dox4/ruby
1
0
Fork 0
mirror of https://github.com/ruby/ruby.git synced 2026-08-01 02:20:48 +08:00
Code Issues Packages Projects Releases Wiki Activity

Improve performance of String#scrub by skipping ASCII runs with search_nonascii (#16359)

Browse Source 
Improve performance of String#scrub by skipping ascii characters after finding a multi byte character using search_nonascii. Create benchmark for String#scrub to show performance difference made by including the new ascii skips.
This commit is contained in:
FletcherDares 2026-05-26 00:52:41 -04:00 committed by GitHub
parent 678a2ef77f
commit effac0a485
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
Notes: git 2026-05-26 04:53:19 +00:00 
Merged-By: ioquatix <samuel@codeotaku.com>
2 changed files with 60 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

48
benchmark/string_scrub.yml Normal file
View File

@ -0,0 +1,48 @@
prelude: |
STRING_SIZE = 1024
def duplicate_to_length(str, target_length)
return "" if target_length <= 0
return str[0, target_length] if str.length >= target_length
(str * ((target_length / str.length) + 1))[0, target_length]
end
base = "Hello \u{1f600} world! \u{00e9}\u{00f1}"
padding = duplicate_to_length(base, STRING_SIZE)
valid_utf8 = (padding.b + "OK".b).force_encoding("UTF-8")
valid_utf8.valid_encoding?
unknown_but_valid_utf8 = valid_utf8.dup.b.force_encoding("UTF-8")
invalid_utf8 = (padding.b + "\x80\xFF".b).force_encoding("UTF-8")
invalid_utf8.valid_encoding?
unknown_but_invalid_utf8 = (padding.b + "\x80\xFF".b).force_encoding("UTF-8")
worst_case_utf8 = duplicate_to_length("\u{1f600}\u{00e9}\u{00f1}", STRING_SIZE).b.force_encoding("UTF-8")
unknown_but_valid_utf8_worst_case = worst_case_utf8.dup.b.force_encoding("UTF-8")
unknown_but_invalid_utf8_worst_case = (worst_case_utf8.b + "\x80\xFF".b).force_encoding("UTF-8")
benchmark:
scrub_known_valid: |
string = valid_utf8.dup
string.scrub!
scrub_known_invalid: |
string = invalid_utf8.dup
string.scrub!
scrub_unknown_but_valid_coderange: |
string = unknown_but_valid_utf8.dup
string.scrub!
scrub_unknown_and_invalid_coderange: |
string = unknown_but_invalid_utf8.dup
string.scrub!
scrub_unknown_but_valid_coderange_worst_case: |
string = unknown_but_valid_utf8_worst_case.dup
string.scrub!
scrub_unknown_and_invalid_coderange_worst_case: |
string = unknown_but_invalid_utf8_worst_case.dup
string.scrub!

12
string.c
View File

@ -11850,6 +11850,18 @@ enc_str_scrub(rb_encoding *enc, VALUE str, VALUE repl, int cr)
else if (MBCLEN_CHARFOUND_P(ret)) {
cr = ENC_CODERANGE_VALID;
p += MBCLEN_CHARFOUND_LEN(ret);
/*
* After a valid multibyte character, skip the following ASCII run.
* If the next byte is already non-ASCII, search_nonascii would only
* rediscover p after its word-at-a-time setup.
*/
if (p < e && ISASCII(*p)) {
p = search_nonascii(p, e);
if (!p) {
p = e;
break;
}
}
}
else if (MBCLEN_INVALID_P(ret)) {
/*