Rocklua was using the full fscanf implementation
to simply read %ld for the file:read("*n") function
wasting 1k on unneeded/unused functionality
Instead, I've implemented a filetol function to duplicate it
without the extra overhead using strtol which as an added bonus
ERANGE errors now resolve to LONG_MIN and LONGMAX instead of
integer overflow
filetol()
reads long int from an open file, skips preceding
whitespaces returns -1 if error, 1 on success.
*num set to LONG_MAX or LONG_MIN on overflow.
If number of digits is > than LUAI_MAXNUMBER2STR
filepointer will continue till the next non digit
but buffer will stop being filled with characters.
Preceding zero is ignored.
Change-Id: Ia42d0f73c63a894625bca4581e9b7e1cc7387fd2
Some of the lcd functions had support for multiple screens but this wasn't
very safe since the screen number wasn't bounded within the screens[] array
This adds support for all the lcd functions along with checking that screen#
is bounded properly, adds around 600 bytes to devices with a remote screen
devices without a remote screen lock to SCREEN_MAIN
Change-Id: I618bbc7b3919c7b0ff375fb2d71949d7cab43c87
rli_marshal and rli_copy allow the use of a custom lua function
instead of the built-ins, this custom function can signal to stop
before the whole image is iterated.
Originally it was checking for a return of 0
which was additional overhead for the built-in functions
(which never stop early) as well.
Now custom_transform sets the x & y deltas to 0 to indicate early
exit removing an extra 'if' per cycle and return from all of the
rli_transform functions
The lcd_xxx_bitmap functions all require x, y, w, h, screen these calls
have been consolidated into a single function get_bmp_bounds()
Change-Id: I88de3149c58d1bfb40e9d1a91341fb86cbd63b51
In particular, this solves seeking glitches seen in ~6 hr mp3 files.
(Patch taken from Igor Poretsky's tree)
Change-Id: Id65b6726146b6d2d1a223e90b88e401d1b2d597a
While luaL_addstring() works perfectly fine for the
final buffer once Lual_pushresult() is called
kbd_input doesn't display a previous input properly
since the buffer hasn't been finalized yet
Change-Id: Ic50acaa8b7b17077dec8750dd2b6382624e8a549
Removes unneeded functions from kbd_input
Consolidates message filling function for gui_syncyesno_run & do_menu
Change-Id: If3c3cea3cbf37a8dc52983c0db174de6d54b35f8
The way to call the playlist functions has changed
rb.playlist("option", var)
rb.playlist_add(filename) = becomes rb.playlist("add", filename)
added playlist.lua to the includes for conversion to old functions
if your script is broken by this change you simply add `require("playlist")`
to the top for the old functionality
added rb.playlist_tracks(dir, filename) to playlist.lua
this will allow you to add all tracks in a playlist.m3u8
to a lua table
Change-Id: I87fcc56be365d8495d214f069331b6ddbfbef1db
audio_play was removed from the rocklib I assume due to inconsistent
behavior I've readded it with a check for audio paused which instead
uses rewind/ff and then resumes audio
the way to call the audio functions has changed as well
rb.audio("option", var)
so rb.audio_play(0, 0) becomes rb.audio("play", 0, 0)
audio_audio_flush_and_reload_tracks becomes
rb.audio("flushandreloadtracks")
all functions except audio("getfilepos") return the previous (or still current)
status
added audio.lua to the includes for conversion to old functions
if your script is broken by this change you simply add `require("audio")`
to the top for the old functionality
Change-Id: I364adf0c85d9c12b98cde29c26fbe5ee05b9d331
Both of these functions modified the string in the lua stack
per lua manual:
'When a C function receives a string argument from Lua,
there are only two rules that it must observe: Not to pop the string
from the stack while accessing it and never to modify the string'
strip_extension will still work with old parameters
and is thus backwards compatible
strip_extension("filename")
create_numbered_filename has changed slightly and IS NOT backwards compatible
create_numbered_filename(path, prefix, suffix, [number])
(number defaults to -1)
Change-Id: I34cf7e2f6f691f33d5ac2b2e995855a171fb99b3
allows rocklib_img to be excluded if needed
stops rocklib_aux from generating redundant prototypes for
lcd_mono_bitmap[_part]
Change-Id: Ie208ad71ab5f9a7deb026dc01a5b0a0631a0d29c
is_diacritic is used for pretty much all of the text display sizing
Timing here matters, so important it uses a
priority buffer in order to cut down on searching the database.
This patch is verified functionally equlivalent but
saves 80 bytes and is ~20% faster for the clip+ I tested it on
Several things were done:
Optimizing the copy loop for the priority buffer
Globbing some operations together
but the main one has to do with the database structure its self
--
Rather than a bit packed struct it combines the lower 8 bytes into
one value [info] and uses flags to access the values instead
Change-Id: I475073419b647639ef1eeaa3b9213e4e1067c0db
change offending bool return to int
warning: cast between incompatible function types from
'_Bool (*)(void)' to 'int (*)(void)' [-Wcast-function-type]
forgot to remove -- typedef int (*menu_function)(void);
Change-Id: Ie4c8d3ddb0fb7843c4ec584203350d658d6bee3e
rocklib_aux is auto generated from plugin.h
there are a few functions that get added automatically that
are unusable without their companion functions or duplicate
functionality already supplied by lua
Duplicated functionality:
rb->rand, rb->srand
-- see math.rand math.srand
rb->remove, rb->rename
-- see os.remove, os.rename
Unusable:
rb->open_utf8
-- this should be added to the lua file open routines (if at all)
rb->codec_run_proc, rb->codec_close
-- without rb->codec_load_file these are pointless
rb->timer_set_period, timer_unregister
-- even with timer_register implemented lua is not
-- reentrant and crashes the state when timer fires
Shouldn't be used!:
rb->strlcpy, rb->strlcat, rb->strcpy, rb->strcat
-- lua reuses strings by hashed values you break this contract if
-- you change strings behind its back plus lua provides a way to
-- do these functions safely within the strings api
Change-Id: I2f65704a90930378cbbceb254e52f61e8074471e
int is an incompatible type for targets that have 32bit fb_data
need to use FB_SCALARPACK for them
Change-Id: Ib3b5ff19c54d8d1bb76af33d0538a17a71301514
removes tslf allocations from do_menu and gui_syncyesno_run in favor of
lua_newuserdata
removes some luaL_opt functions in favor of equivalent lua_to functions
moves some definitions to the rocklib.h file
Change-Id: Iaacc3249b8f1af2c220ce59dead0050c66cb3b04
- increase decoder thread stack size just enough (2KiB) to
avoid stack overflows when seeking in opus files
- only do so on devices with at least 8MiB of RAM
Change-Id: I7e7182ae866338b4aad6ed7e32391ddd667121bd
* Fully Interrupt-driven, with proper task yielding
* Much more robust error handling
* Eliminate duplicate code
* Pile of bugfixes
(Much of this adapted from Igor Poretsky's tree)
Change-Id: I46006412323cba2088b70094635d62a241be1d7e
swr/swl instructions used for word aligning were wrong. This
made memset() terribly broken. I can't imagine how it went
uncaught for soooo long. Spotted by Solomon Peachy.
I run unit tests for alignments 0,1,2,3
size 1, 2, 3, 4, 5, 63, 64, 65, 127, 128, 129;
and fill pattern 0x00 and other (since 0 is special case in this
implementation).
Change-Id: I513a10734335fe97734c10ab5a6c3e3fb3f4687a
On Classic, IRAM1 (second 128Kb of a total of 256KB available IRAM) is
slower than DRAM. Codecs that actually are using regions of IRAM1 runs
faster when DRAM is used, so IRAM1 is disabled and only IRAM0 remains
enabled: 48KB for core and 80KB for codecs/plugins.
The next test_codec results shows how decode time is decreased:
file boosted unboosted
*.ra ~1.5% ~0.5%
*.mpc ~21% ~4.5%
*.ogg ~0.5% ~0%
nero_he*.m4a ~8% ~1%
nero*.m4a ~25% ~7%
wmapro*.wma ~4.5% ~0%
wma*.wma ~25% ~7%
In addition there is a small power save when IRAM1 HW is disabled.
Change-Id: I102adee11458e82037f23076d5d5956e23235de8
For s5l8701, s5l8702 and as3525v2 targets.
The crash occurs when USB is inserted or extracted while the playlist
is being loaded or updated (it could take a few seconds for huge
playlists), at this point all buflib memory is allocated and not freed
before usb_starage_init_connection() is executed.
This workaround mitigates this panic by using static memory for USB
buffers, so this memory cannot be used for other tasks, in addition
the problem still persist when playlist load is 'paused' by USB
insertion and then updated after USB extraction.
Change-Id: Iff1db5a949361fd543e0b494924d1f2906c84b5e
