wikimedia
/
parsoid
/
482
/
4
85%
master: 85%
master: 85%
|
Ran
|
Files
69
|
Run time
4s
|
Badge
Embed ▾
README BADGES
|
coverage: 85.441%. Remained the same
push
travis-ci
Tim Starling
Make tokenizer errors be more vague It's very expensive to collect error position and expectation information: it requires a call to peg$computeLocation() on every subexpression failure, just in case the whole start rule later fails. Wrapping the start rule in a "named" node causes this subexpression failure collection mechanism to be suppressed, improving speed by 30% on top of the previous no-cache patch, for the same test case. Change-Id: I6fd65fed2
11039 of 12920 relevant lines covered (85.44%)
22840.97 hits per line
