Menu
▾
▴
Update on the branch surgery
|
From: <es...@th...> - 2017-10-25 17:16:37
|
I have successfully transplanted the 5.0, 4.6. and 4.4 stable
branches.
Daniel Sebald says:
>First, that oldest branch I couldn't make sense of and I think it actually may
>be bogus. Rather, it might be that those 3.x series commits should actually
>be mixed in with the master branch. We'll come back to this one.
I agree, the 3.7 branch may be unsalvageable. HBB tells me:
>3.7.1 has one check-in to ChangeLog that's tagged but not tarred, and an extra
>pair of quotes around a line in configure.in that tarred but not tagged:
But that's the last thing we need to tackle - rerooting the 4.2 and
4.0 branches comes first.
Here's what HBB and Dan think about the root points:
5.2:
HBB:
> branch-5-2-stable:1.6034.0.2
>revision 1.6034
>date: 2017/05/22 05:05:26; author: sfeam; state: Exp; lines: +5 -4170
>branches: 1.6034.2;
>Tag branchpoint for 5.2; Split off entries prior to 2016-02-07 into
>ChangeLog.5
Dan didn't weigh in on this one. It doesn't need to be rerooted;
the tag set was, apparently, complete and the join was not
displaced.
5.0:
HBB:
> branch-5-0-stable:1.5023.0.2
>revision 1.5023
>date: 2014/08/20 04:42:44; author: sfeam; state: Exp; lines: +6 -0
>branches: 1.5023.2;
>Empirical adjustments to dashlength calculated as a function of linewidth
Dan:
>[8/20/14 1:37 PM] Child: Start of separate branch for version 5 stable
>releases
>git -C branch_reference checkout 66a787634452661f02d7997a2a780165a89cedf0
>---
>MY BEST GUESS AT THE BRANCH POINT IS:
>commit 9c68a0d5f752801b536c03694cceb8e66b1122cc
>Author: Ethan A Merritt <merritt@u.washington.edu>
>Date: Tue Aug 19 21:42:44 2014 -0700
> Empirical adjustments to dashlength calculated as a function of linewidth
Ethan has confirmed that this is the correct branch point for
5-0-stable. So everybody agrees.
4.6:
HBB:
>branch-4-6-stable:1.3943.0.2
>revision 1.3943
>date: 2011/11/22 22:35:30; author: sfeam; state: Exp; lines: +3 -0
>branches: 1.3943.2;
>Try to allow for coordinate offset caused by scrolling (Firefox, Opera,
>chrome?)
Dan:
>[11/21/11 11:26 PM] Child: Bump stable version to 4.6; patchlevel is "alpha"
>git -C branch_reference checkout 88ccf09ba5b9f6042d7f2911dc1be6ae7a9aa308
>---
>MY BEST GUESS AT THE BRANCH POINT IS:
>commit df14294df51035145dc7c5391d603b0d452bcd20
>Author: Ethan A Merritt <merritt@u.washington.edu>
>Date: Tue Nov 22 21:12:06 2011 -0800
> update Bruce Ravel's contact info
>git -C root_reference checkout df14294df51035145dc7c5391d603b0d452bcd20
>diff -ur --exclude=".git" root_reference/ branch_reference/ | grep "^-"| grep
>-v "^---" | wc -l
>diff -ur --exclude=".git" root_reference/ branch_reference/ | grep "^+"| grep
>-v "^+++" | wc -l
>13 + 18
>---
>The branch-to-master diff with this is includes:
>
>diff -ur '--exclude=.git' root_reference/ChangeLog branch_reference/ChangeLog
>--- root_reference/ChangeLog 2017-10-24 21:39:21.267669735 -0500
>+++ branch_reference/ChangeLog 2017-10-24 21:31:56.447665382 -0500
>@@ -1,5 +1,10 @@
> 2011-11-22 Ethan A Merritt <xxxxxx@xxxxxx>
>
>+ * Branchpoint for 4.6
>+ cvs tag -b branch-4-6-stable
>+
>+2011-11-22 Ethan A Merritt <xxxxxx@xxxxxx>
>+
>and the first branch on master after this includes:
>
>diff --git a/ChangeLog b/ChangeLog
>index 90a47dc..d5670d7 100644
>--- a/ChangeLog
>+++ b/ChangeLog
>@@ -1,5 +1,9 @@
> 2011-11-22 Ethan A Merritt <merritt@u.washington.edu>
>
>+ * Branchpoint for 4.6 (cvs tag -b branch-4-6-stable)
>+
>+2011-11-22 Ethan A Merritt <merritt@u.washington.edu>
>+
>
>but the version number is kept at 4.5.
Dan's and HBB's choices are separated by only one commit in the
sequence, which looks like this:
2011-11-22T20:41:15Z :32059 Honor numeric locale when formatting contour lab
2011-11-22T22:35:32Z :32062 Try to allow for coordinate offset caused by scr
2011-11-22T05:26:43Z :32071 Bump stable version to 4.6; patchlevel is "alph
2011-11-23T05:12:06Z :34067 update Bruce Ravel's contact info
2011-11-22T05:08:36Z :34073 *** empty log message ***
2011-11-24T22:59:11Z :34075 LT_SINGLECOLOR flags explicit surface color
I actually wouldn't be surprised if the right root point turned out to
be the one in the middle - /Bump stable version to 4.6/. Perhaps Ethan
can clear up which one of these is right.
Note: Changing this in the conversion script is really trivial. The
relevant command sequence is
# The 4.6 branch:
/Try to allow for coordinate offset caused by/ assign v46root --singleton
/Bump stable version to 4.6;/ assign bump46 --singleton
<v46root>,<bump46> reparent --useorder --rebase
All I have to do to change the root point is edit the
selection expression in the first line.
4.4:
HBB:
> branch-4-4-stable:1.3134.0.2
>revision 1.3134
>date: 2009/05/31 19:20:09; author: sfeam; state: Exp; lines: +5 -0
>branches: 1.3134.2;
>Update
Dan:
>[5/30/09 7:30 PM] Child: Start branch for Release 4.4
>git -C branch_reference checkout e3ff538655e5b5888d9aa60b63398fc62847e46f
>---
>MY BEST GUESS AT THE BRANCH POINT IS:
>commit 65b26b2a31b4de73417d746abe96561b4655c699
>Author: Ethan A Merritt <merritt@u.washington.edu>
>Date: Sat May 30 12:20:09 2009 -0700
> Update
>git -C root_reference checkout 65b26b2a31b4de73417d746abe96561b4655c699
>diff -ur --exclude=".git" root_reference/ branch_reference/ | grep "^-"| grep
>-v "^---" | wc -l
>diff -ur --exclude=".git" root_reference/ branch_reference/ | grep "^+"| grep
>-v "^+++" | wc -l
>1 + 8
>---
>The diff includes the following:
>diff -ur '--exclude=.git' root_reference/ChangeLog branch_reference/ChangeLog
>--- root_reference/ChangeLog 2017-10-24 21:16:50.131656512 -0500
>+++ branch_reference/ChangeLog 2017-10-24 21:00:03.367646659 -0500
>@@ -1,5 +1,10 @@
> 2009-05-31 Ethan A Merritt <xxxxxx@xxxxxx>
>
>+ Start branch for Release 4.4
>+ Set VERSION to 4.4 PATCHLEVEL to alpha
>+
>+2009-05-31 Ethan A Merritt <xxxxxx@xxxxxxx>
>+
>
>and the first checkin after that is called "Branch point for version 4.4".
>Although that message would suggest it's the branchpoint, it actually has
>+ New stuff, i.e. code other than bugfixes or things identified as
>+ being desireable for version 4.4, should be commited in the
>+ development branch only (this one).
>+
>+ Bugfixes should be applied to both the development branch 4.3
>+ (this one) and the new one (branch-4-4-stable). When we release
>+ 4.4, we will bump the version number of the development branch to 4.5.
>which suggests to me that this is the first commit on master after the
>bifurcation.
HBB and Dan agree, and that agrees with what I found.
4.2:
HBB:
> branch-4-2-stable:1.2224.0.2
>revision 1.2224
>date: 2006/10/01 12:11:58; author: broeker; state: Exp; lines: +15 -7
>branches: 1.2224.2;
>Reorganize top-level configuration of automake and autoconf to allow
>version-testing automake.
Dan:
>[10/1/06 10:17 AM] Child: Jump to version 4.2, 1st release candidate.
>git -C branch_reference checkout e14ada2e7b0ae4f80d8abeffb5a27d8587a3bbed
>---
>MY BEST GUESS AT THE BRANCH POINT IS:
>commit e18d4e5da418c2ad5a98a2aff3237088aeee01c8
>Author: Hans-Bernhard Broeker <xxxxxx@xxxxxx>
>Date: Sun Oct 1 17:11:57 2006 +0200
> Jump to version 4.2, 1st release candidate.
>git -C root_reference checkout e18d4e5da418c2ad5a98a2aff3237088aeee01c8
>diff -ur --exclude=".git" branch_reference/ root_reference/ | grep "^-"| grep
>-v "^---" | wc -l
>diff -ur --exclude=".git" branch_reference/ root_reference/ | grep "^+"| grep
>-v "^+++" | wc -l
>2 + 2
>---
>In this case HBB specifically made a note about branching.
I think Dan got this one wrong; /Jump to version 4.2/ looks to me
like branch first commit, not the root point (see the digram).
Matters are slightly complicated by the fact that that content
matches to two different commits.
2006-10-01T15:17:17Z :3688 Jump to version 4.2, 1st release candidate.
2006-10-01T15:11:57Z :22184 Jump to version 4.2, 1st release candidate.
The later one looks like a manual merge onto the mainline, so I
have rooted the earlier one to /Reorganize top-level configuration/.
4.0:
HBB:
> branch-4-0-stable:1.1148.0.2
>revision 1.1148
>date: 2004/07/02 05:03:29; author: sfeam; state: Exp; lines: +4 -0
>branches: 1.1148.2;
>#include gp_time.h
Dan:
>[7/7/04 12:34 PM] Child: Bugfix: Incorrect brace movement cause lost key
>presses.
>git -C branch_reference checkout 536b2f37fdc8438459e7c7b0f44132fcd2204114
>---
>MY BEST GUESS AT THE BRANCH POINT IS:
>commit 093383fb4f9250b58d832fb10e581d90f48abb9c
>Author: Ethan Merritt <xxxxxx@xxxxxx>
>Date: Wed Jun 30 22:03:30 2004 -0700
> #include gp_time.h
>git -C root_reference checkout 093383fb4f9250b58d832fb10e581d90f48abb9c
>diff -ur --exclude=".git" branch_reference/ root_reference/ | grep "^-"| grep
>-v "^---" | wc -l
>diff -ur --exclude=".git" branch_reference/ root_reference/ | grep "^+"| grep
>-v "^+++" | wc -l
>23 + 10
>---
>The distance being quite small and the commit in master after this is
>"Renumbered head trunk version to 4.1.0", which is likely the first commit
>after the bifurcation. Note that Jun 30 is a ways from July 7, but that first
>commit on the new branch is a fix that Ethan also wanted to include in the
>stable release branch.
OK, we have agreement here.
To sum up: everyone agrees on the root points for 5.2, 5.0, 4.0, and 4.4.
I think Dan made an error locating the 4.2 root point, but I could be
wrong. We also have differing ideas about the 4.6 root.
I'm going to to do some more analysis on the 3.7 branch, but the odds
that we'll have to write it off look pretty high.
--
<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>
"Rightful liberty is unobstructed action, according to our will, within limits
drawn around us by the equal rights of others."
-- Thomas Jefferson
|
