Except the raw MIDI queues, since they're MIDI specific

Signed-off-by: falkTX <falktx@falktx.com>

Also document this usecase

Signed-off-by: falkTX <falktx@falktx.com>

Signed-off-by: falkTX <falktx@falktx.com>

Signed-off-by: falkTX <falktx@falktx.com>

Found via `codespell -q 3`

* Set MIDI port pretty names on macOS

* Set MIDI port pretty names on Windows

* Set MIDI port pretty names on Linux/alsarawmidi

* Update alsarawmidi port naming to match CoreMIDI and WinMME

* Rename PortSetPrettyNameProperty to PortSetDeviceName

* Set hardware property to MIDI port names

Signed-off-by: falkTX <falktx@falktx.com>

I was sure I did this already...

Signed-off-by: falkTX <falktx@falktx.com>

Thanks for noticing

Signed-off-by: falkTX <falktx@falktx.com>

Signed-off-by: falkTX <falktx@falktx.com>

aften v0.0.8 is latest released tarball.
https://downloads.sourceforge.net/aften/aften-0.0.8.tar.bz2
Travis builds for OSX use aften v0.0.8 (brew).

More recent aften code has introduced an additional parameter to aften_encode_frame().

The check in wscript compiles a small test program to see if the additional parameter is understood.
If additional parameter 'count' is available, set HAVE_AFTEN_NEW_API.

HAVE_AFTEN_NEW_API is used in common/JackAC3Encoder.cpp to call aften_encode_frame()
either with or without 'count' param.

Problem:
int res = aften_encode_frame(&fAftenContext, fAC3Buffer + SPDIF_HEADER_SIZE, fSampleBuffer);
Fails, because API is
AFTEN_API int aften_encode_frame(AftenContext *s, unsigned char *frame_buffer,const void *samples, int count)

The additional parameter 'count' was added to libaften function aften_encode_frame() here:
e1cbb66628 (diff-d4868af7d235a914aef2ec4f979fcb9fR87)

Solves https://github.com/jackaudio/jack2/issues/455

JackTools is a part of the LGPL licenced jacklib and should not pull in
GPL licensed code.

* Fix generation of client uuid

Signed-off-by: falkTX <falktx@falktx.com>

* Remove all client properties when removed; Cleanup a few things

Signed-off-by: falkTX <falktx@falktx.com>

* Make all uuid function arguments use jack_uuid_t type; Fix warnings

Signed-off-by: falkTX <falktx@falktx.com>

* Rework uuids to never be int, more cleanup

Signed-off-by: falkTX <falktx@falktx.com>

Signed-off-by: falkTX <falktx@falktx.com>

Signed-off-by: falkTX <falktx@falktx.com>

Only REG_EXTENDED regular expressions are understood.

Eg. foo_{1|2} is NOT understood.

This commit solves segfault when using invalid regular expressions.
See https://github.com/jackaudio/jack2/issues/433

* Revert "oss_driver: Use float to S32 conversion if requested"

This reverts commit c5a0f5ea1c.

* Revert "alsa_in/out: Use float to S32 conversion if requested"

This reverts commit 148c8d8e6a.

* Revert "alsa_driver: Use float to S32 conversion if requested"

This reverts commit d017e1fffe.

* Revert "memops: Provide function for float to S32 conversion"

This reverts commit bb99e09b99.

* Revert "memops: Align S24LE and S32LE to float conversion"

This reverts commit b4ea23df6a.

* Revert "memops: Provide function for S32 to float conversion"

This reverts commit 244fc27e29.

* Revert "memops: Align float to S24LE and S32LE conversion"

This reverts commit 4455fe020c.

* Revert "memops: Remove not used conversion macros"

This reverts commit a82f3f2fb4.

* Revert "memops: Use right-aligned S24LE to float conversion"

This reverts commit e753254313.

* Revert "memops: Use right-aligned float to S24LE conversion"

This reverts commit 831163e516.

Problem:

#set two properties for UUID 4294967296

$ jack_property -s 4294967296 "a" 1
$ jack_property -s 4294967296 b "2"

#-l lists them

$ jack_property -l
4294967296
key: b value: 2
key: a value: 1

#-l for UUID doesn't list them  <<<<<<

$ jack_property -l 4294967296
$ jack_property -l 4294967296 a
Value not found for a of 4294967296
$ jack_property -l 4294967296 b
Value not found for b of 4294967296

#it seems that 3 chars is the minimum length for value when querying for UUID

$ jack_property -s 4294967296 a 123
$ jack_property -l 4294967296
key: a value: 123
$ jack_property -l 4294967296 a
123

In example-clients/property.c:

/* list all properties for a given UUID */

                        if ((cnt = jack_get_properties (uuid, &description)) < 0) {
                                fprintf (stderr, "could not retrieve properties for %s\n", subject);
                                exit (1);
                        }

cnt is always 0 for values < 3 chars. Why?

In common/JackMetadata.cpp:

int JackMetadata::GetProperty(jack_uuid_t subject, const char* key, char** value, char** type)
int JackMetadata::GetProperties(jack_uuid_t subject, jack_description_t* desc)

This loop gets results:
while ((ret = cursor->get (cursor, &key, &data, DB_NEXT)) == 0) {

but are dropped because of this check:

/* result must have at least 2 chars plus 2 nulls to be valid
if (data.size < 4) {

This rule isn't understood. Explanations are welcome!
Reducing the check to 2 (1 char + null) will consider single char values.
This makes listing properties for a given UUID the same keys as when listing all UUIDs and keys.
No side-effects of lowering the value has been detected (yet).

Note: there is no problem getting "short" values in this method:
int JackMetadata::GetAllProperties(jack_description_t** descriptions)
(as used by jack_property -l without UUID)

jack_property output after change to (data.size < 2):

$ jack_property -s 4294967296 "a" 1
$ jack_property -s 4294967296 b "2"
$ jack_property -l
4294967296
key: b value: 2
key: a value: 1
$ jack_property -l 4294967296
key: b value: 2
key: a value: 1
$ jack_property -l 4294967296 a
1
$ jack_property -l 4294967296 b
2

Note: This change should be considered also for JACK1 (libjack/metadata.c)
to keep implementations in sync.

See https://github.com/jackaudio/jack2/issues/413

Due to inlining and enabled optimization the compiler will not create
machine instructions for the if-clauses and left shift for
sample_move_dS_s32s() and sample_move_dS_s32().

Signed-off-by: Timo Wischer <twischer@de.adit-jv.com>

Due to the optimization of the compiler this commit does not introduce any
differences in the machine instructions.
Tested with GCC version 6.2.0 on ARM64 and x86_64.

Signed-off-by: Timo Wischer <twischer@de.adit-jv.com>