previous song played will be reused.

git-svn-id: https://svn.musicpd.org/mpd/trunk@5791 09075e82-0dd4-0310-85a5-a0d7c8717e4f

forever without this.

git-svn-id: https://svn.musicpd.org/mpd/trunk@5790 09075e82-0dd4-0310-85a5-a0d7c8717e4f

installing libs/includes, and apparently even the defaults can break cross
compiles (see bug #1454).

git-svn-id: https://svn.musicpd.org/mpd/trunk@5789 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5788 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5787 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5786 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5733 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5729 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5697 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5680 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5679 09075e82-0dd4-0310-85a5-a0d7c8717e4f

pointing it out.

git-svn-id: https://svn.musicpd.org/mpd/trunk@5673 09075e82-0dd4-0310-85a5-a0d7c8717e4f

LOCATE_TAG_KEY_FILE.  Specifying "file: " as an argument to
search/find/list wasn't the point of that patch...

git-svn-id: https://svn.musicpd.org/mpd/trunk@5670 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5669 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5668 09075e82-0dd4-0310-85a5-a0d7c8717e4f

from a branch of 0.12.1.

git-svn-id: https://svn.musicpd.org/mpd/trunk@5666 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5665 09075e82-0dd4-0310-85a5-a0d7c8717e4f

to segfault.  This could be exploited by malicious users to crash other
users' mpd.  But more importantly, I believe clients are doing this
unintentionally, and that this is what is causing mpd to segfault for many
people after running for long periods of time.

git-svn-id: https://svn.musicpd.org/mpd/trunk@5649 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5589 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5492 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5460 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5455 09075e82-0dd4-0310-85a5-a0d7c8717e4f

PKG_PROG_PKG_CONFIG to be done first.  But this is only done if PulseAudio
and JACK aren't explitily disabled.  Therefore, using ./configure
--disable-pulse --disable-jack will make it so that neither LSR nor
Zeroconf can be found.  This is the second time this bug has come up (first
being when JACK was added), so can people please start adding to the test
around PKG_PROG_PKG_CONFIG when they add new features that need
PKG_CHECK_MODULES?

git-svn-id: https://svn.musicpd.org/mpd/trunk@5453 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5422 09075e82-0dd4-0310-85a5-a0d7c8717e4f

playlist.

git-svn-id: https://svn.musicpd.org/mpd/trunk@5420 09075e82-0dd4-0310-85a5-a0d7c8717e4f

later be used for playlist searching.

git-svn-id: https://svn.musicpd.org/mpd/trunk@5419 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5418 09075e82-0dd4-0310-85a5-a0d7c8717e4f

Mixing code and declarations is ugly, anyways.

We could probably get away with using alloca(), but I'm not sure
how good compiler support is for that, either.  It's probably
more supported than mixed declarations and code.  Nevertheless;
we'll trigger memory checkers on exit because we don't free
the buffers; but we won't actually leak because we reuse those
buffers (just like the non-SRC code path).

git-svn-id: https://svn.musicpd.org/mpd/trunk@5397 09075e82-0dd4-0310-85a5-a0d7c8717e4f

Also took out an unnecessary memset in getPlayerErrorStr

git-svn-id: https://svn.musicpd.org/mpd/trunk@5396 09075e82-0dd4-0310-85a5-a0d7c8717e4f

Avoid unnecessary memset to zero, snprintf always puts a
trailing '\0'.  We also have no need to subtract one from the
buffer we're snprintf-ing it to.

We also check the return value of snprintf to ensure it's not
too long.  I have a feeling we might as well avoid snprintf
altogether so we don't have to worry about buffer sizing/stack
overflow and just do a bunch of write(2)s, letting Nagle sort it
out...

Also, centralize some of the exit error handling in with
goto.  This makes the code a bit more consistent and
maintainable as well as reducing code and binary size.

git-svn-id: https://svn.musicpd.org/mpd/trunk@5395 09075e82-0dd4-0310-85a5-a0d7c8717e4f

We need to identify ourselves as HTTP/1.1 so Range: works;
and so the server can return HTTP/1.1 instead of HTTP/1.0.

Tested against lighttpd 1.4.13

git-svn-id: https://svn.musicpd.org/mpd/trunk@5394 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5378 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5376 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5374 09075e82-0dd4-0310-85a5-a0d7c8717e4f

implementation, and fixing it is a big enough job that I don't know when
I'll get around to it.  Probably best just starting from scratch anyhow.

git-svn-id: https://svn.musicpd.org/mpd/trunk@5373 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5369 09075e82-0dd4-0310-85a5-a0d7c8717e4f

this fixes it though.

git-svn-id: https://svn.musicpd.org/mpd/trunk@5338 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5335 09075e82-0dd4-0310-85a5-a0d7c8717e4f

git-svn-id: https://svn.musicpd.org/mpd/trunk@5333 09075e82-0dd4-0310-85a5-a0d7c8717e4f

#2) fix a deadlock condition when attempting to seek if the decoder quit and returned to playerInit()

git-svn-id: https://svn.musicpd.org/mpd/trunk@5325 09075e82-0dd4-0310-85a5-a0d7c8717e4f