How were the constants for the whitespace_table lookup in json_character_block::classify choosen?
#2290
Replies: 1 comment 2 replies
-
|
This code has been around a long time, and I am not certain where these unused numbers come from. There might be some historical reason. We could probably trace it back using the git history. |
Beta Was this translation helpful? Give feedback.
2 replies
-
|
It seems to originate from this commit: 6c168f0#diff-ae56130591292550ceef73821a56532805a46ee2f4115507103d2c7a1c5aebddR622 |
Beta Was this translation helpful? Give feedback.
-
|
Yep, I think that's right, and you can see that the logic was different at the time. |
Beta Was this translation helpful? Give feedback.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
I'm a bit confused why such non-obvious constants were used in the
whitespace_tablelookup table:simdjson/src/icelake.cpp
Line 51 in 9b31492
simdjson/src/icelake.cpp
Lines 84 to 86 in 9b31492
If I understood it correctly, the only criterion for the constants is that they don't contain their index in the lower four bits.
Wouldn't using zeros for all non-whitespace characters be more intuitive?
Beta Was this translation helpful? Give feedback.
All reactions