patch problem with new port

[Scot Hetzel]

On 7/19/05, Paul Schmehl <pauls at utdallas.edu> wrote:
> I need to create some patches for some ports I'm working on, but there's a
> problem.
> 
> Here's an example.  I need to change "wish" to "wish8.3":
> 
> This is an excerpt from the original file:
> #!/bin/sh
> # Run wish from users PATH \
> exec wish "$0" "$@"
> 
> # $Id: sguil.tk,v 1.157 2004/12/06 21:18:21 bamm Exp $ #
> 
> ####################################################################
> # Sguil - A tcl/tk interface for network security monitoring       #
> ####################################################################
> 
> This is what I need:
> #!/bin/sh
> # Run wish from users PATH \
> exec wish8.3 "$0" "$@"
> 
> # $Id: sguil.tk,v 1.157 2004/12/06 21:18:21 bamm Exp $ #
> 
> ####################################################################
> # Sguil - A tcl/tk interface for network security monitoring       #
> ####################################################################
> 
> The problem is, even if I use -C1 with diff, I *still* end up with the $Id
> line included in the patchfile, and that's illegal.  It will break the
> patch when the port is committed.
> 
The $Id shouldn't break the port, as the FreeBSD ports collection uses
$FreeBSD: to indicate the revision of a file.

> Is there a way around this problem?  I have several files that all need to
> be patched this way - both wish and tclsh are used in them.
> 

If you are repeatedly changing only one line in several files, you
could use the ${REPLACE_CMD} variable to replace the line.

USE_REINPLACE=	yes
WISH_CMD?=    wish8.3

post-patch:
.for file in file1 file2
    @${REINPLACE_CMD} -e 's:exec wish:exec ${WISH_CMD}:g' ${WRKDIR}/$file
.endfor

You can use this trick even if you are patching one of the files, as
the post-patch occurs after your patch is applied.

Scot
-- 
DISCLAIMER:
No electrons were mamed while sending this message. Only slightly bruised.