Date: 14 Jul 1995 02:31:50 +0200 From: knarf@nasim.cube.net (Frank Bartels) To: freebsd-ports@freebsd.org Subject: Re: diffs and checksums Message-ID: <3u4dtm$rt5@nasim.nasim.cube.net> References: <3u446o$oih@nasim.nasim.cube.net> <1223.805679516@palmer.demon.co.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
Gary Palmer (gary@palmer.demon.co.uk) wrote:
> In message <3u446o$oih@nasim.nasim.cube.net>, Frank Bartels writes:
> The type of diff supplied in a port is extremely dependant on the
> person doing the port
Oh, I did not know this, I always thought only context diffs are
accepted.
> - I prefer context diffs, and tend to use them,
> even if I am working on an existing port which uses a different diff
> format - I find context diffs a lot easier for ports as you get a
> better idea of what is actually being changed as you have the context
> of the patch...
Huh? With unified diffs you also have the context of the patch. And I
get a *much* better idea of what has changed, because the original and
the modified lines follow immediateley and not many lines later. Also
with the `+' and `-' lines I always see what is added and removed.
With the `!' I always ask me if this is the original or modified line.
Small example (less-290):
--- Makefile.in.orig Mon Mar 13 22:41:54 1995
+++ Makefile.in Fri Jul 14 02:17:15 1995
@@ -15,7 +15,7 @@
LDFLAGS = @LDFLAGS@
O=o
-LIBS = @LIBS@
+LIBS = -ltermcap
prefix = @prefix@
exec_prefix = @exec_prefix@
The context is present and please don't tell me you don't see on the
first view what is changed. ;) BTW: Unified diffs are much smaller as
they don't contain the context twice (this is why the format is called
"unified context diff" - see patch(1)).
> Seeing as the plain tar file is never seen by the ports mechanism, you
> are suggesting we add another step? Currently we go straight from the
> tarball to the extracted source, so there is no real way to check the
> md5 of the tar file.
zcat during the checksum check would do the trick.
> Also, if the file starts as a .Z, you do a znew, surely you get a file
> with a .gz extension? That shouldn't produce a checksum mismatch, you
> should be told that it can't find a checksum to match against that
> file (I can't remember what it says, but it's not a checksum
> mismatch!)
Sure, I either have to modify the Makefile or link the .gz -> .Z.
Bye,
Knarf
--
Frank Bartels | UUCP/ZModem/Fax: + 49 89 5469593 | MiNT is
knarf@nasim.cube.net | Login: nuucp Index: /pub/ls-lR.nasim.gz | Now TOS!
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3u4dtm$rt5>