Re: Patch to avoid using mkfifo under cygwin

Re: Patch to avoid using mkfifo under cygwin

[Nicholas Wourms]

On Fri, Sep 13, 2002 at 11:07:38AM -0700, Keith Seitz wrote:
 >>On Fri, 13 Sep 2002, Mo DeJong wrote:
 >>>This was causing a failure in the Tcl test suite, but it
 >>>should not matter to anyone.  Since Cygwin will add the
 >>>feature soon, it seems fine to just leave it the way it
 >>>is.
 >>
 >>Ah, okay.  Thanks for the clarification.

 >Ditto.  Thanks, Mo.  This should be available soon when we
 >release "cygserver".

Is this mkfifo support "vaporware" or is it something that 
could be checked into the cygdaemon-branch for testing? 
IIRC, nothing was said to indicate that this was actually 
being fixed.  All I remember is that Robert said he was too 
busy and Max said he might look at it.  However, this is the 
first peep regarding the issue I heard since then.  Clue me 
in if I'm wrong.

Cheers,
Nicholas

Patch to avoid using mkfifo under Cygwin

[Mo DeJong]

Here is a patch that avoids calling mkfifo under Cygwin since
it is not functional.

Mo

2002-09-12  Mo DeJong  <supermo@bayarea.net>

	* unix/tclUnixFCmd.c (DoCopyFile): Don't use mkfifo
	when compiling with Cygwin, since it is not supported.

Index: unix/tclUnixFCmd.c
===================================================================
RCS file: /cvs/src/src/tcl/unix/tclUnixFCmd.c,v
retrieving revision 1.2
diff -u -r1.2 tclUnixFCmd.c
--- unix/tclUnixFCmd.c	9 Sep 2001 23:56:09 -0000	1.2
+++ unix/tclUnixFCmd.c	12 Sep 2002 22:37:10 -0000
@@ -391,12 +391,18 @@
 	    }
 	    return CopyFileAtts(src, dst, &srcStatBuf);
 	}
+#ifndef __CYGWIN__
+        /*
+         * mkfifo is not supported under Cygwin even though it is prototyped
+         * in newlib headers.
+         */
         case S_IFIFO: {
 	    if (mkfifo(dst, srcStatBuf.st_mode) < 0) {	/* INTL: Native. */
 		return TCL_ERROR;
 	    }
 	    return CopyFileAtts(src, dst, &srcStatBuf);
 	}
+#endif /* __CYGWIN__ */
         default: {
 	    return CopyFile(src, dst, &srcStatBuf);
 	}

Re: Patch to avoid using mkfifo under cygwin

[Christopher Faylor]

On Thu, Sep 12, 2002 at 03:51:53PM -0700, Mo DeJong wrote:
>Here is a patch that avoids calling mkfifo under Cygwin since
>it is not functional.

Actually mkfifo will soon be useful in cygwin.

cgf

Re: Patch to avoid using mkfifo under Cygwin

[Keith Seitz]

On Thu, 12 Sep 2002, Mo DeJong wrote:

> Here is a patch that avoids calling mkfifo under Cygwin since
> it is not functional.

Ok, so given cgf's latest comments, do we want to hold off on this or 
revert it when the cygwin work is ready? Maybe we can write an autoconf 
test for it?

Keith

> 2002-09-12  Mo DeJong  <supermo@bayarea.net>
> 
> 	* unix/tclUnixFCmd.c (DoCopyFile): Don't use mkfifo
> 	when compiling with Cygwin, since it is not supported.
> 
> Index: unix/tclUnixFCmd.c
> ===================================================================
> RCS file: /cvs/src/src/tcl/unix/tclUnixFCmd.c,v
> retrieving revision 1.2
> diff -u -r1.2 tclUnixFCmd.c
> --- unix/tclUnixFCmd.c	9 Sep 2001 23:56:09 -0000	1.2
> +++ unix/tclUnixFCmd.c	12 Sep 2002 22:37:10 -0000
> @@ -391,12 +391,18 @@
>  	    }
>  	    return CopyFileAtts(src, dst, &srcStatBuf);
>  	}
> +#ifndef __CYGWIN__
> +        /*
> +         * mkfifo is not supported under Cygwin even though it is prototyped
> +         * in newlib headers.
> +         */
>          case S_IFIFO: {
>  	    if (mkfifo(dst, srcStatBuf.st_mode) < 0) {	/* INTL: Native. */
>  		return TCL_ERROR;
>  	    }
>  	    return CopyFileAtts(src, dst, &srcStatBuf);
>  	}
> +#endif /* __CYGWIN__ */
>          default: {
>  	    return CopyFile(src, dst, &srcStatBuf);
>  	}
> 

Re: Patch to avoid using mkfifo under cygwin

[Christopher Faylor]

On Fri, Sep 13, 2002 at 07:47:24AM -0700, Keith Seitz wrote:
>On Thu, 12 Sep 2002, Mo DeJong wrote:
>
>> Here is a patch that avoids calling mkfifo under Cygwin since
>> it is not functional.
>
>Ok, so given cgf's latest comments, do we want to hold off on this or 
>revert it when the cygwin work is ready? Maybe we can write an autoconf 
>test for it?

Given that this is something that will just magically work when the cygwin
DLL is updated, I don't think autoconf is the right way to go.

What's wrong with the current code which returns an error when mkfifo fails?
That seems pretty clear to me.

cgf

>>          case S_IFIFO: {
>>  	    if (mkfifo(dst, srcStatBuf.st_mode) < 0) {	/* INTL: Native. */
>>  		return TCL_ERROR;
>>  	    }

Re: Patch to avoid using mkfifo under cygwin

[Keith Seitz]

On Fri, 13 Sep 2002, Christopher Faylor wrote:

> Given that this is something that will just magically work when the cygwin
> DLL is updated, I don't think autoconf is the right way to go.

Doh! I overlooked that.
> 
> What's wrong with the current code which returns an error when mkfifo fails?
> That seems pretty clear to me.

Hmmm.. So mkfifo is defined in cygwin, but returns errors? If so, I'd say 
we should just leave it alone.

Mo, is this actually causing problems somewhere?
Keith

Re: Patch to avoid using mkfifo under cygwin

[Mo DeJong]

On Fri, 13 Sep 2002 10:00:03 -0700 (PDT)
Keith Seitz <keiths@redhat.com> wrote:

> On Fri, 13 Sep 2002, Christopher Faylor wrote:

...

> Hmmm.. So mkfifo is defined in cygwin, but returns errors? If so, I'd say 
> we should just leave it alone.
> 
> Mo, is this actually causing problems somewhere?
> Keith

This was causing a failure in the Tcl test suite, but it should not
matter to anyone. Since Cygwin will add the feature soon,
it seems fine to just leave it the way it is.

cheers
Mo

Re: Patch to avoid using mkfifo under cygwin

[Keith Seitz]

On Fri, 13 Sep 2002, Mo DeJong wrote:

> This was causing a failure in the Tcl test suite, but it should not
> matter to anyone. Since Cygwin will add the feature soon,
> it seems fine to just leave it the way it is.

Ah, okay. Thanks for the clarification.

Keith

Re: Patch to avoid using mkfifo under cygwin

[Christopher Faylor]

On Fri, Sep 13, 2002 at 11:07:38AM -0700, Keith Seitz wrote:
>On Fri, 13 Sep 2002, Mo DeJong wrote:
>>This was causing a failure in the Tcl test suite, but it should not
>>matter to anyone.  Since Cygwin will add the feature soon, it seems
>>fine to just leave it the way it is.
>
>Ah, okay.  Thanks for the clarification.

Ditto.  Thanks, Mo.  This should be available soon when we release
"cygserver".

cgf

Re: Patch to avoid using mkfifo under Cygwin

[Andrew Cagney]

> On Thu, 12 Sep 2002, Mo DeJong wrote:
> 
> 
>> Here is a patch that avoids calling mkfifo under Cygwin since
>> it is not functional.
> 
> 
> Ok, so given cgf's latest comments, do we want to hold off on this or 
> revert it when the cygwin work is ready? Maybe we can write an autoconf 
> test for it?

What about the 5.3 branch?

Andrew


>> 2002-09-12  Mo DeJong  <supermo@bayarea.net>
>> 
>> 	* unix/tclUnixFCmd.c (DoCopyFile): Don't use mkfifo
>> 	when compiling with Cygwin, since it is not supported.
>> 
>> Index: unix/tclUnixFCmd.c
>> ===================================================================
>> RCS file: /cvs/src/src/tcl/unix/tclUnixFCmd.c,v
>> retrieving revision 1.2
>> diff -u -r1.2 tclUnixFCmd.c
>> --- unix/tclUnixFCmd.c	9 Sep 2001 23:56:09 -0000	1.2
>> +++ unix/tclUnixFCmd.c	12 Sep 2002 22:37:10 -0000
>> @@ -391,12 +391,18 @@
>>  	    }
>>  	    return CopyFileAtts(src, dst, &srcStatBuf);
>>  	}
>> +#ifndef __CYGWIN__
>> +        /*
>> +         * mkfifo is not supported under Cygwin even though it is prototyped
>> +         * in newlib headers.
>> +         */
>>          case S_IFIFO: {
>>  	    if (mkfifo(dst, srcStatBuf.st_mode) < 0) {	/* INTL: Native. */
>>  		return TCL_ERROR;
>>  	    }
>>  	    return CopyFileAtts(src, dst, &srcStatBuf);
>>  	}
>> +#endif /* __CYGWIN__ */
>>          default: {
>>  	    return CopyFile(src, dst, &srcStatBuf);
>>  	}
>> 
> 
> 
>