You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently the terminology section of the spec mentions that the VLQ Base64 encoded are limited to 32-bit quantities.
This definition should probably be made more precise, and describe what "limited" means. For example, it could specify that such VLQs are invalid and implementations may reject them (other possible semantics might include clamping values to 32-bits, but I'm not aware of browser implementations that do this).
There's also a question of if the sign bit of the VLQ should be included in the 32 bits. It seems natural to me to include it, as a reason to pick a 32 bit limit to begin with might be to ensure an implementation can use a 32-bit representation to store the whole VLQ. However, I think Firefox actually doesn't include the sign bit in the check linked above (if it checked against i32::MAX instead of u32::MAX it would IIUC).
The text was updated successfully, but these errors were encountered:
takikawa
added a commit
to takikawa/source-map-spec
that referenced
this issue
Jun 24, 2024
Currently the terminology section of the spec mentions that the VLQ Base64 encoded are limited to 32-bit quantities.
This definition should probably be made more precise, and describe what "limited" means. For example, it could specify that such VLQs are invalid and implementations may reject them (other possible semantics might include clamping values to 32-bits, but I'm not aware of browser implementations that do this).
Firefox seems to consider such VLQs invalid while WebKit & Chrome don't appear to enforce the limit 1, 2.
There's also a question of if the sign bit of the VLQ should be included in the 32 bits. It seems natural to me to include it, as a reason to pick a 32 bit limit to begin with might be to ensure an implementation can use a 32-bit representation to store the whole VLQ. However, I think Firefox actually doesn't include the sign bit in the check linked above (if it checked against
i32::MAX
instead ofu32::MAX
it would IIUC).The text was updated successfully, but these errors were encountered: