Compress all translated strings with Huffman coding. #1121
Conversation
This saves code space in builds which use link-time optimization. The optimization drops the untranslated strings and replaces them with a compressed_string_t struct. It can then be decompressed to a c string. Builds without LTO work as well but include both untranslated strings and compressed strings. This work could be expanded to include QSTRs and loaded strings if a compress method is added to C. Its tracked in micropython#531.
| @@ -242,7 +242,7 @@ freedos: | |||
|
|
|||
| # build an interpreter for coverage testing and do the testing | |||
| coverage: | |||
| $(MAKE) V=2 \ | |||
| $(MAKE) \ | |||
There was a problem hiding this comment.
check: this change is on purpose?
There was a problem hiding this comment.
Ya, it reduces the verbosity which is my preference.
| @@ -102,20 +102,6 @@ const mp_obj_module_t mp_module_uerrno = { | |||
| }; | |||
|
|
|||
| const char* mp_errno_to_str(mp_obj_t errno_val) { | |||
There was a problem hiding this comment.
should this call mp_common_errno_to_str()? or remove gutted function so it doesnt get called by accident?
There was a problem hiding this comment.
It can't because we only want to allocate to the stack. decompressing here would require a heap allocation since the stack disappears on return. I think we do really want to call both.
There was a problem hiding this comment.
np - plz grep to check all calls to this are now calling mp_common_ :)
There was a problem hiding this comment.
Done!
/V/E/circuitpython (huffman|✔) $ ag -Q mp_errno_to_str .
py/moduerrno.c
104:const char* mp_errno_to_str(mp_obj_t errno_val) {
127:const char* mp_errno_to_str(mp_obj_t errno_val) {
py/mperrno.h
143:const char* mp_errno_to_str(mp_obj_t errno_val);
py/objexcept.c
124: msg = mp_errno_to_str(o->args->items[0]);
| const char* msg = mp_errno_to_str(o->args->items[0]); | ||
| const compressed_string_t* common = mp_common_errno_to_str(o->args->items[0]); | ||
| const char* msg; | ||
| char decompressed[50]; |
There was a problem hiding this comment.
50 is arbitrary since we don't know the length yet. That's why its checked below against the actual length (which includes NULL). It has to be declared in this scope so it survives down to the print call below.
There was a problem hiding this comment.
50 is ok, we subtract 1 here
https://github.com/adafruit/circuitpython/pull/1121/files#diff-3c0b0b9ab47034d1e9ce2494287f5b68R70
This saves code space in builds which use link-time optimization.
The optimization drops the untranslated strings and replaces them
with a compressed_string_t struct. It can then be decompressed to
a c string.
Builds without LTO work as well but include both untranslated
strings and compressed strings.
This work could be expanded to include QSTRs and loaded strings if
a compress method is added to C. Its tracked in #531.