0 & 1003.16.0.0

2002-11-25 Thread Steve Fairbairn


Hi,

I have come across a problem with cygwin which relates to the Windows
CreateProcess() function.

Tested on 
Cygwin DLLs: 1003.15.0.0 & 1003.16.0.0
Windows 2000 sp3

I have a Visual C++ program which uses CreateProcess to launch an executable
much like the example provided in the attachment.  The example has been hard
coded to execute a cygwin compiled binary (ls) as this is what I was
attempting to do in the real thing.

When the program is run from a dos command prompt, it all works perfectly.
However when the program is run from cygwin bash, the launched program (ls)
fails with a read exception somewhere in the cygwin dll.

If the program is changed to execute a non cygwin-built binary, it all seems
to work fine.

I also bring your attention to the following lines from the example...

/* Something to do with bInheritHandles being FALSE, as when it is TRUE,
it works. */
CreateProcess(0, cmd, 0, 0, FALSE,
DETACHED_PROCESS | NORMAL_PRIORITY_CLASS,
0,
cwd,
&StartupInfo,
&ProcessInformation);

When the CreateProcess call is changed so handles are inherited, it all
works perfectly.

If this is the wrong place to report an issue like this, then please let me
know where I should email instead.  I couldn't spot any other likely
candidates.

Regards,

Steve Fairbairn.

 <> 
PS. Full VC++ Workspace and binaries are available, I just didn't want to
attach it on a mail to the whole list.
PPS. Apologies for the disclaimer, nothing I can do to stop it :(.




*** 
This email has originated from Perwill plc (Registration No. 1906964) 
Office registered at: 13A Market Square, Alton, Hampshire, GU34 1UR, UK 
Tel: +44 (0)1420 545000 
Fax: +44 (0)1420 545001 
www.perwill.com 
*** 
Privileged, confidential and/or copyright information may be contained 
in this email, and is only for the use of the intended addressee. 
To copy, forward, disclose or otherwise use it in any way if you are not 
the intended recipient or responsible for delivering to him/her is
prohibited.
If you receive this email by mistake, please advise the sender immediately, 
by using the reply facility in your email software.

We may monitor the content of emails sent and received via our network 
for the purposes of ensuring compliance with policies and procedures. 
This message is subject to and does not create or vary any contractual 
relationships between Perwill plc and the recipient. 
*** 
Any opinions expressed in the email are those of the sender and not 
necessarily of Perwill plc.
*** 
This email has been scanned for known viruses using 
McAfee WebShield 4.5 MR1a 
*** 





test.c
Description: Binary data
--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

[Patch] skipping import libraries for performance reasons - direct auto-import of dll's

2002-11-25 Thread Ralf Habacker

Hi all,

the recent ld contains support importing of symbols located in other dll's
through import libraries.

Especially for big libraries using import libraries has some big disadvantages:

1. ld need much time and very much memory to process them  - This is because
each symbol is located in a seperate objectfile and ld opens a bfd for each.

2. The size of import libraries could be very huge (for example 10 MB for
qt-3) - This is because each symbol is located in a seperate objectfile.

current ld releases are able to skip import libraries by grabbing the needed
informations directly from a dll. The only issue which was open until now, is
that ld has no auto-import support for linking directly to a dll, but this is be
fixed with the appended patch.

Using dll's directly instead of import libraries has several advantages:

1. linking goes much faster (for qt-3 about 50%: from 2 minutes down to one
minute)

2. ld uses much less memory (for qt-3 about 70%: from 500 MB to 150 MB)

3. ld works more like the linux version. There are only static archives and
shared libraries which could be linked directly without the indirection of using
import libraries. This simplifies for example libtool handling.

How does it works:

See the example path layout. There are two dll's xxx.dll and yyy.dll in an
applications bin dir. In the lib dir there are the import library for xxx.dll
and a link to the dll cygyyy.dll.

../bin/
cygxxx.dll
cygyyy.dll

../lib/
libxxx.dll.a
[cyg|lib|]yyy.dll   (may be the dll  or a symbolic link to 
../bin/cygyyy.dll)


Now the link line (you can see that using the dll directly is the same as using
the import libraries, there is no difference)

gcc -Wl,-verbose  -o a.exe -L../lib/ -lxxx -lyyy ...

The ld log below shows, that for xxx the import library is used and for yyy the
dll is directly used. This means, the only task of using the dll instead of the
import library is to replace the import library with a symbolic link to the dll.


==
attempt to open /usr/lib/crt0.o succeeded
/usr/lib/crt0.o
attempt to open client.o succeeded
client.o
attempt to open ../lib/libxxx.dll.a succeeded !!!
...
attempt to open ../lib/libyyy.dll.a failed
attempt to open ../lib/yyy.dll.a failed
attempt to open ../lib/libyyy.a failed
attempt to open ../lib/cygyyy.dll failed
attempt to open ../lib/libyyy.dll failed
attempt to open ../lib/yyy.dll succeeded !!!

Info: resolving _var by linking to __imp__var (auto-import)
Info: resolving _foo by linking to __imp__foo (auto-import)
Info: resolving _func_ptr by linking to __imp__func_ptr (auto-import)

I have tested this patch with recent qt-2/3 and kde2 sources without any
problems. A sample test case is appended too.

Please give this patch a try and report problems to me o0r to this list.

Thank you

Ralf Habacker

KDE on cygwin http://cygwin.kde.org







ld-auto-import-dll.patch
Description: Binary data


ld-test.tar.bz2
Description: Binary data
--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

who succed to compile freetype 2.X ?

2002-11-25 Thread CHAUVIERE Jean-Raymond

I didn't succed cause of the apparently wellknown problem of _WinMain@16.
I read all the messages on different lists, the faq (but don't want to add a 
main() )
the -Wl,-u, ... has not effect.
I need the freetype as a pre requisite to compile the gd. 2.0
to get imagefontheight in php4

But if someone has a previous version of gd and other image lib as dll, i will 
be happy if he can give me the url link.

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

Re: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)

2002-11-25 Thread Frank-Michael Moser

(Danilo: Sorry for accidently sending my previous mail directly to you)
Danilo Turina wrote:

 > Since 6/7 months I have 0x000a errors with my PC (Win2K SP3), the
 > problem is very simple: when XFree is running and there is a write to
 > /etc I got a blue screen.
 > Here are the steps I perform:
 >
 > * start XFree (XWin+WindowMaker+ all that is launched by
 > startxwin.bat);
 > * open a xterm;
 > * touch /etc/pippo (or any other file you want, within /etc).
 >

Unfortunately I can exactly reproduce your problem without any McAffe or
other AV stuff. I'm running the latest Cygwin on XP-Home. I attached the
cygcheck -s and in a previous mail I wanted to attach the mini-dump but
I was not allowed to send it:
<[EMAIL PROTECTED]>:
ezmlm-reject: fatal: Sorry, I don't accept messages larger than 10
bytes (#5.2.3)

My system unfortunately did restart at once so I did not find any
information from the stack but if I could provide some more useful
information or any debug support (but I do not know anything about C!)
please let me know.

Frank-Michael

Cygwin Win95/NT Configuration Diagnostics
Current System Time: Mon Nov 25 13:25:36 2002

Windows XP Home Edition Ver 5.1 Build 2600 

Path:   .
c:\home\moser\scripts\cygwin
c:\home\moser\scripts
c:\home\moser\bin\cygwin
c:\home\moser\bin
c:\home\moser\work\bin
.
C:\jdk\1.4.1\bin
C:\cygwin\usr\X11R6\bin
C:\cygwin\bin\X11
C:\cygwin\usr\local\bin
C:\cygwin\bin
C:\cygwin\bin
c:\Programme\system
c:\WINDOWS\system32
c:\WINDOWS
c:\WINDOWS\System32\Wbem
c:\Programme\Network Associates\PGPcmdln
C:\cygwin\usr\X11R6\bin

SysDir: C:\WINDOWS\System32
WinDir: C:\WINDOWS

HOME = `c:\home\moser'
MAKE_MODE = `unix'
PWD = `/home/moser'
USER = `moser'

Use `-r' to scan registry

a:  fd   N/AN/A
c:  hd  NTFS   28607Mb  83% CP CS UN PA FC PIP
d:  cd   N/AN/A
e:  cd   N/AN/A
h:  hd  NTFS   28607Mb  83% CP CS UN PA FC PIP
i:  net NTFS   51095Mb  91% CP CSPAWeb_Installers
m:  hd  NTFS   28607Mb  83% CP CS UN PA FC PIP
n:  net NTFS   16042Mb  90% CP CSPAILVM-Bilder
p:  net NTFS   16042Mb  90% CP CSPAExport
s:  net NTFS   51095Mb  91% CP CSPAsoftware
t:  net NTFS   16042Mb  90% CP CSPATransfer
u:  net NTFS   51095Mb  91% CP CSPAmoser
w:  net NTFS   51095Mb  91% CP CSPAdecodon
x:  hd  NTFS   28607Mb  83% CP CS UN PA FC PIP
y:  hd  NTFS   28607Mb  83% CP CS UN PA FC PIP
z:  hd  NTFS   28607Mb  83% CP CS UN PA FC PIP

C:\cygwin  / system  binmode
\\dagobert\moser   /home/gustav/mosersystem  binmode
\\dagobert\software/home/public/software system  binmode
\\dagobert\transfer/home/public/transfer system  binmode
C:\cygwin/bin  /usr/bin  system  binmode
C:\cygwin/lib  /usr/lib  system  binmode
\\dagobert\cvsroot /usr/local/cvsrootsystem  binmode
C:\jdk\1.4.1   /usr/local/java   system  binmode
C:\cygwin\usr\X11R6\lib\X11\fonts  /usr/X11R6/lib/X11/fonts  system  binmode
.  /cygdrive userbinmode,cygdrive

Found: C:\cygwin\bin\bash.exe
Found: C:\cygwin\bin\cat.exe
Found: C:\cygwin\bin\cpp.exe
Found: C:\cygwin\bin\find.exe
Found: C:\cygwin\bin\gcc.exe
Found: C:\cygwin\bin\gdb.exe
Found: C:\cygwin\bin\ld.exe
Found: C:\cygwin\bin\ls.exe
Found: C:\cygwin\bin\make.exe
Found: C:\cygwin\bin\sh.exe

   41k 2002/05/14 C:\cygwin\usr\X11R6\bin\cygPropList-0.dll
   58k 2002/05/07 C:\cygwin\bin\cygbz2-1.dll
   54k 2002/01/27 C:\cygwin\bin\cygbz21.0.dll
6k 2002/06/24 C:\cygwin\bin\cygcharset-1.dll
  643k 2002/11/09 C:\cygwin\bin\cygcrypto.dll
  493k 2002/11/19 C:\cygwin\bin\cygcurl-2.dll
   50k 2002/03/17 C:\cygwin\bin\cygexslt-0.dll
   45k 2001/04/25 C:\cygwin\bin\cygform5.dll
   35k 2002/01/09 C:\cygwin\bin\cygform6.dll
   19k 2002/02/20 C:\cygwin\bin\cyggdbm.dll
   17k 2001/06/28 C:\cygwin\bin\cyghistory4.dll
   20k 2002/10/10 C:\cygwin\bin\cyghistory5.dll
  929k 2002/06/24 C:\cygwin\bin\cygiconv-2.dll
   22k 2001/12/13 C:\cygwin\bin\cygintl-1.dll
   28k 2002/09/20 C:\cygwin\bin\cygintl-2.dll
   21k 2001/06/20 C:\cygwin\bin\cygintl.dll
   81k 2000/12/05 C:\cygwin\bin\cygitcl30.dll
   35k 2000/12/05 C:\cygwin\bin\cygitk30.dll
   45k 2002/02/08 C:\cygwin\bin\cygjbig1.dll
  119k 2002/02/09 C:\cygwin\bin\cygjpeg6b.dll
   59k 2002/09/20 C:\cygwin\bin\cygkpathsea-3-3-7.dll
   25k 2002/07/16 C:\cygwin\bin\cygltdl-3.dll
   26k 2001/04/25 C:\cygwin\bin\cygmenu5.dll
   20k 2002/01/09 C:\cygwin\bin\cygmenu6.dll
  156k 2001/04/25 C:\cygwin\bin\cygncurses++5.dll
  175k 2002/0

Re: net use and sshd

2002-11-25 Thread Corinna Vinschen

On Mon, Nov 25, 2002 at 11:51:05AM +0100, Romain Berrendonner wrote:
> sshd version: OpenSSH_3.4p1
> cugwin version: 1.3.12
> OS: Windows XP
> 
> If I use a password authentication with sshd, I'm able to map a network drive
> by mean of the windows 'net use' command. If I try to map the very same network
> drive after having authenticated through the public key method, mapping 
> doesn't work any longer, and I get the following message:
> 
> System error 5 has occured.
> Access is denied.
> 
> Is this behaviour expected ? Is there a workaround ?

  Yes. No.

Corinna

-- 
Corinna Vinschen  Please, send mails regarding Cygwin to
Cygwin Developermailto:[EMAIL PROTECTED]
Red Hat, Inc.

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

net use and sshd

2002-11-25 Thread Romain Berrendonner

sshd version: OpenSSH_3.4p1
cugwin version: 1.3.12
OS: Windows XP

If I use a password authentication with sshd, I'm able to map a network drive
by mean of the windows 'net use' command. If I try to map the very same network
drive after having authenticated through the public key method, mapping 
doesn't work any longer, and I get the following message:

System error 5 has occured.
Access is denied.

Is this behaviour expected ? Is there a workaround ?

-- 
Romain

--
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ:   http://cygwin.com/faq/

RE: Cygwin causes 0x00000024 Stop Error (BLUE SCREEN)