-
For example, openInput should check for "is file still being written" before it tries to actually open it.
This gives a better exception in case something screwy is happening.
---
Migrated from […
-
regardless of operating system.
We should take a good default value, like the lowest one we consider 'reasonable' macosX? and it should throw exception on openInput if openFiles hashmap is too large.…
-
I think this is confusing: the javadocs say you should not call it, if this is the case,
can we make it so that DocValues coming from IR is not actually closeable?
---
Migrated from [LUCENE-3861](h…
-
This assert was called all the time by all tests, causing a bottleneck. I disabled it in #10227, but it would be nice to add a subclass or package-private method or something to still test it (without…
-
Looks like FileSwitchDirectory has the same bugs in it as #4447.
We should randomly enable this guy in tests and flush them all out the same way.
---
Migrated from [LUCENE-3380](https://issues.apa…
-
unrelated but i thought about this looking at #5211: we should check this doesnt make a corrupt index but instead that IW throws a reasonable exception.
---
Migrated from [LUCENE-4141](https://issu…
-
This reproduces on 4.10:
ant test -Dtestcase=IndexAndTaxonomyReplicationClientTest -Dtests.method=testConsistencyOnExceptions -Dtests.seed=DFD235D80B3C7F11 -Dtests.multiplier=3 -Dtests.slow=true -Dt…
-
Currently MockAnalyzer's checks are implemented with assert. But this is just for convenience, we should not rely upon asserts being turned on: this is the whole point of the class.
---
Migrated fr…
-
SparseFixedBItSet is no longer used by DocIdSetBuilder, but the class didn't get cleaned up and removed.
We only need one sparse representation and don't need to just litter up the codebase with unne…
-
Today the classloader is somewhat hidden, you don't have a way to pass this. But since we are generating classes on the fly, we should allow it to be optionally specified.
---
Migrated from [LUCENE…