Bug#345557: source: debian/patches applies not clean.
On Wed, Jan 04, 2006 at 08:16:25AM +0100, Christian Perrier wrote: -configure: patch-stamp configure-stamp +configure: patch configure-stamp configure-stamp: dh_testdir I haven't worked with quilt enough yet to know the answer to this: is there not an advantage to having a patch-stamp with quilt? I.e., does quilt store enough state that we don't need to worry about additional calls to the patch target wasting time? (patch to samba package switching to quilt) I have to admit that I'm not completely sure. Here I applied what we used in shadow.Let's ask Martin who made the switch of the shadow package to quilt and who maintains the quilt packages. A patch-stamp will give you a little gain, but really little because quilt maintains a whole bunch of status. It knows where it stands in the patch set at every time. Try to travel into the .pc directory if you want to visit it. So, a patch-stamp would save you the time for quilt to get the patch list from debian/patches/serie and the list of applied patches from somewhere in .pc/ and compare them to notice that they match. But that's really quick, I don't see the point of deploying this extra mecanism. Bye, Mt. -- Dans le passé, il y avait plus de futur que maintenant. -- Le Chat signature.asc Description: Digital signature
Bug#345557: source: debian/patches applies not clean.
A patch-stamp will give you a little gain, but really little because quilt maintains a whole bunch of status. It knows where it stands in the patch set at every time. Try to travel into the .pc directory if you want to visit it. So, Steve, do you think I can commit the change? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#345557: source: debian/patches applies not clean.
On Thu, Jan 05, 2006 at 06:59:05AM +0100, Christian Perrier wrote: A patch-stamp will give you a little gain, but really little because quilt maintains a whole bunch of status. It knows where it stands in the patch set at every time. Try to travel into the .pc directory if you want to visit it. So, Steve, do you think I can commit the change? I think so, yes. :) Cheers, -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ signature.asc Description: Digital signature
Bug#345557: source: debian/patches applies not clean.
Indeed, I was just working on the switch to quilt...:-) Here's a first patch. No more comments? So, will someone scream out if I happen to apply this patch? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#345557: source: debian/patches applies not clean.
On Mon, Jan 02, 2006 at 10:06:20AM +0100, Christian Perrier wrote: diff -Nru debian.old/rules debian/rules --- debian.old/rules 2006-01-02 09:59:47.312854758 +0100 +++ debian/rules 2006-01-02 09:29:53.215082689 +0100 @@ -39,6 +39,7 @@ INSTALL += -s endif +include /usr/share/quilt/quilt.make DESTDIR=`pwd`/debian/tmp @@ -86,18 +87,7 @@ conf_args += --build $(DEB_BUILD_GNU_TYPE) --host $(DEB_HOST_GNU_TYPE) endif -patch: patch-stamp -patch-stamp: - dh_testdir - if [ ! -f patch-stamp ]; then /bin/sh debian/scripts/patch-source; fi - touch patch-stamp - -unpatch: - dh_testdir - if [ -f patch-stamp ]; then /bin/sh debian/scripts/unpatch-source; fi - rm -f patch-stamp - -configure: patch-stamp configure-stamp +configure: patch configure-stamp configure-stamp: dh_testdir I haven't worked with quilt enough yet to know the answer to this: is there not an advantage to having a patch-stamp with quilt? I.e., does quilt store enough state that we don't need to worry about additional calls to the patch target wasting time? Everything else looks reasonable to me. Thanks, -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ signature.asc Description: Digital signature
Bug#345557: source: debian/patches applies not clean.
-configure: patch-stamp configure-stamp +configure: patch configure-stamp configure-stamp: dh_testdir I haven't worked with quilt enough yet to know the answer to this: is there not an advantage to having a patch-stamp with quilt? I.e., does quilt store enough state that we don't need to worry about additional calls to the patch target wasting time? (patch to samba package switching to quilt) I have to admit that I'm not completely sure. Here I applied what we used in shadow.Let's ask Martin who made the switch of the shadow package to quilt and who maintains the quilt packages. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#345557: source: debian/patches applies not clean.
Switching to dpatch would give me hives. OTOH, switching to quilt might be worth considering, precisely because it gives a better toolset for fuzzies and patch dependencies. I fully agree here. I use quilt in shadow and geneweb and it makes life *very* comfortable (though I don't actually use the patch dependencies). Now we need to find out *who* can do this..:) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#345557: source: debian/patches applies not clean.
Am Montag, den 02.01.2006, 08:53 +0100 schrieb Christian Perrier: Switching to dpatch would give me hives. OTOH, switching to quilt might be worth considering, precisely because it gives a better toolset for fuzzies and patch dependencies. I fully agree here. Now we need to find out *who* can do this..:) Yesterday while building 3.0.21a I wanted to switch to dpatch, too. I will work on this point/bug:). -- Noèl Köthe noel debian.org Debian GNU/Linux, www.debian.org signature.asc Description: This is a digitally signed message part
Bug#345557: source: debian/patches applies not clean.
There's a small error in my patch : --- rules.old 2006-01-02 12:23:45.436501110 +0100 +++ rules 2006-01-02 11:57:26.028735838 +0100 @@ -99,7 +99,7 @@ touch configure-stamp -build: patch-stamp configure-stamp build-stamp +build: configure-stamp build-stamp build-stamp: dh_testdir
Bug#345557: source: debian/patches applies not clean.
Package: samba Version: 3.0.21a-1 Severity: normal Tags: patch Checked the new Debian version. If one have a look at the build-file you will see many hunks and fuzzes applying the patches. Please find attached patch made of completely reworked single patches. Note: The number of original patches doesn't change. Working on that, I found a very complicated patch in patch system. So switching to dpatch could make maintaining the patches more efficient, Isn't it? Elimar -- Excellent day for drinking heavily. Spike the office water cooler;-) Nur in patches.hunked: README_nosmbldap-tools.patch. diff -ur patches.hunked/VERSION.patch patches/VERSION.patch --- patches.hunked/VERSION.patch2006-01-01 19:33:11.154811000 +0100 +++ patches/VERSION.patch 2006-01-01 19:34:19.792248917 +0100 @@ -1,6 +1,6 @@ --- samba-3.0.0rc2/source/VERSION.orig 2003-09-02 21:56:11.0 -0400 +++ samba-3.0.0rc2/source/VERSION 2003-09-02 21:56:30.0 -0400 -@@ -78,4 +78,4 @@ +@@ -90,4 +90,4 @@ # e.g. SAMBA_VERSION_VENDOR_SUFFIX=vendor_version()# # - CVS 3.0.0rc2-VendorVersion# diff -ur patches.hunked/cups.patch patches/cups.patch --- patches.hunked/cups.patch 2006-01-01 19:33:11.158811000 +0100 +++ patches/cups.patch 2006-01-01 19:25:08.564610884 +0100 @@ -1,6 +1,6 @@ --- samba-3.0.10.orig/source/include/includes.h +++ samba-3.0.10/source/include/includes.h -@@ -940,10 +940,7 @@ +@@ -1049,10 +1049,7 @@ #endif #ifndef DEFAULT_PRINTING diff -ur patches.hunked/documentation.patch patches/documentation.patch --- patches.hunked/documentation.patch 2006-01-01 19:33:11.158811000 +0100 +++ patches/documentation.patch 2006-01-01 19:25:08.564610884 +0100 @@ -1,6 +1,6 @@ --- samba-3.0.4/docs/manpages/swat.8.orig 2004-05-05 10:22:50.0 -0400 +++ samba-3.0.4/docs/manpages/swat.8 2004-05-10 13:25:53.0 -0400 -@@ -80,6 +80,13 @@ +@@ -84,6 +84,13 @@ .SH INSTALLATION .PP @@ -14,7 +14,7 @@ Swat is included as binary package with most distributions\. The package manager in this case takes care of the installation and configuration\. This section is only for those who have compiled swat from scratch\. .PP -@@ -87,13 +94,13 @@ +@@ -91,13 +98,13 @@ .TP 3 \(bu @@ -31,7 +31,7 @@ .LP .SS Inetd Installation -@@ -102,7 +109,7 @@ +@@ -106,7 +113,7 @@ You need to edit your \fI/etc/inetd\.conf \fR and \fI/etc/services\fR to enable SWAT to be launched via \fBinetd\fR\. .PP @@ -40,7 +40,7 @@ .PP \fBswat 901/tcp\fR -@@ -114,10 +121,10 @@ +@@ -118,10 +125,10 @@ the choice of port number isn't really important except that it should be less than 1024 and not currently used (using a number above 1024 presents an obscure security hole depending on the implementation details of your\fBinetd\fR daemon)\. .PP @@ -53,7 +53,7 @@ .PP Once you have edited \fI/etc/services\fR and \fI/etc/inetd\.conf\fR you need to send a HUP signal to inetd\. To do this use \fBkill \-1 PID \fR where PID is the process ID of the inetd daemon\. -@@ -141,8 +148,8 @@ +@@ -145,8 +152,8 @@ This file must contain a mapping of service name (e\.g\., swat) to service port (e\.g\., 901) and protocol type (e\.g\., tcp)\. .TP diff -ur patches.hunked/fhs.patch patches/fhs.patch --- patches.hunked/fhs.patch2006-01-01 19:33:11.162811000 +0100 +++ patches/fhs.patch 2006-01-01 19:25:08.568611201 +0100 @@ -1,7 +1,7 @@ -diff -uNr samba-3.0.10.orig/source/Makefile.in samba-3.0.10/source/Makefile.in samba-3.0.10.orig/source/Makefile.in 2004-12-17 03:50:08.0 -0800 -+++ samba-3.0.10/source/Makefile.in2004-12-17 03:55:29.0 -0800 -@@ -90,6 +90,13 @@ +diff -Naur samba-3.0.21.orig/source/Makefile.in samba-3.0.21/source/Makefile.in +--- samba-3.0.21.orig/source/Makefile.in 2005-12-24 14:03:21.777015449 +0100 samba-3.0.21/source/Makefile.in2005-12-24 14:01:25.935827183 +0100 +@@ -91,6 +91,13 @@ # the directory where lock files go LOCKDIR = @lockdir@ @@ -14,8 +14,8 @@ + # the directory where pid files go PIDDIR = @piddir@ - # man pages language(s) -@@ -114,7 +121,7 @@ + +@@ -120,7 +127,7 @@ PATH_FLAGS4 = $(PATH_FLAGS3) -DSWATDIR=\$(SWATDIR)\ -DLOCKDIR=\$(LOCKDIR)\ -DPIDDIR=\$(PIDDIR)\ PATH_FLAGS5 = $(PATH_FLAGS4) -DLIBDIR=\$(LIBDIR)\ \ -DLOGFILEBASE=\$(LOGFILEBASE)\ -DSHLIBEXT=\@[EMAIL PROTECTED] @@ -24,7 +24,7 @@ PATH_FLAGS = $(PATH_FLAGS6) $(PASSWD_FLAGS) # Note that all executable programs now provide for an optional executable suffix. -@@ -1319,7 +1326,7 @@ +@@ -1431,7 +1438,7 @@ @$(SHELL) $(srcdir)/script/installscripts.sh $(INSTALLPERMS) $(DESTDIR)$(BINDIR) $(SCRIPTS) installdat: installdirs @@ -33,9 +33,9 @@ installmsg: installdirs @$(SHELL) $(srcdir)/script/installmsg.sh $(DESTDIR)$(LIBDIR) $(srcdir) -diff -uNr samba-3.0.10.orig/source/configure.in
Bug#345557: source: debian/patches applies not clean.
severity 345557 minor thanks On Sun, Jan 01, 2006 at 08:26:39PM +0100, Elimar Riesebieter wrote: Package: samba Version: 3.0.21a-1 Severity: normal Tags: patch Checked the new Debian version. If one have a look at the build-file you will see many hunks and fuzzes applying the patches. Please find attached patch made of completely reworked single patches. Note: The number of original patches doesn't change. Working on that, I found a very complicated patch in patch system. So switching to dpatch could make maintaining the patches more efficient, Isn't it? Switching to dpatch would give me hives. OTOH, switching to quilt might be worth considering, precisely because it gives a better toolset for fuzzies and patch dependencies. FWIW, I don't imagine it makes much sense to apply this patch as-is, as opposed to having someone with commit access go through and clean them up directly (one way or another). -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ signature.asc Description: Digital signature