[MITgcm-support] genmake2: going with the flow

Tue Sep 23 10:57:52 EDT 2003

Hello Ed,

sorry, that I made so much noise. I think, our emails are now getting 
ahead of each other, but I have figured out how to fix linux_ifc (see 
my last email-posting).

More generally, I think that something like
$TOOLSDIR/genmake2 -mods=../code
should still work, that genmake2 should be able to figure out some 
default configuration, which it couldn't in my case. But maybe that's 
not your policy? I won't complain.

The nice thing about the .genmakerc is/was, that it made it possible to 
run genmake with the same options without specifying the options on the 
command line (source of error for me), but again, if that's a matter of 
policy, I won't complain.

The usage of genmake2 is probably not less intuitive than the usage of 
genmake, I am just used to the latter. I was refering to the code of 
genmake2, but I am not a bash-person (do you believe in reincarnation 
(o: ). genmake I could understand, and thus successfully manipulate, 
genmake2 is beyond my capabilities, but in the end, you don't want 
people to edit that file anyway, right? So, I won't complain.

I will now write a build_options file for Mac OS X (o:

Martin

On Tuesday, September 23, 2003, at 04:13 PM, Ed Hill wrote:

> On Tue, 2003-09-23 at 03:26, Martin Losch wrote:
>> Hi,
>>
>> I am trying to keep up with your pace and want to learn how to use
>> genmake2, or is that too early? I am having all sorts of problems:
>> 1. when I specify -optfile=$TOOLSDIR/build_options/linux_ifc I still
>> have to specify -fc=ifc to make things work. Is that intended?
>
> Hi Martin,
>
> Thank you for trying genmake2!  We have it working with a number of
> compilers/platforms, so its probably not too early for others to try
> it.  I'll try to answer the specific problems below:
>
>
>> 2. with the command
>>>> ../../tools/genmake2 -optfile=../../tools/build_options/linux_ifc
>> -mods=../code -fc=ifc
>> I get makefile
>>>> make depend
>> ... (skipping the usual stuff)
>> ../../tools/f90mkdepend >> Makefile
>> echo: No match
>>>> make
>> ... compile, compile, compile ...
>> ... timer_stats.o -lPEPCF90
>> make: o: Command not found
>> make: [mitgcmuv] Error 127 (ignored)
>>>>
>> There's no executable.
>
> This error indicates that the name of either the compiler or linker is
> missing and should be specified in the "optfile".
>
>
>> 3. Apparently there no more .genmakerc. That's too bad, will that
>> possibility be put back into genmake2?
>
> OK, I'll add this to the documentation.  The new "optfiles" are 
> intended
> to replace the old .genmakerc files.  And the optfiles given in
> $ROOTDIR/tools/build_options/ are intended primarily as "good starting
> points", not necessarily the best for your particular platform.  Some 
> of
> them have been tested and work on some machines and some are still in
> need of work.  The trick is to take whatever options you were using in
> your old .genmakerc file and convert them into one or more of the
> new-style optfiles.
>
> I'll bet that the options file at:
>
>   MITgcm/tools/build_options/linux_ifc_ieee
>
> is probably closer to what you want.
>
>
>> 4. Do you already have documentation on how to use genmake2. The -help
>> option is not that help full. And simple options don't work, e.g.:
>>>> ../../tools/genmake2 -mods=../code
>> ....
>>>>  make depend
>> make[1]: Entering directory `/home/lincl/mlosch/head/channel/build2'
>> ../../pkg/debug/chksum_tiled.F chksum_tiled.F
>> make[1]: execvp: ../../pkg/debug/chksum_tiled.F: Permission denied
>> make[1]: *** [chksum_tiled.F] Error 127
>> make[1]: Leaving directory `/home/lincl/mlosch/head/channel/build2'
>> make: *** [depend] Error 2
>
> Documentation is available at:
>
>   http://mitgcm.org/dev_docs/devel_HOWTO/
>
> and I'll be adding to it in the coming week--especially based on your
> comments here!
>
>
>> If found it hard to understand genmake, but genmake2 is absolutely
>> impossible for me to comprehend, so I cannot track down these things
>> myself, or include, eg. a .genmakerc switch.
>
> I'm very sorry that you find it counter-intuitive.  I'll do everything 
> I
> can to make the old .genmakerc to "optfiles" transition more
> understandable in the documentation.
>
> thanks,
> Ed
>
>
>> PS. my platform, etc.:
>>>> uname -a
>> Linux lincl0 2.4.20-18.9bigmem #1 SMP Thu May 29 06:41:01 EDT 2003 
>> i686
>> i686 i386 GNU/Linux
>>>> g77 -v
>> Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/3.2.2/specs
>> Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
>> --infodir=/usr/share/info --enable-shared --enable-threads=posix
>> --disable-checking --with-system-zlib --enable-__cxa_atexit
>> --host=i386-redhat-linux
>> Thread model: posix
>> gcc version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)
>>>> ifc -V
>> Intel(R) Fortran Compiler for 32-bit applications, Version 7.1   Build
>> 20030813Z
>> Copyright (C) 1985-2003 Intel Corporation.  All rights reserved.
>> FOR NON-COMMERCIAL USE ONLY
>>
>> GNU ld version 2.13.90.0.18 20030206
>>    Supported emulations:
>>     elf_i386
>>     i386linux
>>
>> _______________________________________________
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org
>> http://dev.mitgcm.org/mailman/listinfo/mitgcm-support
> -- 
> Edward H. Hill III, PhD
> office:  MIT Dept. of EAPS;  Room 54-1424;  77 Massachusetts Ave.
>             Cambridge, MA 02139-4307
> email:   eh3 at mit.edu,  ed at eh3.com
> URL:     http://web.mit.edu/eh3/
> phone:   617-253-0098
> fax:     617-253-4464
> <signature.asc>_______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-support