If V8 disambiguates keyword candidates from identifiers by keeping track of uppercase characters, would it make sense for minifies to prefer A-Z over a-z for identifiers?
The cornerstone of parser performance is a fast scanner. @tverwaes explains how V8’s JavaScript scanner got up to 2.1× faster: https://v8.dev/blog/scanner pic.twitter.com/n35oVWWcpU
-
-
-
Either use 1-char identifiers since all keywords are 2 or more chars, or identifiers that don't contain a keyword start or keyword continuation char. The simplest rule right now is one of A-Z, 0-9, $, _ as continuation char. A-Z as single char is probably the fastest :-)
- Još 3 druga odgovora
Novi razgovor -
-
-
Yeah! By the huge amount of modern-day javascript the parser and scanner cannot be fast enough
Great work @tverwaes and thx for all the fun. -
It was great having you! By now the (eager) parser is almost twice as fast as the bytecode compiler...
Kraj razgovora
Novi razgovor -
-
-
One-day, even people have not open the brower, The V8 is already run complete.
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
Regarding script streaming , What about gzipped js ? Can the parser start parsing the js before the buffering completes ?
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.