options: configurable timeout for MPD connections
command.c: comment an #endif
command.c: change find_key_command for MAX_COMMAND_KEYS != 3
The performance penalty shouldn't be too high (loop unrolling could
help here).
The performance penalty shouldn't be too high (loop unrolling could
help here).
command.c: turn while loops into for loops
command.c: rework loop logic in find_key_command
screen_interface: document the cmd callback
command: make Ctrl-X/Alt-X/F1 etc. translatable
screen_lyrics: new key to edit lyrics
Thanks for the idea and an initial patch go to Jitka Novotna <jitka@ucw.cz>
Thanks for the idea and an initial patch go to Jitka Novotna <jitka@ucw.cz>
screen_lyrics: put code for reloading in a function
Add option "text-editor"
The lyrics screen is going to spawn a text editor to edit lyrics.
Put the infrastructure in place to make it configurable.
The lyrics screen is going to spawn a text editor to edit lyrics.
Put the infrastructure in place to make it configurable.
Merge remote branches 'jn/cosmetics', 'jn/doxygen' and 'jn/renames'
screen_song: make Path a proper label
command.c: use for instead of while in check_key_bindings
command.c: put curly braces around a single large statement
main: don't show "connecting to (null)"
This patch is a follow-up to commit b774eac8d9 ("main: don't show
(null) as the server name").
This patch is a follow-up to commit b774eac8d9 ("main: don't show
(null) as the server name").
screen_song: don't show unknown kbit rates
(calling mpd_status_get_kbit_rate twice in a row shouldn't impact
performance, it's declared as "pure")
(calling mpd_status_get_kbit_rate twice in a row shouldn't impact
performance, it's declared as "pure")
rename screen_message to screen_status
Its two functions are screen_status_message and screen_status_printf
so this just seems logical to me.
Its two functions are screen_status_message and screen_status_printf
so this just seems logical to me.
doxygen: don't sort functions in the file documentation
add doc/doxygen.conf and doc/api/ to .gitignore
mention doxygen support in NEWS
add doxygen support
use mpdclient_finish_command at the obvious places
mpdclient: export mpdclient_finish_command
(and include mpdclient.h in mpdclient.c, so it can be used there)
(and include mpdclient.h in mpdclient.c, so it can be used there)
options.c: rewrite two #ifndefs as one #if
configure.ac: (default: enabled) -> [default=yes]
apply whitespace cosmetics to the license comments
ncmpc.h: document sigstop
Makefile.am: Add @CURSES_LIB@ to link the test program
mpdclient: define MPD_IDLE_ALL
configure.ac: Remove duplicate output screen merge error.
configure.ac: Change to the more correct variable for wide support.
screen_list: use G_N_ELEMENTS()
Merge remote branches 'avuton/master' and 'jn/dev'
m4: Add robust curses detection and support
This commit adds the ax_with_curses.m4 macro from the autoconf-archive
project. This also adds support for more plain curses, proper color detection
and compartmentalizes the curses autoconf stuff.
This commit adds the ax_with_curses.m4 macro from the autoconf-archive
project. This also adds support for more plain curses, proper color detection
and compartmentalizes the curses autoconf stuff.
screen.c: Add pdcurses mouse support
configure.ac: Add proper CFLAG checking and appending
Add proper CFLAG checking and appending using standard autoconf macros. This
fixes the warnings generated during autoconf.
Add proper CFLAG checking and appending using standard autoconf macros. This
fixes the warnings generated during autoconf.
wreadln.c: Win32 doesn't support polling, so disable this for now
Win32 doesn't support polling, disabling it for now. Not exactly sure
what this breaks, but win32 appears to run fine with this code disabled.
Win32 doesn't support polling, disabling it for now. Not exactly sure
what this breaks, but win32 appears to run fine with this code disabled.
main.c: Don't try to compile signal support on win32
conf.c: When on win32, mkdir() doesn't use permissions
screen.c: If compiling with PDCURSES, use resize_term() to resize
gidle.c: Removed unused header
configure.ac: Link to winsock when compiling on win32
configure.ac: Replace AM_CONFIG_HEADER deprecated macro
configure.ac: Remove disabled browse screen options
configure.ac: Add AC_MSG_RESULT for outputs screen
configure.ac: Amend -Werror using the cflag framework
m4: Update nls.m4
configure.ac: AC_HELP_STRING depreciated, move to AS_HELP_STRING
list_window: scroll down exactly one page at PGDN
main: use mpdclient_is_connected in assert
screen: restore newline character
It was removed in commit 389a1ad8:
- fprintf(stderr, _("Error: Screen to small\n"));
+ fprintf(stderr, "%s", _("Error: Screen too small"));
exit(EXIT_FAILURE);
It was removed in commit 389a1ad8:
- fprintf(stderr, _("Error: Screen to small\n"));
+ fprintf(stderr, "%s", _("Error: Screen too small"));
exit(EXIT_FAILURE);
mpdclient: handle mpd_send_add's return value
configure.ac: add missing AC_MSG_RESULT
mpdclient_connect: fix parenthesis style
screen_outputs.c: fix indentation
title_bar.c: include config.h directly
options: remove double-semicolon
NEWS: screen_queue: follow songs being moved
Done in commit d71be00d7bbd135c1cd7ec2cfe38774a8a0b373f.
Done in commit d71be00d7bbd135c1cd7ec2cfe38774a8a0b373f.
convert NEWS to UTF-8
resolve.c: Only include netdb.h when not on WIN32
screen.c: re-enable the leaveok optimization
Of course only if we don't need the cursor.
Of course only if we don't need the cursor.
screen_lyrics: fix lyrics_title on small screens
It turns out snprintf returns the length of the resulting string, not
the actual number of bytes copied into the buffer, e.g. the following
call will return 33, and not 7 (or 8):
snprintf(buffer, 8, "this string is 33 characters long");
This patch also adds empty lines around the code dealing with the
plugin name to help readabiliy.
It turns out snprintf returns the length of the resulting string, not
the actual number of bytes copied into the buffer, e.g. the following
call will return 33, and not 7 (or 8):
snprintf(buffer, 8, "this string is 33 characters long");
This patch also adds empty lines around the code dealing with the
plugin name to help readabiliy.
remove unused typedef screen_functions_t
<@cirrus> legacy, I think.
<@cirrus> legacy, I think.
screen.c: don't disable CMD_SCREEN_SWAP on !OUTPUTS_SCREEN
Makefile.am: fix a typo ("smalles")
screen_queue: fix a typo: "revceived"
filelist: simplify some code
clang --analyze pointed me to this code:
src/filelist.c:146:3: warning: Value stored to 'iter' is never read
clang --analyze pointed me to this code:
src/filelist.c:146:3: warning: Value stored to 'iter' is never read
document screen_find, make it return bool
po: update Ukrainian translation
Minor grammar errors and inconsistent punctuation in the ncmpc man page
Merge branch 'regr' of git://git.musicpd.org/jn/ncmpc
screen_queue: fix CMD_LIST_MOVE_*
The code now tries to keep as much as possible of the selection being
moved visible.
The code now tries to keep as much as possible of the selection being
moved visible.
list_window: export scroll-after-cursor code
This moves the list_window_check_origin code into a new, more generic
function, which is exported.
This moves the list_window_check_origin code into a new, more generic
function, which is exported.
options: fix the --key-file help text
Revert "screen_queue: use list_window_move_cursor"
I hadn't quite understood what list_window_move_cursor does, and didn't
test the change.
This reverts commit aba25b543d7a34ce46964a7f22e495cdc3b57852.
I hadn't quite understood what list_window_move_cursor does, and didn't
test the change.
This reverts commit aba25b543d7a34ce46964a7f22e495cdc3b57852.
configure.ac: update comment "Check for glib-2.4"
main: avoid gcc warning for LOCALE && !NLS
Mark charset as G_GNUC_UNUSED if ncmpc hasn't been configured with NLS.
Mark charset as G_GNUC_UNUSED if ncmpc hasn't been configured with NLS.
doc/ncmpc.lirc: fix a typo ("ncmcp")
options: remove misleading MAX_LONGOPT_LENGTH
list_window: "if ( x)" -> "if (x)"
manpage: update the search-mode documentation
Merge remote branches 'jn/doc', 'jn/stuff' and 'jn/stuff1'
screen_queue: use list_window_move_cursor
It ensures the cursor stays visible.
It ensures the cursor stays visible.
list_window: document list_window_center
fix a typo ("connectiond")
plugin: fix plugin_callback_t documentation
screen_lyrics: clarify some code
list_window.h: "non-zero" -> "true", we use bool
command: use RET instead of 13 for '\r'
command: use KEY_CTL
And add the convenience macro C().
And add the convenience macro C().
command: add macro KEY_CTL()
I copied it from tig[1], which is licensed under the GPLv2+, so we
may use the code.
[1] http://jonas.nitro.dk/tig/
I copied it from tig[1], which is licensed under the GPLv2+, so we
may use the code.
[1] http://jonas.nitro.dk/tig/
command: white space cosmetics
Restore the old idention style of the key code macros, it was
destoyed by commit f81333d (code style, indent with tabs XI).
(That's a long time ago, I know.)
Restore the old idention style of the key code macros, it was
destoyed by commit f81333d (code style, indent with tabs XI).
(That's a long time ago, I know.)
screen_artist: document add_query (some more)
screen_artist: don't use artist directly in add_query
screen_artist: add_query: add a "const"
screen_search: make SEARCH_ARTIST_TITLE part of an enum
conf.c: unbreak two short lines, remove parentheses
increase connection timeout to 5 seconds
1.5 seconds was not long enough for some MPD operations. The
disadvantage is that the user may have to wait up to 5 seconds without
screen updates, because ncmpc doesn't do full asynchronous I/O yet.
1.5 seconds was not long enough for some MPD operations. The
disadvantage is that the user may have to wait up to 5 seconds without
screen updates, because ncmpc doesn't do full asynchronous I/O yet.
mpdclient: pass integer timeout to mpdclient_connect()
Use milliseconds instead of float seconds. Reduces overhead.
Use milliseconds instead of float seconds. Reduces overhead.
screen_lyrics: prevent a double free/use after free
And a memory leak.
And a memory leak.