Re: Changelogs on kernel.org

[Tomas Szepe]

> > > Alright, another one.
> > > 
> > > $ ./fmtcl2.pl /usr/src/ChangeLog-2.5.14
> > > 
> > > Summary of changes from v2.5.13 to v2.5.14
> > > ============================================
> > > 
> > > o  kd.h                                                                             ([e-mail])
> > > o  NTFS: Release 2.0.6 - Major bugfix to make compatible with other kernel changes. ([e-mail])
> > > o  mm/memory.c:                                                                     ([e-mail])
> > > o  suppress allocation warnings for radix-tree allocations                          ([e-mail])
> > 
> > <snip>
> > 
> > Ah - thats it!
> 
> Not quite.  In mutt, it looks like this:
> 
> o  kd.h
> +([e-mail])
> o  NTFS: Release 2.0.6 - Major bugfix to make compatible with other kernel
> +changes. ([e-mail])
> o  mm/memory.c:
> +([e-mail])
> o  suppress allocation warnings for radix-tree allocations
> +([e-mail])
> o  radix-tree locking fix
> +([e-mail])
> o  Allow truncate to discard unmapped buffers
> +([e-mail])
> o  decouple swapper_space treatment from other address_spaces
> 
> NOT easy to read.

Hmm, now that's true.

How's this for another try? (logwidth = 38, adjustable per taste)
Myself I'd say this is crap and a nice 140-char xterm will do much
better (you're going to see "long" lines anyway, look at the NTFS
entry below).

$ ./fmtcl3.pl /usr/src/ChangeLog-2.5.14

Summary of changes from v2.5.13 to v2.5.14
============================================

o  kd.h                                ([e-mail])
o  NTFS: Release 2.0.6 - Major bugfix to make compatible with other kernel changes.
                                       ([e-mail])
o  mm/memory.c:                        ([e-mail])
o  suppress allocation warnings for radix-tree allocations
                                       ([e-mail])
o  radix-tree locking fix              ([e-mail])
o  Allow truncate to discard unmapped buffers
                                       ([e-mail])
o  decouple swapper_space treatment from other address_spaces
                                       ([e-mail])
o  Allow ext3 pages to be written back by VM pressure
                                       ([e-mail])
o  Fix SMP race in truncate            ([e-mail])
o  handle concurrent block_write_full_page and set_page_dirty
                                       ([e-mail])
o  Fix PG_launder                      ([e-mail])
o  Fix concurrent writepage and readpage
                                       ([e-mail])
<snip>

#!/usr/bin/perl -w

use strict;

my $logwidth = 38;

my %people = ();
my $addr = "";
my @cur = ();

sub append_item()
{
	if (!$addr) { return; }
	if (!$people{$addr}) { @{$people{$addr}} = (); }
	push @{$people{$addr}}, [@cur];

	@cur = ();
}

sub print_terse_items($)
{
	my @items = @{$people{$_[0]}};
	while ($_ = shift @items) {
		my $len = length(@$_[0]);
		my $fill = ($len <= $logwidth) ? (" " x ($logwidth - $len + 1)) : ("\n" . (" " x ($logwidth + 1)));
		print "@$_[0]$fill($_[0])\n";
	}
}

while (<>) {
	# Match address
	if (/^\s*<([^>]+)>/) {
		# Add old item (if any) before beginning new
		append_item();
		$addr = $1;
	} elsif ($addr) {
		# Add line to patch
		# strip whitespace at start & end
		s/^\s*(.*)\s*$/$1/;
		# kill starting 'PATCH' captions
		s/\s*\[?PATCH\]?\s*//;
		# insert a bullet
		s/^(.*)$/o  $1/;
		# Only save 1 line
		if (scalar(@cur) < 1) { push @cur, "$_"; }
	} else {
		# Header information
		print;
	}
}

append_item();
foreach $addr (sort keys %people) {
	print_terse_items($addr);
}