Do as with fdprintf and avoid filesystem calls for every single
byte. If it overreads, just put the excess back with lseek, which
does no I/O itself.
Change-Id: Ifd5d21b5dca7183346e44d365d3f7d45e8cc6438
Paths are stored after the structure at their actual length plus
any aligment padding. In principle, any type of auxilliary data
could go there.
Change-Id: Ic5487dc4089781b5cc52414d1691ba6d9dc1893c
The location of the handle cannot be kept across calls to
shrink_handle() since it may move the structure. The error was
there in one place at the inception, corrected, then reintroduced.
Make shrink_handle() return the new location and use it, which
makes the side effects of the function clearer.
Change-Id: Icae6a0ad6f7bb0d6645b044cccfa4aef88db42ad
There's no need for it any longer since the list is now doubly-
linked. As a bonus, stack limits pose no barrier to the length of
the list.
Change-Id: I41c567f946b640ef1e3c2d93da2f5aef9a763c66
Does away the statically-allocated track list which frees quite
a fair amount of in-RAM size.
There's no compile-time hard track limit.
Recommended TODO (but not right away): Have data small enough use
the handle structure as its buffer data area. Almost the entire
handle structure is unused for simple allocations without any
associated filesystem path.
Change-Id: I74a4561e5a837e049811ac421722ec00dadc0d50
Just use long so the compiler potentially doesn't complain about
use of other values not in the enum. It's also the type used
around the system for event ids.
Increase min codec API version.
No functional changes.
Change-Id: If4419b42912f5e4ef673adcdeb69313e503f94cc
If in the middle of a manual skip, playback would try to seek, and
therefore start, the codec before the audio handle was available.
This wasn't really a problem since the codec would just bail out
and be retried later. But, it is a problem for a change I was working
on with seeking where the codec could get caught in a full-speed
loop trying to seek itself (stoppable, not lockup).
The main side effect of this change that you may notice, if using an
HDD with dircache turned on and the disk is not spinning, is that you
can keep holding down prev/next if dir skipping and the WPS will
start FF/RW mode. By the time the new track shows up, you will have
seeked into it some amount.
Well, the PBE is getting the info ASAP anyway and as far as it's
concerned, the next track is under way. On that end of things, it's
correct. Perhaps WPS should lock out its own seek mode at certain
times.
Change-Id: Ifc7409a886df399cec189d1bae2adba3872e857a
It isn't necessary to explicitly stop and restart playback to
force it to update something that must cause rebuffering.
Change-Id: I6ff5394fcafc7374af67ef9fbf9022bb4a79b773
This brings a various small changes to the drawing and input code,
as well as a brand new "mouse mode", where input goes to a virtual
mouse cursor. Only Loopy has this mouse mode enabled by default,
while other games have it hidden away under the debug menu. Some
changes by me to Palisade were required to make it playable; those
are included here as well. Right now, sgt-net is pushing the c200v2's
upper limit on size and may have to be dropped in a future commit.
Change-Id: I495d2a2125462c2985aec1ffbc54bbe3fe5133bd
The new vuprintf makes unnecessary workarounds due to formatting
limitations. I checked grep output for whatever appeared to fit
but it's possible I missed some instances because they weren't
so obvious.
Also, this means sound settings can dynamically work with any
number of decimals rather than the current assumption of one or
two. Add an ipow() function to help and take advantage of dynamic
field width and precision. Consolidate string formatting of sound
settings.
Change-Id: I46caf534859dfd1916cd440cd25e5206b192fcd8
FIX - softlock not allowing power button to shutdown device when active
FIX - autolock not working unless Selective backlight enabled
Rework - get_action_worker split into more managable chunks
Rework - moved all static variables to a single struct action_last
to be passed through all internal functions
Rework - current action state moved to struct action_current
to be initialized at each get_action /get_action_custom call
Rework - action lookup in button maps changed slightly to prefer fully formed actions
rather than blindly returning action with BUTTON_NONE as pre_button
this should allow actions defined in keymaps positional independence
Rework / Clean-up - moved conditionals on internal functions to inside the functions
-Allows much cleaner code & empty functions should be optimized
out by compiler
Rework / Clean-up - softlock and backlight exemptions code
Clean-up - remove a lot of (now) unused defines
Clean-up - organize / move all exported functions to their own section in the file
Comments added to all internal functions to better explain their purpose
Change-Id: I0ac02d1a2b45826843611513fea3c20a819d2fed
- enable various features in radio screen (like mode switch)
- improve keyboard (short play to select, long play to finish, power to abort)
- enable volume keys in radio screen if available
Change-Id: I01e71291cfd783763ffd55d4fa60610b6900c726
None of the Sony up to A15 seem to support RDS (they use either Si4702 or Si4708),
thus I did not add any code to support RDS.
Change-Id: I64838993b9705b36b94665f8470c7a89c772c961
Both blitter_save() and blitter_load() functioned incorrectly when
zoomed in -- blitter_save() would copy from the wrong location, and
blitter_load() would ignore the y-coordinate of the destination.
Change-Id: I7c85debf5953575f72c4a81e3dbcf514202c3aed
This adds no functionality, but instead cleans up the source and adds
some more substantial documentation.
Change-Id: I77328c171a61db7729bdf928ba094cfbed4ec0dd
This fixes two separate bugs: The first was that the drawmode was not
reliably set when drawing the title, occasionally leading to a black
rectangle where the title should be; the second that the title bar's
space wasn't being cleared before drawing, leading to artifacts when
the length of the title string changed.
Change-Id: I3859125b14d568e1098b095ab134645a504b2d45
This makes it possible to play the game while zoomed in. Read the
manual entry if you want to know more.
Change-Id: Iff8bab12f92ebd2798047c25d1fde7740aa543ce
Adds boot data to as3525 devices Sansa C200v2 E200v2 Clip Clipv2 Clip+ ClipZip
fuze, fuzev2 m200v4
Adds boot_data to features.txt
default arm crt0.s now had boot data if HAVE_BOOTDATA is defined
Change-Id: I614a556696540511a69fc12a4520b01c268bf8a9
Bootdata is a special location in the Firmware marked by a magic header
The bootloader is able to copy information to the firmware by locating
this struct and passing data to the firmware when it is loaded but
before it is actually executed
Data is verified by a crc of the bootdata
Change-Id: Ib3d78cc0c3a9d47d6fe73be4747a11b7ad6f0a9e
This adds colored font rendering, as well as a workaround for font
loading while zoomed. Additionally, the frontend has been modified to
match the new upstream API.
Change-Id: I8c3fe57e6854f176485bf792cf4778cd54a21674
This adds a "Zoom In" option to the pause menu of each puzzle, which
displays the puzzle at triple size (subject to change). This should
help with tiny screens, modulo memory concerns associated with
allocating the temporary framebuffer to which drawing operations are
redirected. Coincidentally, there's an upstream bug with Map that
causes the cursor's positioning to be incorrectly displayed when
zoomed.
Change-Id: Ic8b7c2942acf558e295f4271dd7dc458cd336895
One mustn't assume a plugin will only call plugin_get_audio_buffer one
time or that the buffer_size pointer is always non-NULL. At least one
plugin, pacbox, will call it each time it (re)starts audio, with a NULL
param (which is intentional because it only wants to kill audio
playback), and leak away all the RAM because the handle gets clobbered
by further calls and the memory can't be released.
Change-Id: Ic5b94dbc0277c42964ea85b4e9d0302a7c6f1fe4
Playlist was CRC-ing the path from the id3, which may have been
modified to remove "bogus dirs". This would cause a CRC mismatch
in the resume information.
Now, just use the current playlist's current index and call
playlist_get_filename_crc32() to get the original path when
updating resume info.
While technically correct, if this causes any issue(s) it's just
a one-line change and painless to revert.
Change-Id: Ie595ef6c40349c342bd7acac8c542829f9cd5d76
sonynwz: quirk for cpufreq broken driver
There was some redundancy between frequency_linux(cpu, true) and
current_scaling_frequency(), also I see no reason to compile the cpuinfo stuff
unconditionally and the scaling info only on DX since it was already printed
some partial scaling info anyway. Thus compile all the code unconditionally
and simplify the logic in the debug menu. Also avoid putting buffers of size
PATH_MAX on stack since it can be quite big and we only requires 64 bytes
for those paths.
On Sony NWZ, the cpu driver reports frequency in MHz instead of kHz thus we need
to make the cpuinfo code aware of that bug.
Change-Id: I61af45ab5f179ecc909b4841b9137a915a60193a
Simply extends the current isqrt() to be able to do fractional bits
and improves the initial estimate using clz(). iqrt() itself is
no more and is equivalent to fp_sqrt(x, 0). The original also had
a small bug where the guess comparision should have been >=, not >.
Uses no large integer math or division and is very accurate
(simply returns a truncated fraction).
Change-Id: I2ae26e6505df1770dc01e56220f7385369f90ae9
Aims to provide a lib/keymaps.h for plugins needing simple button
functionality beyond that provided by PLA. Currently used by puzzles
and xworld.
Change-Id: Icb3493aaf176d401762de834dd48fc76a3824c5a
The bug is due to a stupid make misfeature. The article [1] contains much more
information but in a nutshell, the following code:
a b: c
bla
is equivalent to:
a: c
bla
b: c
bla
This is bad because in parallel runs (make -j typically), "bla" can be run
TWICE and even worse, twice in PARALLEL. Obviously the result will be
completely unexpected. This is a real bummer because on the other hand,
the following code:
%.c %.h: %:in
bla
actually expresses the fact that bla produces two files. For some reasons,
pattern rules work differently from implicit rules.
This commit attempts to fix the problem with lang.h by rewriting (simplified):
lang.c lang.h: lang.in
genlang
as
lang.h: lang.in
genlang
lang.c: lang.h
This works (it correctly expresses the dependency chain and ensures genlang runs
once) but as one drawback: if one manually removes lang.c, then genlang will not
re-run since the second rule does nothing. This is minor drawback since no one
ever removes lang.c manually and "clean" removes lang.h which triggers a rebuild.
[1]: https://www.gnu.org/software/automake/manual/html_node/Multiple-Outputs.html
Change-Id: Ic0bf7c7c203dc599b00fde457946d2316c70474e
Hopefully this will help narrow down the problem with LANG_* error message,
I have the feeling that lang.h is not generated in time but it's very hard to
reproduce.
Change-Id: I02b6c98ed9c7e7168fad10dcf1142e307fbc6093
Keylock is now toggled by the key combo User&Power for WPS and FMS.
Pitchscreen has been added as long press of User.
While in bookmark screen, short press of Power deletes bookmark.
While in (most) menus or tree, first press of MENU takes you back to the
main menu, second press to WPS (aka button_context_tree).
Removed ACTION_WPSAB_SINGLE as long press of USER is now Pitchscreen.
Change-Id: I6aee7f8655b1073ed8d74dabe73895a7c0e54bfe
