1Changes(3) User Contributed Perl Documentation Changes(3)
2
6 REVISION HISTORY - Set::Object
7
9 • Use meta-spec 2, Moose -> dev.requires not runtime.recommends (GH
10 #6, RT #134034) by mohawk
11
13 • Fixed perl5.28 regressions (not cperl), for @$set args to other set
14 functions. (RT #131345). Fixed by adding missing SvGETMAGIC
15 expansions to all args.
16 • Bumped minimal required perl version to 5.8 for arrayref @$set
18 behavior.
19 • Fixed the smokers and some author tests.
21
23 • Skip installing Changes.pod (PR #5, kentfredric)
24
26 • Add test_cover and release targets
27
29 • Test fixes for -Ddefault_inc_excludes_dot, RT #120540
30 • Skip failing threads test <= 5.10.0
32 • Add CI hooks
34
36 • Davem fixed get-magic for the new arguments, fixing NetHack-Item
37 [perl #132152]
38 <https://rt.perl.org/Public/Bug/Display.html?id=132152>. See [cpan
39 #123582] <https://rt.cpan.org/Ticket/Display.html?id=123582>
40 • Stabilize t/misc/kwalitee.t test
42
44 • Change to Artistic License 2.0
45 • Add more misc tests. Add use strict to Set::Object::Weak
47
49 • Make synopsis examples work (tadzik)
50
52 • Fix RT #90070: Improve pod for sorted @$set in list context vs
53 unsorted ->elements (ftobin)
54 • Fix RT #87658: Typo in pod (dsteinbrunner)
56
58 • Reverts Changes.pod patch RT#85244 as it conflicts with the
59 Set::Object pod on metacpan. Also RT #85246
60
62 • Changes.pod patch by Gregor Hermann, debian perl group - RT#85244
63
65 • Fixes for 5.16 and newer - RT#83426 Sort by member names, not the
66 refs. Fix wrong weak test with globals. Changed to lexicals and
67 use strict/warnings.
68 • Fix t/misc/segfault.t, check SvMAGICAL before doing "mg_find"
70 • Fixed wrong C format types in warnings and
72 "SET_OBJECT_MAGIC_backref". Added PREREQ_PMs, make Test::LeakTrace
73 and Moose an optional dependency for t/misc/more_leaks.t
74 • Improve thread-safety (forbid concurrent writes), but still dealing
76 with Attempt to free non-existent shared string and Unbalanced
77 string table refcount: (1) for "8" during global destruction. - RT
78 #22760
79 • Add missing typemap entry const char * for 5.6
81
83 • Fixed a typo in a function which really shouldn't exist - RT#79653
84 • Marked a couple of functions as deprecated for removal (including
86 the above function).
87 • Fixed the magic cleanup code to use the appropriate macro/function
89 to clean up the magic list instead of trying to do it itself.
90 Unsurprisingly, fixes a memory leak. (RT#69967, also RT#67289)
91
93 • Fixed a portability concern for some platforms' C compilers (that
94 do not support //-style comments) - RT#52690
95 • Valid SIGNATURE file - RT#50492
97
99 • Fix a typo that affected building against Perls with debugging
100 enabled.
101 • A minor #define refactor (Yuval).
103 • Squash some long-standing warnings I had been ignoring for ages and
105 Havard Eidnes prompted me to fix by logging an RT ticket for it.
106 • Squash some global clean-up warnings.
108
110 • Methods which destroyed $@ due to internal use of "eval" now
111 properly call "local($@)" (Yuval)
112 • Fix a leak and a corner case with weak set magic, and squash some
114 warnings (Yuval)
115 • Define behaviour when dealing with return values from operations on
117 weak sets. No longer hard-coded "Set::Object" - may affect
118 Set::Object sub-classes (Sam).
119 • (post-release changelog entry) related to the previous change,
121 Set::Object::Weak::set did not use to shift the invocant off the
122 stack, which caused result sets to have the inserted sets as
123 members. It now does, fixing this bug.
124
126 • Portability: support Perls before 5.7.3, and make this dependency
127 explicit in the Makefile.PL
128
130 • Remove threads test which had inadvertantly slipped into the
131 MANIFEST from source control, and add comprehensive documentation
132 on the state of thread support in this module.
133 • Correct other silly MANIFEST mistakes.
135
137 • Make behaviour when dealing with items which were "undef" defined.
138 Previously it might segfault (or Bus Error on Mac OS X) on some
139 operations and Perl versions, or result in an empty string being
140 inserted.
141 • Fix a minor compile problem on IRIX, reported by David Cantrell.
143
145 • Sub-classing interface added; it is now much easier to make
146 Set::Object subclasses that return objects other than more
147 Set::Objects when subclassing. From Brían Mach Aon Innéirghthe.
148
150 • Fix false negative when Test::Pod::Coverage isn't installed.
151 Reported by Anna Bernathova of SuSE.
152
154 • The "->compare" function was returning "disjoint" for empty sets.
155 RT#24965. (Nigel Metheringham)
156 • Document lots of methods that were previously not documented.
158 • Fix "Set::Object::Weak::set()". It was not passing its arguments
160 to "Set::Object::Weak->new()", which was very broken.
161
163 • remove bogus inclusions of Data::Dumper and Devel::Peek
164 • New class "Set::Object::Weak", which all weak sets should get re-
166 blessed into when you call "->weaken". Also added alternative
167 methods of constructing weak sets.
168 • Fix a memory leak with scalar members (the internal hash used to
170 store the items was never being freed). RT#24508.
171
173 • We could only build on gcc due to the use of variadic macros. use
174 "#ifdef" to hide this GCC-specific extension (used only for
175 debugging) from non-GCC compilers.
176 • Add note to man page on compatibility of "Storable" serializations
178 of "Set::Object"s between revisions.
179
181 • fix compile problem affecting some platforms
182
184 • weak reference support
185
187 • Dumping sets with scalars in them via Storable now works.
188
190 • Fix build problem on Win32 (thanks Steffen Goeldner)
191
193 • Fix Makefile.PL so this Changes.pod is not delivered as
194 "Set::Changes" :-}
195 • Restore compatibility with 5.6.0
197 • Remove some compile-time warnings, please let me know if this
199 version does not build on any platform!
200
202 • added patch from Sergey Skvortsov for backwards compatibility with
203 Perl versions before 5.8 (include >100k ppport.h!)
204 • removed '"set()"' constructor from default export list, as its name
206 conflicts with too many other modules.
207
209 • removed debugging &_ func (was causing strange errors on 5.9.1)
210
212 • added '"set()"' constructor and use-as-array-ref interface
213
215 • added 'bool' overload operator to "Set::Object". For backwards
216 compatibility, always returns true.
217 1.08_02, 14 Jan 2005
219 • Cutting out the "Universe" representation; this module gives plain
220 sets only.
221 • documentation fixes
223 1.08_01, 12 Jan 2005
225 • First attempt at adding support for scalars. This version tries to
226 work as closely to "Set::Scalar" as possible.
227 This approach will not be continued unless there are a lot of
229 requests for it to be implemented. I think it's overly
230 complicated, and not what people expect when they want a Set.
231
233 • another segfault bites the dust; this time, avoids a segfault when
234 taking a difference between sets of exactly 31 and 0 size. Same
235 root fault as the previous bug, this time I have a test case for
236 it, too.
237
239 • yet another segfault fix, this time due to incorrect use of PUSHs
240 vs XPUSHs. There is no test case for this bug unfortunately.
241
243 • Fixed segfault when attempting to remove items from an empty set
244 (fix from Richard Sterling Cox)
245
247 • some segfault fixes with Storable and Perl 5.6.1 (Perl 5.8.1 should
248 be OK)
249
251 • added various functions similar to some found in Scalar::Util, but
252 with an extra one that performs a function that lays some
253 groundwork for Container::Object ("ish_int").
254 Some of these functions were in the XS code anyway, and they are
256 extremely small, so I didn't see a problem with duplicating them -
257 saves an extra dependency. Plus, Graham Barr won't let me put
258 "ish_int" or "is_key" in his module.
259 Knowing that they are available will also assist in fixing some
261 longer standing bugs in Tangram (eg, using "$obj + 0" to get a
262 unique ID when "refaddr($obj)" would be better)
263 • lots more input marshalling; if you ever end up doing funny things
265 to a "Set::Object", then at least now you'll get a clearer error
266 message.
267 • more tests; Devel::Cover test coverage now at 91% (stmt), 67%
269 (branch)
270 • erroneously putting scalars into "Set::Object"s now no longer dumps
272 core.
273 • serialisation of "Set::Object" objects is now possible via the
275 Storable module (Data::Dumper still does the wrong thing, but there
276 is nothing I can do about that).
277
279 • fixed overload related bug
280
282 From original Changes file.
283 Revision history for Perl extension Set::Object.
285 0.01 Sat Jun 6 13:36:56 1998
287 - original version; created by h2xs 1.18
288 0.02 Sun Jul 12 1998
290 - fixed refcount bug that leaked object when hash was
291 reorganized
292 - ensured compatibility with Perl 5.004_71
293 1.00 mid-99
295 - fixed some bug (related to subsets?)
296 - first mature release
297 1.01 8 Nov 1999
299 - no longer use Perl arrays internally, now compatible with
300 Perl 5.005_62
301 1.02 8 May 08 2000
303 - fixed bug that miscomputed new bucket while reindexing
304perl v5.34.0 2021-07-22 Changes(3)