From knarf at cct.lsu.edu  Mon Feb  9 12:54:06 2009
From: knarf at cct.lsu.edu (Frank Loeffler)
Date: Mon, 9 Feb 2009 12:54:06 -0600
Subject: [Developers] absolute paths in THORNLIST
Message-ID: <20090209185406.GR3179@numrel07.cct.lsu.edu>

Hi,

the attaches patch allows for absolute paths in THORNLIST. It requires
'grep' to be available.

Frank


Index: Makefile
===================================================================
RCS file: /cactusdevcvs/Cactus/Makefile,v
retrieving revision 1.178
diff -u -r1.178 Makefile
--- Makefile	26 Jan 2009 23:27:22 -0000	1.178
+++ Makefile	9 Feb 2009 18:44:22 -0000
@@ -64,9 +64,14 @@
 export TPARFLAGS FPARFLAGS
 
 # Directory for configuration options
+
+# set $(THORNLIST_DIR) to "." if it is not set already and if
+# $(THORNLIST) does not start with a slash
 ifeq ($(strip $(THORNLIST_DIR)), )
+ifeq ($(shell echo $(strip $(THORNLIST)) | grep '^/'),)
 THORNLIST_DIR = "."
 endif
+endif
 
 # End of parallel make stuff
 


--- StripMime Report -- processed MIME parts ---
multipart/mixed
  text/plain (text body -- kept)
  text/plain (text body -- kept)
---

From schnetter at cct.lsu.edu  Mon Feb  9 13:14:52 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Mon, 9 Feb 2009 13:14:52 -0600
Subject: [Developers] [Patches] absolute paths in THORNLIST
In-Reply-To: <20090209185406.GR3179@numrel07.cct.lsu.edu>
References: <20090209185406.GR3179@numrel07.cct.lsu.edu>
Message-ID: <9F5E0A4F-6569-434E-9471-C78BB04FA8D7@cct.lsu.edu>

On Feb 9, 2009, at 12:54:06, Frank Loeffler wrote:

> Hi,
>
> the attaches patch allows for absolute paths in THORNLIST. It requires
> 'grep' to be available.


grep is not a problem; it's used elsewhere as well.

Question: Is the strip around the THORNLIST necessary?  Doesn't the  
shell strip it off when it parses the arguments for the echo?

(There are other cases in the Makefile as well where strip and echo  
are used together.)

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 194 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/developers/attachments/20090209/46a03258/attachment-0001.bin 

From knarf at cct.lsu.edu  Mon Feb  9 13:55:45 2009
From: knarf at cct.lsu.edu (Frank Loeffler)
Date: Mon, 9 Feb 2009 13:55:45 -0600
Subject: [Developers] [Patches] absolute paths in THORNLIST
In-Reply-To: <9F5E0A4F-6569-434E-9471-C78BB04FA8D7@cct.lsu.edu>
References: <20090209185406.GR3179@numrel07.cct.lsu.edu>
	<9F5E0A4F-6569-434E-9471-C78BB04FA8D7@cct.lsu.edu>
Message-ID: <20090209195544.GT3179@numrel07.cct.lsu.edu>

Hi,

On Mon, Feb 09, 2009 at 01:14:52PM -0600, Erik Schnetter wrote:
> grep is not a problem; it's used elsewhere as well.

Good.

> Question: Is the strip around the THORNLIST necessary?  Doesn't the shell 
> strip it off when it parses the arguments for the echo?

It can be on my linux system. I included it because of the other cases of
strip and echo, thinking that maybe this might not the case always.

If you think it can always be removed, please do so.

Frank


From knarf at cct.lsu.edu  Mon Feb  9 17:03:35 2009
From: knarf at cct.lsu.edu (Frank Loeffler)
Date: Mon, 9 Feb 2009 17:03:35 -0600
Subject: [Developers] doc: latin9 -> utf8
Message-ID: <20090209230335.GC3179@numrel07.cct.lsu.edu>

Hi,

this patch changes the input encoding for the documentation from latin9
to uft8. I could not find problems when building the documentation with
this. It enables the use of uft8 textfiles, which is (or is becoming)
the standard nowadays.

Ok to apply?

Frank


Index: doc/latex/cactus.sty
===================================================================
RCS file: /cactusdevcvs/Cactus/doc/latex/cactus.sty,v
retrieving revision 1.25
diff -u -r1.25 cactus.sty
--- doc/latex/cactus.sty	30 May 2006 13:13:22 -0000	1.25
+++ doc/latex/cactus.sty	9 Feb 2009 23:01:54 -0000
@@ -42,7 +42,7 @@
 
 % Load in other packages.
 
-\RequirePackage[latin9]{inputenc}
+\RequirePackage[utf8]{inputenc}
 \RequirePackage{fancyhdr}
 \RequirePackage{graphicx}
 \RequirePackage{latexsym}


--- StripMime Report -- processed MIME parts ---
multipart/mixed
  text/plain (text body -- kept)
  text/plain (text body -- kept)
---

From tradke at aei.mpg.de  Tue Feb 10 04:47:00 2009
From: tradke at aei.mpg.de (Thomas Radke)
Date: Tue, 10 Feb 2009 11:47:00 +0100
Subject: [Developers] doc: latin9 -> utf8
In-Reply-To: <20090209230335.GC3179@numrel07.cct.lsu.edu>
References: <20090209230335.GC3179@numrel07.cct.lsu.edu>
Message-ID: <49915B24.6090806@aei.mpg.de>

Frank Loeffler wrote:
> Hi,
> 
> this patch changes the input encoding for the documentation from latin9
> to uft8. I could not find problems when building the documentation with
> this. It enables the use of uft8 textfiles, which is (or is becoming)
> the standard nowadays.
> 
> Ok to apply?

I've tested it on Damiana with the Beta16 thornlist 
(http://www.cactuscode.org/Releases/beta16.txt), and had no problems 
building the UsersGuide and the ThornGuide.

-- 
Cheers, Thomas.

From tradke at aei.mpg.de  Tue Feb 10 11:02:14 2009
From: tradke at aei.mpg.de (Thomas Radke)
Date: Tue, 10 Feb 2009 18:02:14 +0100
Subject: [Developers] [Patches] absolute paths in THORNLIST
In-Reply-To: <20090209195544.GT3179@numrel07.cct.lsu.edu>
References: <20090209185406.GR3179@numrel07.cct.lsu.edu>	<9F5E0A4F-6569-434E-9471-C78BB04FA8D7@cct.lsu.edu>
	<20090209195544.GT3179@numrel07.cct.lsu.edu>
Message-ID: <4991B316.1090702@aei.mpg.de>

Frank Loeffler wrote:
> Hi,
> 
> On Mon, Feb 09, 2009 at 01:14:52PM -0600, Erik Schnetter wrote:
>> grep is not a problem; it's used elsewhere as well.
> 
> Good.
> 
>> Question: Is the strip around the THORNLIST necessary?  Doesn't the shell 
>> strip it off when it parses the arguments for the echo?
> 
> It can be on my linux system. I included it because of the other cases of
> strip and echo, thinking that maybe this might not the case always.
> 
> If you think it can always be removed, please do so.

Wouldn't something like the following be sufficient ?

# Directory for configuration options
ifeq ($(strip $(THORNLIST_DIR)), )
THORNLIST_DIR = $(dir $(THORNLIST))
endif

-- 
Cheers, Thomas.

From knarf at cct.lsu.edu  Tue Feb 10 12:24:00 2009
From: knarf at cct.lsu.edu (Frank Loeffler)
Date: Tue, 10 Feb 2009 12:24:00 -0600
Subject: [Developers] [Patches] absolute paths in THORNLIST
In-Reply-To: <4991B316.1090702@aei.mpg.de>
References: <20090209185406.GR3179@numrel07.cct.lsu.edu>
	<9F5E0A4F-6569-434E-9471-C78BB04FA8D7@cct.lsu.edu>
	<20090209195544.GT3179@numrel07.cct.lsu.edu>
	<4991B316.1090702@aei.mpg.de>
Message-ID: <20090210182359.GF22451@numrel07.cct.lsu.edu>

Hi,

On Tue, Feb 10, 2009 at 06:02:14PM +0100, Thomas Radke wrote:
> Wouldn't something like the following be sufficient ?
> 
> # Directory for configuration options
> ifeq ($(strip $(THORNLIST_DIR)), )
> THORNLIST_DIR = $(dir $(THORNLIST))
> endif

When specifying just the file: THORNLIST=bla.th, THORNLIST_DIR needs to
be set to "." because later in the file the path is always used as
THORNLIST_DIR/THORNLIST and I wanted to keep the patch small.

It would also fail if you specify a THORNLIST_DIR and a THORNLIST
including a path like

THORNLIST_DIR=/home/knarf/thornlists
THORNLIST=project1/thornlist1.th

Frank


From schnetter at cct.lsu.edu  Tue Feb 10 16:55:10 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Tue, 10 Feb 2009 16:55:10 -0600
Subject: [Developers] [Patches] absolute paths in THORNLIST
In-Reply-To: <20090210182359.GF22451@numrel07.cct.lsu.edu>
References: <20090209185406.GR3179@numrel07.cct.lsu.edu>
	<9F5E0A4F-6569-434E-9471-C78BB04FA8D7@cct.lsu.edu>
	<20090209195544.GT3179@numrel07.cct.lsu.edu>
	<4991B316.1090702@aei.mpg.de>
	<20090210182359.GF22451@numrel07.cct.lsu.edu>
Message-ID: <6CFB78B2-A6E5-4388-A376-2138CB6CB295@cct.lsu.edu>

On Feb 10, 2009, at 12:24:00, Frank Loeffler wrote:

> Hi,
>
> On Tue, Feb 10, 2009 at 06:02:14PM +0100, Thomas Radke wrote:
>> Wouldn't something like the following be sufficient ?
>>
>> # Directory for configuration options
>> ifeq ($(strip $(THORNLIST_DIR)), )
>> THORNLIST_DIR = $(dir $(THORNLIST))
>> endif
>
> When specifying just the file: THORNLIST=bla.th, THORNLIST_DIR needs  
> to
> be set to "." because later in the file the path is always used as
> THORNLIST_DIR/THORNLIST and I wanted to keep the patch small.
>
> It would also fail if you specify a THORNLIST_DIR and a THORNLIST
> including a path like
>
> THORNLIST_DIR=/home/knarf/thornlists
> THORNLIST=project1/thornlist1.th

I applied the original patch with a slightly modified comment.

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 194 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/developers/attachments/20090210/c738bc75/attachment.bin 

From tradke at aei.mpg.de  Thu Feb 12 10:42:16 2009
From: tradke at aei.mpg.de (Thomas Radke)
Date: Thu, 12 Feb 2009 17:42:16 +0100
Subject: [Developers] print process ID in the Cactus banner
Message-ID: <49945168.6000405@aei.mpg.de>

The attached tiny patch outputs the process ID in the Cactus banner 
printed at simulation startup. This information is useful eg. so that 
one can associate core.<pid> files with MPI tasks 
(CCTK_Proc<task-ID>.out files) while post-mortem debugging a parallel run.

-- 
Cheers, Thomas.


--- StripMime Report -- processed MIME parts ---
multipart/mixed
  text/plain (text body -- kept)
  text/x-patch
---

From jtao at cct.lsu.edu  Thu Feb 12 10:49:28 2009
From: jtao at cct.lsu.edu (Jian Tao)
Date: Thu, 12 Feb 2009 10:49:28 -0600
Subject: [Developers] print process ID in the Cactus banner
In-Reply-To: <49945168.6000405@aei.mpg.de>
References: <49945168.6000405@aei.mpg.de>
Message-ID: <49945318.5010005@cct.lsu.edu>

Hi Thomas,

This reminds me some ideas we had long time ago about outputting
  more environmental (both hardware and software) info in some
parser-friendly files, which in turn can be used for documentation,
debugging, and optimization.

Regards,
Jian

Thomas Radke wrote:
> The attached tiny patch outputs the process ID in the Cactus banner 
> printed at simulation startup. This information is useful eg. so that 
> one can associate core.<pid> files with MPI tasks 
> (CCTK_Proc<task-ID>.out files) while post-mortem debugging a parallel run.
> 


From schnetter at cct.lsu.edu  Thu Feb 12 13:25:39 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Thu, 12 Feb 2009 13:25:39 -0600
Subject: [Developers] [Patches] print process ID in the Cactus banner
In-Reply-To: <49945168.6000405@aei.mpg.de>
References: <49945168.6000405@aei.mpg.de>
Message-ID: <DDED4FA5-EC3E-41FC-A4D5-6B5AA139966F@cct.lsu.edu>

On Feb 12, 2009, at 10:42:16, Thomas Radke wrote:

> The attached tiny patch outputs the process ID in the Cactus banner  
> printed at simulation startup. This information is useful eg. so  
> that one can associate core.<pid> files with MPI tasks  
> (CCTK_Proc<task-ID>.out files) while post-mortem debugging a  
> parallel run.
>
> -- 
> Cheers, Thomas.
> Index: Banner.c
> ===================================================================
> RCS file: /cactusdevcvs/Cactus/src/main/Banner.c,v
> retrieving revision 1.35
> diff -u -r1.35 Banner.c
> --- Banner.c	16 Feb 2006 16:05:46 -0000	1.35
> +++ Banner.c	12 Feb 2009 16:34:26 -0000
> @@ -102,7 +102,7 @@
>  Util_CurrentTime (DATALENGTH, buffer);
>  printf (" (%s)\n", buffer);
>  Util_GetHostName (buffer, DATALENGTH);
> -  printf ("Run host:          %s\n", buffer);
> +  printf ("Run host:          %s (pid=%d)\n", buffer, (int)  
> getpid());
>  if (getcwd (buffer, DATALENGTH))
>  {
>    printf ("Working directory: %s\n", buffer);


I agree.

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 194 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/developers/attachments/20090212/fad02cf6/attachment.bin 

From schnetter at cct.lsu.edu  Thu Feb 12 20:31:03 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Thu, 12 Feb 2009 20:31:03 -0600
Subject: [Developers] [Patches] doc: latin9 -> utf8
In-Reply-To: <20090209230335.GC3179@numrel07.cct.lsu.edu>
References: <20090209230335.GC3179@numrel07.cct.lsu.edu>
Message-ID: <EA292EC5-B99C-41AD-95E1-F8CAA01E8E06@cct.lsu.edu>

On Feb 9, 2009, at 17:03:35, Frank Loeffler wrote:

> Hi,
>
> this patch changes the input encoding for the documentation from  
> latin9
> to uft8. I could not find problems when building the documentation  
> with
> this. It enables the use of uft8 textfiles, which is (or is becoming)
> the standard nowadays.
>
> Ok to apply?


Yes, ok to apply.

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 194 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/developers/attachments/20090212/b5460944/attachment.bin 

From tradke at aei.mpg.de  Fri Feb 13 04:37:52 2009
From: tradke at aei.mpg.de (Thomas Radke)
Date: Fri, 13 Feb 2009 11:37:52 +0100
Subject: [Developers] print process ID in the Cactus banner
In-Reply-To: <49945318.5010005@cct.lsu.edu>
References: <49945168.6000405@aei.mpg.de> <49945318.5010005@cct.lsu.edu>
Message-ID: <49954D80.8080000@aei.mpg.de>

Jian Tao wrote:
> Hi Thomas,
> 
> This reminds me some ideas we had long time ago about outputting
>   more environmental (both hardware and software) info in some
> parser-friendly files, which in turn can be used for documentation,
> debugging, and optimization.

Hi Jian,

sorry, I can't recall such a discussion, it must have been too long ago.

There is AEIThorns/Formaline which can collect certain metadata about a 
run, and save it in a file or send it off to an external information 
service. Did you have this type of functionality in mind ?

-- 
Cheers, Thomas.

From knarf at cct.lsu.edu  Fri Feb 13 09:59:59 2009
From: knarf at cct.lsu.edu (Frank Loeffler)
Date: Fri, 13 Feb 2009 09:59:59 -0600
Subject: [Developers] disable strict ansi support for c99 and the intel C
	compiler > 8
Message-ID: <20090213155959.GN22451@numrel07.cct.lsu.edu>

Hi,

this patch changes '-std=gnu99' to '-std=c99 -U__STRICT_ANSI__' for the
intel C compiler versions > 8.

'gnu99' was never officially supported and does not work as expected
starting with version 11. The official way to achieve c99 extensions
like M_PI is to undefine __STRICT_ANSI__ (which is turned on by
'-std=c99) [1]. This should then be equivalent to the -std=gnu99
option of the gnu c compiler.

This change should not be necessary for version 9, because this version
does not define __STRICT_ANSI__ for 'c99', but including version 9 here
keeps the patch smaller and the known-architecture file smaller and more
readable and it should work just as well.

[1]
http://www.intel.com/software/products/compilers/docs/clin/Release_Notes.htm :

Behavior Change for -std=c99 and -std=c89

When the -std=c99 or -std=c89 options are used, the 10.1 compiler's
behavior is now compatible with gcc and the __STRICT_ANSI__ macro is
predefined. If your application compiles with errors and you wish to use
the previous version's behavior, add -U__STRICT_ANSI__ to the command
line. 

Frank


Index: lib/make/known-architectures/linux
===================================================================
RCS file: /cactusdevcvs/Cactus/lib/make/known-architectures/linux,v
retrieving revision 1.132
diff -u -r1.132 linux
--- lib/make/known-architectures/linux	26 Jan 2009 20:53:27 -0000	1.132
+++ lib/make/known-architectures/linux	13 Feb 2009 15:49:32 -0000
@@ -380,7 +380,7 @@
         icc_opt=''
       fi
       if test $icc_version -ge 8; then
-        icc_c99='-std=gnu99'
+        icc_c99='-std=c99 -U__STRICT_ANSI__'
       else
         icc_c99='-c99'
       fi


--- StripMime Report -- processed MIME parts ---
multipart/mixed
  text/plain (text body -- kept)
  text/plain (text body -- kept)
---

From tradke at aei.mpg.de  Fri Feb 13 10:43:20 2009
From: tradke at aei.mpg.de (Thomas Radke)
Date: Fri, 13 Feb 2009 17:43:20 +0100
Subject: [Developers] disable strict ansi support for c99 and the intel
 C	compiler > 8
In-Reply-To: <20090213155959.GN22451@numrel07.cct.lsu.edu>
References: <20090213155959.GN22451@numrel07.cct.lsu.edu>
Message-ID: <4995A328.1010603@aei.mpg.de>

Frank Loeffler wrote:
> Hi,
> 
> this patch changes '-std=gnu99' to '-std=c99 -U__STRICT_ANSI__' for the
> intel C compiler versions > 8.

It works me with Intel 10 and 11. Ok to apply.

-- 
Cheers, Thomas.