Conversation
|
@alexdima are these before/after measurements in milliseconds or something else? Do you happen to have millisecond measurements, or info on how you measured? Curious how this approach compares to https://github.com/trishume/syntect :) |
|
Also curious if you have any ideas why bootstrap.min.css became 4.1x slower on VS Code Desktop? |
|
Came here from link in release notes. I have massive respect for you @alexdima + vscode team's ever prioritization of performance. Really cool that this uses webassembly under the hood. Wow! |
|
This is insane! Keep it up!! |
|
@slimsag Yes, the measurements are in millis. I got them by running https://github.com/trishume/syntect looks like a great project. I'm not sure how times can be compared objectively. The current JS TM grammar that VS Code ships with is not written with speed in mind (it is basically the entire TS TM grammar with some minor things removed). Philosophically, it is also written more like a parser, which leads sometimes to some extraordinarily complex regular expressions -- e.g.). Matching such regular expressions is no breeze for any regex engine: Regarding |

All measurements done with the setting:
VS Code Desktop
VS Code Web (Chrome Version 81.0.4044.122)