mirror of
https://github.com/jlfwong/speedscope.git
synced 2024-11-23 14:37:09 +03:00
aaac0ad7e3
I knew early on that `integeruniquer.index` could be used to index into `integeruniquer.data`, but I initially thought it was an optimization rather than a necessity. It seems like if there's data past the 1MB threshold in `integeruniquer.data`, then `integeruniquer.index` is actually quite useful. The file seems to contain `[byte offset, MB offset]` pairs encoded as two 32 bit unsigned little endian integers. Using that to decode the integer arrays encoded in `integeruniquer.data` allows the file in #63 to load. Fixes #63 |
||
---|---|---|
.. | ||
integer-uniquer-data.grammar | ||
integer-uniquer-index.grammar | ||
object-allocations.grammar | ||
simple-allocations.trace.zip | ||
simple-time-profile.trace.zip | ||
time-sample-bulkstore.grammar |