Add serialization and blockstore stats for common tests #33
Conversation
austinabell
force-pushed
the
ser/bsstats
branch
from
24d8973
to
5bd6d05
Compare
| } | ||
|
|
||
| assert.Equal(t, "bafy2bzaced44wtasbcukqjqicvxzcyn5up6sorr5khzbdkl6zjeo736f377ew", c.String()) | ||
| assert.Equal(t, bsStats{r: 12, w: 12, br: 573, bw: 573}, trackingBs.stats) |
There was a problem hiding this comment.
these stats 😱 are why I think we need to work on a collapsable form #17 (comment) .. I don't have any idea how common this kind of perverse sparseness is in actual Filecoin data though
There was a problem hiding this comment.
the only stat that isn't used in Filecoin was bytes read, all other 3 are relevant
|
these lgtm but needs a rebase and to have the commented code dealt with - can
would be very interested to hear more on this if you still have these top of mind @austinabell |
Sorry, I don't have the bandwidth to do this anytime soon, the motivation was to make sure blockstore usages and roots of the amt is consistent for all cases as there were some strange patterns, but I don't know the current status or if this would be useful for other impls now that this is probably stabilized. Would be good to make sure things don't change in any case, but I don't have an opinion or motivation to push for this myself so feel free to close or modify this PR. |
|
closing because stale |
The stats include reads, writes, bytes read, bytes written to use to compare against since Filecoin uses reads, writes, and bytes written for gas usage.
Rust equivalent test additions: ChainSafe/forest#716
There are a few cases where there are many more writes than necessary, but I will open up an issue to explain