[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: dirty but most simple method by which pkg install will be successful
From: |
Tatsuro MATSUOKA |
Subject: |
Re: dirty but most simple method by which pkg install will be successful in mingw octave |
Date: |
Fri, 9 Nov 2007 06:01:33 +0900 (JST) |
Dear Michael
> Actually, backslash / forward slash problem is already taken into
> account in pkg.m: I also use MSYS to compile octave-forge packages,
> and basically pkg.m expects MSYS. The code of pkg.m has been
> modified such that it executes its command through a shell (see
> "shell" subfunction): in that function, any backslash is converted into
> forward slash. So "configure" and "make" work normally. This is not
> the problem here.
>
> I compared my verbose output when compiling a package and I noticed
> a difference in the "copyfile" call: the paths used only contain backslashes,
> no forward slash:
>
> copyfile
> C:\DOCUME~1\goffioul\LOCALS~1\Temp\oct-15\sockets-1.0.2\src\sockets.oct
> C:\DOCUME~1\goffioul\LOCALS~1\Temp\oct-15\sockets
> -1.0.2\inst\i686-pc-msdosmsvc-api-v26
>
> What I suspect is that "cp.exe" (which I also use when compiling
> octave-forge packages, and which is used internally by "copyfile")
> cannot deal with mixed slashes in paths. It works for me, because it
> only contains backslashes. The temporary dir used during compilation
> is obtained from "tmpnam". For me, it returns a path with backslashes
> only. What does it return for you?
Thank you for your reply.
I noticed here my verbose log after my modification pkg.m.
Without it,
octave.exe:4> pkg install -local -verbose io-1.0.3.tar.gz
cp: missing destination file
Try `cp --help' for more information.
mkdir (C:/DOCUME~1/Tatsu/LOCALS~1/Temp\oct-2)
untar (io-1.0.3.tar.gz, C:/DOCUME~1/Tatsu/LOCALS~1/Temp\oct-2)
make: Entering directory `/tmp/oct-2/io-1.0.3/src'
mkoctfile dlmread.cc
mkoctfile textread.cc
make: Leaving directory `/tmp/oct-2/io-1.0.3/src'
warning: package io is empty
The pkg stopped at preparation stage.
You said that the current pkg.m consider msys.
But for my system, it does not seem to work at all.
So I have been preparing patch for pkg.m.
What do you think about this ?
Is this the system problem ?
Tatsuro
--------------------------------------
New Design Yahoo! JAPAN 2008/01/01
http://pr.mail.yahoo.co.jp/newdesign/
- dirty but most simple method by which pkg install will be successful in mingw octave, Tatsuro MATSUOKA, 2007/11/06
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, David Bateman, 2007/11/07
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, Benjamin Lindner, 2007/11/07
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, John W. Eaton, 2007/11/07
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, David Bateman, 2007/11/07
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, Benjamin Lindner, 2007/11/07
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, Tatsuro MATSUOKA, 2007/11/07
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, Michael Goffioul, 2007/11/08
- Re: dirty but most simple method by which pkg install will be successful in mingw octave,
Tatsuro MATSUOKA <=
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, Michael Goffioul, 2007/11/09
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, Tatsuro MATSUOKA, 2007/11/09
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, Tatsuro MATSUOKA, 2007/11/07
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, John W. Eaton, 2007/11/07
- Re: dirty but most simple method by which pkg install will be successful in mingw octave, Tatsuro MATSUOKA, 2007/11/07