Use unsaved in-memory commits during key list re-construction #3486
Conversation
| @@ -635,23 +636,77 @@ protected void validateHashExists(OP_CONTEXT ctx, Hash hash) throws ReferenceNot | |||
| */ | |||
| protected abstract List<CommitLogEntry> fetchPageFromCommitLog(OP_CONTEXT ctx, List<Hash> hashes); | |||
|
|
|||
| private List<CommitLogEntry> fetchPageFromCommitLog( | |||
There was a problem hiding this comment.
Mind renaming this to fetchMultipleFromCommitLog? I think, the term "page" is wrong here (and actually always was - it was just used to fetch a whole page).
There was a problem hiding this comment.
renamed (in a separate commit for ease of review)
Codecov Report
@@ Coverage Diff @@
## main #3486 +/- ##
============================================
+ Coverage 86.62% 86.66% +0.03%
- Complexity 3032 3043 +11
============================================
Files 376 376
Lines 16377 16411 +34
Branches 1207 1213 +6
============================================
+ Hits 14186 14222 +36
+ Misses 1746 1745 -1
+ Partials 445 444 -1
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
* Allow the unsaved in-memory commits built during merge / transplant operations to be found during key list re-construction. * Skip the in-memory lookup in the common case of non-merge call paths to reduce the amount of transient lists (parameters and results). * Parameterize merge / transplant test cases for full coverage of related code. Fixes projectnessie#3466
dimas-b
force-pushed
the
key-list-on-merge
branch
from
49f3e83 to
1a86023
Compare
dimas-b
changed the title
| BranchName main = BranchName.of("main"); | ||
| BranchName branch = BranchName.of("branch"); | ||
| BranchName conflict = BranchName.of("conflict"); | ||
|
|
||
| databaseAdapter.create(branch, databaseAdapter.hashOnReference(main, Optional.empty())); | ||
|
|
||
| Hash[] commits = new Hash[3]; | ||
| Hash[] commits = new Hash[numCommits]; |
There was a problem hiding this comment.
I remember that my PR for same issue got blocked because there was no strict validation. But this PR also don't have any validation but went ahead.
I am not sure I understand the process here.
There was a problem hiding this comment.
Your change was complex and changed the algorithm/implementation to generate the key-lists.
Dmitri's change did not change that algorithm.
There was a problem hiding this comment.
I also checked code coverage in IntelliJ (manually) - all branches in the new if statements were covered :)
Allow the unsaved in-memory commits built during merge / transplant
operations to be found during key list re-construction.
Skip the in-memory lookup in the common case of non-merge call
paths to reduce the amount of transient lists (parameters and
results).
Parameterize merge / transplant test cases for full coverage
of related code.
Fixes #3466