Improve model ChainSel to fix ChainDB QSM test flakiness #3990
+22
−13
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
amesgen
approved these changes
Sep 15, 2022
ouroboros-consensus-test/test-storage/Test/Ouroboros/Storage/ChainDB/Model.hs
Outdated
Show resolved
Hide resolved
Comment on lines
880
to
883
| sortChains $ pruneKnownInvalid (invalid m) <$> chains bs | ||
| where | ||
| pruneKnownInvalid invalidBlocks chain = | ||
| Chain.takeWhileChain (\b -> blockHash b `Map.notMember` invalidBlocks) chain |
There was a problem hiding this comment.
Let's include a comment why we do this, sth along the lines of
We don't want known-invalid blocks to influence the sorting of candidate chains.
Also a small style thing: What about making invalidBlocks a where binding under pruneKnownInvalid?
bartfrenk
changed the title
jasagredo
approved these changes
Sep 15, 2022
| @@ -876,8 +876,12 @@ validChains cfg m bs = | |||
| -- first, which results in the valid chain A->B', which is then chosen | |||
| -- over the equally preferable A->B as it will be the first in the list | |||
| -- after a stable sort. | |||
There was a problem hiding this comment.
I would suggest enlarging this comment to explain the new behavior. It took me some passes to actually see how things are now solved.
bartfrenk
force-pushed
the
bartfrenk/3689/fix-tests
branch
from
93a13ef to
9db8c3f
Compare
bartfrenk
force-pushed
the
bartfrenk/3689/fix-tests
branch
from
9db8c3f to
f6df3ae
Compare
|
bors r+ |
amesgen
added a commit
to IntersectMBO/ouroboros-consensus
that referenced
this pull request
Aug 10, 2023
…SM test" This reverts commit f6df3ae. IntersectMBO/ouroboros-network#3990
amesgen
added a commit
to IntersectMBO/ouroboros-consensus
that referenced
this pull request
Aug 15, 2023
…SM test" This reverts commit f6df3ae. IntersectMBO/ouroboros-network#3990
github-merge-queue bot
pushed a commit
to IntersectMBO/ouroboros-consensus
that referenced
this pull request
Aug 17, 2023
…repancy (#284) This PR is the third round of fighting with a ChainDB q-s-m model vs SUT discrepancy for `getIsValid`, see #123 for the history. The two previous rounds either introduced new bugs or left certain bugs unaddressed. The approach of this PR is to highlight the *cumulative* change over all of these different fixes compared to the initial implementation is. For this, we first revert IntersectMBO/ouroboros-network#3651, IntersectMBO/ouroboros-network#3743 and IntersectMBO/ouroboros-network#3990 in the first four commits. It is recommended to skip these while reviewing. The fifth commit in this PR is then fixing the bug in one go with the benefit of hindsight, see its commit description for details. Closes #123
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Fix ChainDB QSM test by incorporating already detected invalid blocks into the model's logic that orders the candidate chains according to the way the actual implementation would try them.
Suppose that we have the candidate chains [GABCd, GAEf] in which lower case letters indicate invalid blocks, and in which the ChainDB already knows that d is invalid, and f is the last block that has been added.
The model implementation would:
It would not detect valid blocks in GAEf, in particular E.
The actual implementation already knows that d is invalid and would start from [GABC, GAEf], and the order in which it would try chains is [GAEf, GABC]. In the process it would detect that E is valid.
The solution proposed here changes 1. to already take into account known invalid blocks (we are keeping track of those in the model anyway).
Closes #3689.
Checklist