tweak change#4745 to make ebcdic output match for chars <= 037
[p5sagit/p5-mst-13.2.git] / ext / Data / Dumper / Changes
CommitLineData
823edd99 1=head1 NAME
2
3HISTORY - public release history for Data::Dumper
4
5=head1 DESCRIPTION
6
7=over 8
8
a2126434 9=item 2.11 (unreleased)
10
11C<0> is now dumped as such, not as C<'0'>.
12
13qr// objects are now dumped correctly (provided a post-5.005_58)
14overload.pm exists).
15
16Implemented $Data::Dumper::Maxdepth, which was on the Todo list.
17Thanks to John Nolan <jpnolan@Op.Net>.
18
19=item 2.101 (30 Apr 1999)
20
21Minor release to sync with version in 5.005_03. Fixes dump of
22dummy coderefs.
23
7820172a 24=item 2.10 (31 Oct 1998)
25
26Bugfixes for dumping related undef values, globs, and better double
27quoting: three patches suggested by Gisle Aas <gisle@aas.no>.
28
29Escaping of single quotes in the XS version could get tripped up
30by the presence of nulls in the string. Fix suggested by
31Slaven Rezic <eserte@cs.tu-berlin.de>.
32
33Rather large scale reworking of the logic in how seen values
34are stashed. Anonymous scalars that may be encountered while
35traversing the structure are properly tracked, in case they become
36used in data dumped in a later pass. There used to be a problem
37with the previous logic that prevented such structures from being
38dumped correctly.
39
40Various additions to the testsuite.
41
823edd99 42=item 2.09 (9 July 1998)
43
44Implement $Data::Dumper::Bless, suggested by Mark Daku <daku@nortel.ca>.
45
46=item 2.081 (15 January 1998)
47
48Minor release to fix Makefile.PL not accepting MakeMaker args.
49
50=item 2.08 (7 December 1997)
51
52Glob dumps don't output superflous 'undef' anymore.
53
54Fixes from Gisle Aas <gisle@aas.no> to make Dumper() work with
55overloaded strings in recent perls, and his new testsuite.
56
57require 5.004.
58
59A separate flag to always quote hash keys (on by default).
60
61Recreating known CODE refs is now better supported.
62
63Changed flawed constant SCALAR bless workaround.
64
65=item 2.07 (7 December 1996)
66
67Dumpxs output is now exactly the same as Dump. It still doesn't
68honor C<Useqq> though.
69
70Regression tests test for identical output and C<eval>-ability.
71
72Bug in *GLOB{THING} output fixed.
73
74Other small enhancements.
75
76=item 2.06 (2 December 1996)
77
78Bugfix that was serious enough for new release--the bug cripples
79MLDBM. Problem was "Attempt to modify readonly value..." failures
80that stemmed for a misguided SvPV_force() instead of a SvPV().)
81
82=item 2.05 (2 December 1996)
83
84Fixed the type mismatch that was causing Dumpxs test to fail
85on 64-bit platforms.
86
87GLOB elements are dumped now when C<Purity> is set (using the
88*GLOB{THING} syntax).
89
90The C<Freezer> option can be set to a method name to call
91before probing objects for dumping. Some applications: objects with
92external data, can re-bless themselves into a transitional package;
93Objects the maintain ephemeral state (like open files) can put
94additional information in the object to facilitate persistence.
95
96The corresponding C<Toaster> option, if set, specifies
97the method call that will revive the frozen object.
98
99The C<Deepcopy> flag has been added to do just that.
100
101Dumper does more aggressive cataloging of SCALARs encountered
102within ARRAY/HASH structures. Thanks to Norman Gaywood
103<norm@godel.une.edu.au> for reporting the problem.
104
105Objects that C<overload> the '""' operator are now handled
106properly by the C<Dump> method.
107
108Significant additions to the testsuite.
109
110More documentation.
111
112=item 2.04beta (28 August 1996)
113
114Made dump of glob names respect C<Useqq> setting.
115
116[@$%] are now escaped now when in double quotes.
117
118=item 2.03beta (26 August 1996)
119
120Fixed Dumpxs. It was appending trailing nulls to globnames.
121(reported by Randal Schwartz <merlyn@teleport.com>).
122
123Calling the C<Indent()> method on a dumper object now correctly
124resets the internal separator (reported by Curt Tilmes
125<curt@ltpmail.gsfc.nasa.gov>).
126
127New C<Terse> option to suppress the 'C<VARI<n> = >' prefix
128introduced. If the option is set, they are output only when
129absolutely essential.
130
131The C<Useqq> flag is supported (but not by the XSUB version
132yet).
133
134Embedded nulls in keys are now handled properly by Dumpxs.
135
136Dumper.xs now use various integer types in perl.h (should
137make it compile without noises on 64 bit platforms, although
138I haven't been able to test this).
139
140All the dump methods now return a list of strings in a list
141context.
142
143
144=item 2.02beta (13 April 1996)
145
146Non portable sprintf usage in XS code fixed (thanks to
147Ulrich Pfeifer <pfeifer@charly.informatik.uni-dortmund.de>).
148
149
150=item 2.01beta (10 April 1996)
151
152Minor bugfix (single digit numbers were always getting quoted).
153
154
155=item 2.00beta (9 April 1996)
156
157C<Dumpxs> is now the exact XSUB equivalent of C<Dump>. The XS version
158is 4-5 times faster.
159
160C<require 5.002>.
161
162MLDBM example removed (as its own module, it has a separate CPAN
163reality now).
164
165Fixed bugs in handling keys with wierd characters. Perl can be
166tripped up in its implicit quoting of the word before '=>'. The
167fix: C<Data::Dumper::Purity>, when set, always triggers quotes
168around hash keys.
169
170Andreas Koenig <k@anna.in-berlin.de> pointed out that handling octals
171is busted. His patch added.
172
173Dead code removed, other minor documentation fixes.
174
175
176=item 1.23 (3 Dec 1995)
177
178MLDBM example added.
179
180Several folks pointed out that quoting of ticks and backslashes
181in strings is missing. Added.
182
183Ian Phillips <ian@pipex.net> pointed out that numerics may lose
184precision without quotes. Fixed.
185
186
187=item 1.21 (20 Nov 1995)
188
189Last stable version I can remember.
190
191=back
192
193=cut