aboutsummaryrefslogtreecommitdiffstats
path: root/lib-src/ChangeLog
diff options
context:
space:
mode:
Diffstat (limited to 'lib-src/ChangeLog')
-rw-r--r--lib-src/ChangeLog73
1 files changed, 73 insertions, 0 deletions
diff --git a/lib-src/ChangeLog b/lib-src/ChangeLog
index 8d2c95e671c..0bb24c375e7 100644
--- a/lib-src/ChangeLog
+++ b/lib-src/ChangeLog
@@ -1,3 +1,76 @@
12015-03-27 Paul Eggert <eggert@cs.ucla.edu>
2
3 Port etags to -DDEBUG
4 * etags.c (xnew, xrnew) [DEBUG]: Don't include chkmalloc.h, which
5 is not part of Emacs and is typically not installed.
6 Instead, just invoke xmalloc and xrealloc as usual.
7 Problem reported by Nicolas Richard in:
8 http://bugs.gnu.org/20191#20
9 (xrnew): Avoid no-longer-needed cast to 'char *'.
10 (xrealloc): First arg is now void *, not char *.
11
122015-03-06 Paul Eggert <eggert@cs.ucla.edu>
13
14 Random minor fixes for movemail
15 * movemail.c: Include <stdbool.h> and <signal.h>.
16 (waitpid) [WINDOWSNT]: New macro.
17 (wait) [WINDOWSNT]: Remove.
18 (main, popmail, pop_retr, mbx_write, mbx_delimit_begin)
19 (mbx_delimit_end): Use bool for boolean.
20 (main): Simplify #if usage a bit.
21 (main): Don't assume EOF == -1. Prefer 'return' to 'exit'. Don't
22 possibly unlink lockname twice, as that's a race condition. Set
23 SIGCHLD to SIG_DFL to work around SysV misfeature. Check for fork
24 failure. Use waitpid, not wait, to avoid a race condition in the
25 unlikely case where we start up with a child.
26 (NOTOK, OK): Remove, in favor of plain boolean.
27 (popmail, pop_retr): Don't get confused about errno, e.g., ferror
28 need not set errno.
29 (popmail): Use fclose (mbf), not close (fileno (mbf)), to also
30 detect any stream-related errors (e.g., memory exhaustion).
31 (pop_retr): Report pop errors separately, since caller now does
32 errno reporting.
33 (mbx_write, mbx_delimit_begin, mbx_delimit_end): Check < 0, not ==
34 EOF, as it's a bit faster and (in theory) pickier.
35
362015-02-27 Mark Laws <mdl@60hz.org>
37
38 Support daemon mode on MS-Windows (bug#19688)
39 * emacsclient.c (decode_options) [WINDOWSNT]: Don't reject empty
40 arguments for --alternate-editor.
41 (print_help_and_exit) [WINDOWSNT]: Don't refrain from advertising
42 empty arguments for --alternate-editor.
43 (start_daemon_and_retry_set_socket) [WINDOWSNT]: MS-Windows
44 specific code to start Emacs in daemon mode and wait for it to be
45 ready for client connections.
46
472015-02-23 Pete Williamson <petewil0@googlemail.com> (tiny change)
48
49 Use ${EXEEXT} more uniformly in makefiles
50 * Makefile.in (EMACS): Append ${EXEEXT}.
51
522015-02-20 Paul Eggert <eggert@cs.ucla.edu>
53
54 Simplify binary I/O configuration
55 * etags.c: Include <sysstdio.h> rather than <stdio.h>.
56 (process_file_name, analyze_regex): Use FOPEN_BINARY rather than
57 hard-coded "b".
58
592015-02-19 Eli Zaretskii <eliz@gnu.org>
60
61 * etags.c (process_file_name) [!DOS_NT]: Use "r", not "rb" in the
62 call to 'popen'. (Bug#19735)
63
642015-02-13 Paul Eggert <eggert@cs.ucla.edu>
65
66 Better support for future plugins
67 See the thread containing:
68 http://lists.gnu.org/archive/html/emacs-devel/2015-02/msg00720.html
69 * make-docfile.c (write_globals): Generate code that #defines
70 Qxxx macros other than Qnil only if DEFINE_NONNIL_Q_SYMBOL_MACROS.
71 Qnil is safe to define even in plugins, since it must be zero for
72 other reasons.
73
12015-01-24 Paul Eggert <eggert@cs.ucla.edu> 742015-01-24 Paul Eggert <eggert@cs.ucla.edu>
2 75
3 Fix a couple of AM_V_GEN bugs 76 Fix a couple of AM_V_GEN bugs