jepler/circuitpython
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
circuitpython/py
History
Jim Mussared 154b4eb354 py: Implement "common word" compression scheme for error messages. 
The idea here is that there's a moderate amount of ROM used up by exception
text.  Obviously we try to keep the messages short, and the code can enable
terse errors, but it still adds up.  Listed below is the total string data
size for various ports:

    bare-arm 2860
    minimal 2876
    stm32 8926  (PYBV11)
    cc3200 3751
    esp32 5721

This commit implements compression of these strings.  It takes advantage of
the fact that these strings are all 7-bit ascii and extracts the top 128
frequently used words from the messages and stores them packed (dropping
their null-terminator), then uses (0x80 | index) inside strings to refer to
these common words.  Spaces are automatically added around words, saving
more bytes.  This happens transparently in the build process, mirroring the
steps that are used to generate the QSTR data.  The MP_COMPRESSED_ROM_TEXT
macro wraps any literal string that should compressed, and it's
automatically decompressed in mp_decompress_rom_string.

There are many schemes that could be used for the compression, and some are
included in py/makecompresseddata.py for reference (space, Huffman, ngram,
common word).  Results showed that the common-word compression gets better
results.  This is before counting the increased cost of the Huffman
decoder.  This might be slightly counter-intuitive, but this data is
extremely repetitive at a word-level, and the byte-level entropy coder
can't quite exploit that as efficiently.  Ideally one would combine both
approaches, but for now the common-word approach is the one that is used.

For additional comparison, the size of the raw data compressed with gzip
and zlib is calculated, as a sort of proxy for a lower entropy bound.  With
this scheme we come within 15% on stm32, and 30% on bare-arm (i.e. we use
x% more bytes than the data compressed with gzip -- not counting the code
overhead of a decoder, and how this would be hypothetically implemented).

The feature is disabled by default and can be enabled by setting
MICROPY_ROM_TEXT_COMPRESSION at the Makefile-level.
2020-04-05 14:20:57 +10:00
..
argcheck.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
asmarm.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
asmarm.h py: Add support to save native, viper and asm code to .mpy files. 2019-03-08 15:53:05 +11:00
asmbase.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
asmbase.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
asmthumb.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
asmthumb.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
asmx64.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
asmx64.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
asmx86.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
asmx86.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
asmxtensa.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
asmxtensa.h py/asmxtensa: Add support for Xtensa with windowed registers. 2019-10-05 13:44:08 +10:00
bc.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
bc.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
bc0.h py/bc0.h: Shift comment to start of line to improve format consistency. 2020-02-28 10:29:32 +11:00
binary.c all: Fix implicit casts of float/double, and signed comparison. 2020-03-30 12:04:21 +11:00
binary.h py/binary: Change mp_uint_t to size_t for index, size, align args. 2019-09-02 13:14:27 +10:00
builtin.h extmod/uasyncio: Add optional implementation of core uasyncio in C. 2020-03-26 01:25:45 +11:00
builtinevex.c py/builtinevex: Support passing in a bytearray/buffer to eval/exec. 2020-02-28 12:45:36 +11:00
builtinhelp.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
builtinimport.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
compile.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
compile.h py: Add global default_emit_opt variable to make emit kind persistent. 2019-08-28 12:47:58 +10:00
dynruntime.h py/dynruntime.mk: Set MICROPY_ENABLE_DYNRUNTIME instead of per module. 2020-04-05 14:13:02 +10:00
dynruntime.mk py/dynruntime.mk: Set MICROPY_ENABLE_DYNRUNTIME instead of per module. 2020-04-05 14:13:02 +10:00
emit.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
emitbc.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
emitcommon.c py/scope: Optimise scope_find_or_add_id to not need "added" arg. 2018-10-28 00:38:18 +11:00
emitglue.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
emitglue.h py: Integrate sys.settrace feature into the VM and runtime. 2019-08-30 16:44:12 +10:00
emitinlinethumb.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
emitinlinextensa.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
emitnarm.c py/emitnative: Add support for return/break/continue in try and with. 2018-09-04 14:31:28 +10:00
emitnative.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
emitnthumb.c py/emitnative: Add support for return/break/continue in try and with. 2018-09-04 14:31:28 +10:00
emitnx64.c py/emitnative: Add support for return/break/continue in try and with. 2018-09-04 14:31:28 +10:00
emitnx86.c py/nativeglue: Add new header file with native function table typedef. 2019-12-12 20:15:28 +11:00
emitnxtensa.c py/emitnative: Add support for return/break/continue in try and with. 2018-09-04 14:31:28 +10:00
emitnxtensawin.c py: Add new Xtensa-Windowed arch for native emitter. 2019-10-05 13:44:53 +10:00
formatfloat.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
formatfloat.h all: Use the name MicroPython consistently in comments 2017-07-31 18:35:40 +10:00
frozenmod.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
frozenmod.h various: Update early copyright years to match actual edit history. 2019-05-17 18:06:11 +10:00
gc.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
gc.h py/gc: Adjust gc_alloc() signature to be able to accept multiple flags. 2018-12-20 17:52:16 +11:00
grammar.h all: Add *FORMAT-OFF* in various places. 2020-02-28 10:31:07 +11:00
lexer.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
lexer.h py: Remove 3 obsolete commented-out lines from header files. 2019-11-26 21:36:41 +11:00
makecompresseddata.py py: Implement "common word" compression scheme for error messages. 2020-04-05 14:20:57 +10:00
makemoduledefs.py all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
makeqstrdata.py all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
makeqstrdefs.py py: Implement "common word" compression scheme for error messages. 2020-04-05 14:20:57 +10:00
makeversionhdr.py all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
malloc.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
map.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
misc.h py: Implement "common word" compression scheme for error messages. 2020-04-05 14:20:57 +10:00
mkenv.mk py/mkenv.mk: Move usage of 32-bit flags to py.mk. 2020-01-12 10:34:10 +11:00
mkrules.mk py: Implement "common word" compression scheme for error messages. 2020-04-05 14:20:57 +10:00
modarray.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
modbuiltins.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
modcmath.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
modcollections.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
modgc.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
modio.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
modmath.c all: Convert exceptions to use mp_raise_XXX helpers in remaining places. 2020-03-18 17:26:19 +11:00
modmicropython.c py/modmicropython: Add heap_locked function to test state of heap. 2020-03-11 16:54:16 +11:00
modstruct.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
modsys.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
modthread.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
moduerrno.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
mpconfig.h extmod/uasyncio: Add optional implementation of core uasyncio in C. 2020-03-26 01:25:45 +11:00
mperrno.h py/mperrno: Define MP_EWOULDBLOCK as EWOULDBLOCK, not EAGAIN. 2018-05-01 15:53:25 +10:00
mphal.h lib/utils/sys_stdio_mphal: Add support to poll sys.stdin and sys.stdout. 2019-07-01 17:10:12 +10:00
mpprint.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
mpprint.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
mpstate.c all: Use the name MicroPython consistently in comments 2017-07-31 18:35:40 +10:00
mpstate.h py/qstr: Don't include or init qstr_mutex when GIL is enabled. 2020-01-23 13:29:11 +11:00
mpthread.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
mpz.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
mpz.h tools/codeformat.py: Eliminate need for sizeof fixup. 2020-03-11 14:34:40 +11:00
nativeglue.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
nativeglue.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
nlr.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
nlr.h all: Add *FORMAT-OFF* in various places. 2020-02-28 10:31:07 +11:00
nlrpowerpc.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
nlrsetjmp.c py/nlr: Factor out common NLR code to macro and generic funcs in nlr.c. 2017-12-28 16:46:30 +11:00
nlrthumb.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
nlrx64.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
nlrx86.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
nlrxtensa.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
obj.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
obj.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objarray.c all: Fix implicit casts of float/double, and signed comparison. 2020-03-30 12:04:21 +11:00
objarray.h py/objarray: Introduce "memview_offset" alias for "free" field of object 2018-12-20 17:40:48 +11:00
objattrtuple.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objbool.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objboundmeth.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objcell.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objclosure.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objcomplex.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objdeque.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objdict.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objenumerate.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objexcept.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objexcept.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objfilter.c py: Downcase all MP_OBJ_IS_xxx macros to make a more consistent C API. 2019-02-12 14:54:51 +11:00
objfloat.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objfun.c tools/codeformat.py: Eliminate need for sizeof fixup. 2020-03-11 14:34:40 +11:00
objfun.h py: Integrate sys.settrace feature into the VM and runtime. 2019-08-30 16:44:12 +10:00
objgenerator.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objgenerator.h all: Use the name MicroPython consistently in comments 2017-07-31 18:35:40 +10:00
objgetitemiter.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objint.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objint.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objint_longlong.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objint_mpz.c all: Remove spaces inside and around parenthesis. 2020-03-28 23:36:44 +11:00
objlist.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objlist.h py/obj.h: Move declaration of mp_obj_list_init to objlist.h. 2018-03-13 14:03:15 +11:00
objmap.c py: Downcase all MP_OBJ_IS_xxx macros to make a more consistent C API. 2019-02-12 14:54:51 +11:00
objmodule.c extmod/uasyncio: Add optional implementation of core uasyncio in C. 2020-03-26 01:25:45 +11:00
objmodule.h py/objmodule.h: Remove obsolete mp_builtin_module_weak_links_map decl. 2020-02-11 15:43:13 +11:00
objnamedtuple.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
objnamedtuple.h py/objnamedtuple: Allow to reuse namedtuple basic functionality. 2017-11-20 09:30:06 +02:00
objnone.c py/obj: Add MICROPY_OBJ_IMMEDIATE_OBJS option to reduce code size. 2020-01-13 01:01:45 +11:00
objobject.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objpolyiter.c all: Remove inclusion of internal py header files. 2017-10-04 12:37:50 +11:00
objproperty.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objrange.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objreversed.c py: Downcase all MP_OBJ_IS_xxx macros to make a more consistent C API. 2019-02-12 14:54:51 +11:00
objset.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objsingleton.c py/objsingleton: Use mp_generic_unary_op for singleton objects. 2019-12-27 12:53:36 +11:00
objslice.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objstr.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
objstr.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objstringio.c py/objstringio: Expose tell() on StringIO and BytesIO objects. 2020-03-11 14:43:03 +11:00
objstringio.h py/objstringio: If created from immutable object, follow copy on write policy. 2017-06-09 17:33:01 +03:00
objstrunicode.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objtuple.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objtuple.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
objtype.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
objtype.h py/runtime: Don't allocate iter buf for user-defined types. 2019-12-27 12:34:22 +11:00
objzip.c py: Downcase all MP_OBJ_IS_xxx macros to make a more consistent C API. 2019-02-12 14:54:51 +11:00
opmethods.c py: Make mp_obj_get_type() return a const ptr to mp_obj_type_t. 2020-01-09 11:25:26 +11:00
pairheap.c py/pairheap: Properly unlink node on pop and delete. 2020-03-26 01:21:04 +11:00
pairheap.h py/pairheap: Add helper function to initialise a new node. 2020-03-26 01:21:04 +11:00
parse.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
parse.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
parsenum.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
parsenum.h all: Use the name MicroPython consistently in comments 2017-07-31 18:35:40 +10:00
parsenumbase.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
parsenumbase.h all: Use the name MicroPython consistently in comments 2017-07-31 18:35:40 +10:00
persistentcode.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
persistentcode.h py/persistentcode: Make ARM Thumb archs support multiple sub-archs. 2019-12-12 20:15:28 +11:00
profile.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
profile.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
py.mk py: Implement "common word" compression scheme for error messages. 2020-04-05 14:20:57 +10:00
pystack.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
pystack.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
qstr.c py: Implement "common word" compression scheme for error messages. 2020-04-05 14:20:57 +10:00
qstr.h py: Implement "common word" compression scheme for error messages. 2020-04-05 14:20:57 +10:00
qstrdefs.h all: Add *FORMAT-OFF* in various places. 2020-02-28 10:31:07 +11:00
reader.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
reader.h py: Allow lexer to raise exceptions during construction. 2017-03-14 11:52:05 +11:00
repl.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
repl.h all: Use the name MicroPython consistently in comments 2017-07-31 18:35:40 +10:00
ringbuf.c py/ringbuf: Add peek16 method. 2019-11-21 12:04:53 +11:00
ringbuf.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
runtime.c py: Use preprocessor to detect error reporting level (terse/detailed). 2020-04-05 14:11:51 +10:00
runtime.h all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
runtime0.h py/nativeglue: Add new header file with native function table typedef. 2019-12-12 20:15:28 +11:00
runtime_utils.c py: mp_call_function_*_protected(): Pass-thru return value if possible. 2017-12-05 00:38:41 +02:00
scheduler.c py/scheduler: Move clearing of kbd traceback to mp_keyboard_interrupt. 2020-02-07 16:08:31 +11:00
scope.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
scope.h py/scope: Optimise scope_find_or_add_id to not need "added" arg. 2018-10-28 00:38:18 +11:00
sequence.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
showbc.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
smallint.c all: Use the name MicroPython consistently in comments 2017-07-31 18:35:40 +10:00
smallint.h py: Extend nan-boxing config to have 47-bit small integers. 2017-12-11 22:39:12 +11:00
stackctrl.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
stackctrl.h all: Use the name MicroPython consistently in comments 2017-07-31 18:35:40 +10:00
stream.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
stream.h py/stream.h: Include sys/types.h to get size_t and off_t for POSIX API. 2020-03-25 01:00:52 +11:00
unicode.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
unicode.h py/objstr: Add check for valid UTF-8 when making a str from bytes. 2017-09-06 16:43:09 +10:00
vm.c all: Add *FORMAT-OFF* in various places. 2020-02-28 10:31:07 +11:00
vmentrytable.h all: Add *FORMAT-OFF* in various places. 2020-02-28 10:31:07 +11:00
vstr.c all: Reformat C and Python source code with tools/codeformat.py. 2020-02-28 10:33:03 +11:00
warning.c py: Update my copyright info on some files. 2019-02-06 00:19:00 +11:00