1
1
Version Changes for Hypermail
2
2
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3
3
============================
4
- HYPERMAIL VERSION 2.3.1 :
4
+ HYPERMAIL VERSION 2.4.0 :
5
5
============================
6
6
7
- Jose Kahan (June 19, 2018)
8
- - Changed mispelled configuration option warn_surpressions to the
9
- correct spelling warn_suppressions.
10
-
11
- Jose Kahan (June 14, 2018)
12
- - Hypermail would segfault or have an incorrect thread view if it was
13
- handling an archive with a message (msg1) that was a reply to a message not
14
- in the archive, if msg1 had a reply to it in the archive (msg2), msg2 was
15
- archived before msg1 and both msg1 and msg2 shared the same subject
16
- (regardless of Re: prefixes).
17
-
18
- Bill Shannon (June 7, 2018)
19
- - Add charset alias for Thai and Chinese
20
- - The markup for access key "j" to jump to the start of a message
21
- was only being added when the showhtml configuration option
22
- was disabled
23
- - Setting the describe_folder configuration option has no effect on the
24
- display of the folder index page
25
- - In some cases hypermail failed to null terminate the charset name
26
- when copying it, resulting in bogus charset names
27
- - Hypermail would segfault if given the bogus From line "From: <a@b.com"
28
- - In some cases, when no new message was added to the archive,
29
- would segfault because of a null pointer
7
+ 2019-12-20 Jose Kahan
8
+ * src/getdate.y
9
+ A message's sent or received date with a specified TZ but without any
10
+ seconds value was using the local date seconds value instead of 00.
30
11
31
- ============================
32
- HYPERMAIL VERSION 2.3.0:
33
- ============================
12
+ 2019-12-19 Jose Kahan
13
+ * src/parse.c
14
+ parsemail(): a Content-Transfer-Encoding header with a missing value
15
+ would result in an unitialized variable being used to output an unknown
16
+ encoding warning message. Hypermail now skips this header if it's empty.
17
+
18
+ 2019-11-22 Jose Kahan
19
+ * src/print.c
20
+ printsubject() / printdates(): the "by suject" and "by dates" views
21
+ were not escaping markup present in mail's Suject and From headers.
22
+
23
+ 2019-11-19 Jose Kahan
24
+ * src/pcre
25
+ Upgraded to upstream version 8.43. This fixes some autoconf issues we
26
+ had on newer debian releases. This is only relevant for people that
27
+ are not linking against a system libpcre.
28
+
29
+ * tests/diff_hypermail_archives.pl
30
+ Rewrote diff_hypermail.pl to make it more flexible and powerful for
31
+ finding the differences in two hypermail archives.
32
+
33
+ 2018-12-07 Jose Kahan
34
+ * If the antispamdomain config option was used, hypermail was replacing
35
+ everything following a "@" character by the antispam domain. This fix
36
+ makes antispamdomain only apply for the domain part of valid email
37
+ addresses.
38
+
39
+ 2018-11-29 Bill Shannon
40
+ * string.c
41
+ spamify_replacedomain(): fix another greedy "@" replacement in the
42
+ middle of a string of iso-2022-jp encoded Japanese characters.
43
+
44
+ 2018-11-04 Bill Shannon
45
+ * string.c
46
+ i18n_convstring(): When converting a utf-8 string to iso-2022-jp,
47
+ hypermail always adds the "return to ASCII" escape sequence, even if
48
+ the string ends with ASCII.
49
+
50
+ parseemail(): hypermail attempts to transform anything that looks like
51
+ an email address by replacing the at sign (and optionally the domain
52
+ name) with a string (default "at") to confuse screen-scraping
53
+ programs. But if there's an "@" character in the middle of a string of
54
+ iso-2022-jp encoded Japanese characters, hypermail still converts it,
55
+ breaking the Japanese character.
56
+
57
+ 2018-10-12 Jose Kahan
58
+ * setup.c PreConfig(): removed a 64-bit warning
59
+
60
+ * hypermail.c
61
+ main(): if the en_US locale is not available, try en_US.UTF-8
62
+
63
+ 2018-10-11 Jose Kahan
64
+
65
+ * string.c
66
+ parseemail(): for some reason if a ',' char was concatenaned to an
67
+ email address, it was being parsed as part of the email username.
68
+
69
+ 2018-10-10 Jose Kahan
70
+
71
+ * string.c
72
+ parseurl(): remove deprecated URLs. Improve support for URLs that don't
73
+ have slashes following the 'protocol:' schema. Improve support for tel:
74
+ URLs.
75
+
76
+ 2018-10-09 Bill Shannon
77
+
78
+ * string.c
79
+ parseurl(): non-URL text can be misinterpreted as a URL, causing segfault
80
+ Fixes issue #39
81
+
82
+ 2018-10-08 Jose Kahan
83
+
84
+ * print.c, string.c, proto.h
85
+ The inreplyto_command could generate invalid links if the parser interpolated the
86
+ in-reply-to from the subject. If a message's in-reply-to is interpolated,
87
+ hypermail won't honor it anymore for that message.
88
+
89
+ 2018-10-07 Jose Kahan
90
+
91
+ * print.c printauthors(), printthreads(), printsubjects, printdates()
92
+ Even if set_i18n was enabled, Indexes were written with mixed charsets
93
+ instead of using utf-8 throughout.
94
+
95
+ * configure.ac, Makefile.in, src/Makefile.in
96
+ If the system has a recent libpcre, compile against it instead of using
97
+ the bundled one. This check is done by default when launching configure.
98
+
99
+ New configure options to allow to link against an external pcre library
100
+ (--with-external-pcre) or to force the build and link against the
101
+ bundled one (--enable-bundled-pcre).
102
+
103
+ New configure option to allow to compile against a system libtrio
104
+ (--enable-system-libtrio). Contrary to libpcre, it's not possible to find the
105
+ version of libtrio so we cannot compile automatically against it.
106
+
107
+ 2018-10-04 Jose Kahan
108
+
109
+ * src/pcre
110
+ Upgraded to upstream version 8.42
111
+
112
+ * src/trio
113
+ Upgraded to upstream version 1.16
114
+
115
+ * parse.c
116
+ Only use the headers charset for text/plain if its absent. If no charset
117
+ is available in the other parts, do not add it.
118
+
119
+ 2018-10-04 Jose Kahan
120
+
121
+ * Updated Changelog format for entries newer than version 2.3.0
122
+
123
+ * hypermail.h, parse.c, struct.c, struct.h
124
+ Charset handling for multipart messages was being handled wrong, giving
125
+ priority to the headers charset (if found) over that of the displayed body.
126
+ Sometimes the last found charset was the one being used throughout in the
127
+ generated body.
128
+
129
+ The metafile for attachments was sometimes inheriting the charset of the whole
130
+ body, even when not necessary or wrong. Now the meta doesn't include a charset
131
+ if the attachment doesn't explicitly give one.
132
+
133
+ * parse.c
134
+ References header was processed multiple times as it was not being marked as parsed
135
+ after being processed.
136
+ The epilogue of MIME parts was not being ignored.
137
+
138
+ 2018-06-19 Jose Kahan
139
+
140
+ * Changed mispelled configuration option warn_surpressions to the
141
+ correct spelling warn_suppressions.
142
+
143
+ 2018-06-14 Jose Kahan
144
+
145
+ * Hypermail would segfault or have an incorrect thread view if it was
146
+ handling an archive with a message (msg1) that was a reply to a message
147
+ not in the archive, if msg1 had a reply to it in the archive (msg2),
148
+ msg2 was archived before msg1 and both msg1 and msg2 shared the same
149
+ subject (regardless of Re: prefixes).
150
+
151
+ 2018-06-07 Bill Shannon
152
+
153
+ * Add charset alias for Thai and Chinese
154
+
155
+ * The markup for access key "j" to jump to the start of a message was
156
+ only being added when the showhtml configuration option was disabled
157
+
158
+ * Setting the describe_folder configuration option has no effect on the
159
+ display of the folder index page
160
+
161
+ * In some cases hypermail failed to null terminate the charset name
162
+ when copying it, resulting in bogus charset names
163
+
164
+ * Hypermail would segfault if given the bogus From line "From:
165
+ <a@b.com"
166
+
167
+ * In some cases, when no new message was added to the archive, would
168
+ segfault because of a null pointer
169
+
170
+ 2015-04-03 Ivan Kuraj
34
171
35
- Ivan Kuraj (Apr 3, 2015)
36
172
* Correct includes for new glibc
37
173
38
- Jose Kahan (Jun 10, 2013)
174
+ 2013-06-10 Jose Kahan
175
+
39
176
* Even if a message was annotated as spam/deleted, its Attachments were
40
177
still being created
41
178
42
- Jose Kahan (Apr 18, 2013)
179
+ 2013-04-18 Jose Kahan
180
+
43
181
* Removed commented out code in printfile.c.
182
+
44
183
* Added missing ';' in generated css rules in the same file
45
184
46
- Jose Kahan (Apr 11, 2013)
185
+ 2013-04-11 Jose Kahan
186
+
47
187
* Extended the configure options so that a user can decide if hypermail
48
188
should be compiled and statically linked against the bundled pcre lib
49
189
or if it should be dynamically linked with an external (or system) pcre
50
190
library (if available).
191
+
51
192
* The new configure option is --with-external-pcre[=dir].
193
+
52
194
* The default value for dir is /usr and the include and libraries
53
195
directories will be searched from that starting path. This option is
54
196
disabled by default.
197
+
55
198
* The make clean in libfnv wasn't removing the libfnv.a file
56
199
57
- Jose Kahan (Mar 29, 2013)
200
+ 2013-03-29 Jose Kahan
58
201
59
202
* Updated configure to latest autoconf syntax conventions.
60
203
61
204
* Renamed configure.in to configure.ac and cleaned it up partially.
62
205
63
- Jose Kahan (Mar 20, 2013)
64
-
206
+ 2013-03-20 Jose Kahan
207
+
65
208
* Updated pcre to pcre-8.32 from pcre-4.3. This may have broken the
66
209
Windows lcc compilation but I don't have the means to verify it.
67
210
68
- Jose Kahan (Mar 15, 2013)
69
-
211
+ 2013-03-15 Jose Kahan
212
+
70
213
* Updated trio code to trio version 1.14. Moved all the trio code to
71
214
src/trio. Now compiling trio as a static library and linking against
72
215
it, which makes binary a bit smaller. See src/trio/README.hypermail for
73
216
details.
74
217
75
- Jose Kahan (Mar 12, 2013)
76
-
218
+ 2013-03-12 Jose Kahan
219
+
77
220
* New experimental support for RFC 3676 format=flowed. There are two
78
221
new related configuration options: format_flowed to enable it and
79
222
format_flowed_disable_quotes to disable it in specific cases. See the
80
223
documentation and the RFC for further info.
81
224
82
- Jose Kahan (Mar 08, 2013)
83
-
225
+ 2013-03-08 Jose Kahan
226
+
84
227
* Error: string.c:parseurl() assumes that all URL motifs it can match
85
228
end with a :// string. However, the URL match table it uses included
86
229
some URL motifs that didn't have that string. These have been commented
87
230
as they may cause parseulr to SIGSEVs in some cases.
88
231
89
- Jose Kahan (Feb 26, 2013)
90
-
232
+ 2013-02-26 Jose Kahan
233
+
91
234
* New configuration option, noindex_onindexes, for associating a
92
235
"noindex" robot metadata value with hypermail generated indexes.
93
236
94
- Jose Kahan (Feb 26, 2013)
95
-
237
+ 2013-02-26 Jose Kahan
238
+
96
239
* New configuration option, userobotmeta, for Associatating robot
97
240
annotations with attachments, using the experimental X-Robots-Tag HTTP
98
241
header
@@ -104,28 +247,28 @@ Jose Kahan (Feb 26, 2013)
104
247
* Make sure the delete stubs are added only when the delete_level
105
248
option says we're not to preserve the original message body.
106
249
107
- Jose Kahan (Jan 30, 2013)
108
-
250
+ 2013-01-30 Jose Kahan
251
+
109
252
* Migrated source code repository from cvs.hypermail.org to
110
253
http://sourceforge.net/projects/hypermail/
111
-
254
+
112
255
* Added a mirror of the repository to github:
113
256
http://github.com/hypermail-project/hypermail
114
-
257
+
115
258
* Cleared up the markup errors in docs/hmrc.html by using tidy.
116
-
259
+
117
260
* New configuration directive "annotated" which allows a mailing list
118
261
archive editor to annotate a given message before converting it into
119
262
HTML. Associated additional new directives are
120
263
"htmlmessage_deleted_spam", "htmlmessage_deleted_other",
121
264
"htmlmessage_edited". Consult docs/hmrc.html for more info.
122
-
265
+
123
266
* "htmlmessage_deleted" is deprecated in favor of
124
267
"htmlmessage_deleted_spam"
125
-
268
+
126
269
* "deleted" is superseded by "annotated", although it will be supported
127
270
to take into account legacy archives.
128
-
271
+
129
272
* Regenerated the dependencies for the src/Makefile.in file.
130
273
131
274
============================
0 commit comments