Use code unit instead of code point #961
Conversation
| lit_read_code_point_from_hex (lit_utf8_byte_t *buf_p, /**< buffer with characters */ | ||
| lit_utf8_size_t number_of_characters, /**< number of characters to be read */ | ||
| lit_code_point_t *out_code_point_p) /**< [out] decoded result */ | ||
| lit_read_code_unit_from_hex (lit_utf8_byte_t *buf_p, /**< buffer with characters */ |
There was a problem hiding this comment.
The doc still refers to code points.
There was a problem hiding this comment.
With the help of the non-character markers (as in my previous comment), couldn't we just return a code unit from this function and compare it to U+FFFF to see whether we succeeded or not? (I'm really against out parameters given with pointers, so prefer to replace them wherever possible.)
There was a problem hiding this comment.
No. The helper must be able to read "FFFF" and does not have to deal with it a valid code unit or not. Currently it might work on the test suite but this change would be dangerous.
There was a problem hiding this comment.
I just sweeped through the call sites of this function, and it seems that only three functions call it with number_of_characters being 4: ecma_builtin_json_parse_string, re_parse_char_class, and re_parse_next_token. (Where the param is 2, there can be no possible way of getting a valid value of 0xFFFF.) Also, in a valid json, there must not be a U+FFFF according to the unicode specs. Where I'm not completely sure is the regexes. :/ While googling around I've seen some artificial examples with regex character classes where the upper limit was \uFFFF. What I couldn't find out with certainty is whether they are valid examples.
There was a problem hiding this comment.
JS supports all characters between 0-0xffff. JS was not designed to be UTF compatible, every character is a two byte number.
LaszloLango
force-pushed
the
code-unit-opt
branch
from
5aa4211 to
0698912
Compare
|
@akiss77, @zherczeg, updated. Please check. |
|
@zherczeg, related discussion: #938 (diff) |
|
Ok. LGTM |
JerryScript-DCO-1.0-Signed-off-by: László Langó [email protected]
LaszloLango
force-pushed
the
code-unit-opt
branch
from
0698912 to
e1f20ad
Compare
JerryScript-DCO-1.0-Signed-off-by: László Langó [email protected]