Dan Luu

Memory leak on Slice creation in IngestorTest (valgrind)

2

``` ==20327== 104 bytes in 1 blocks are definitely lost in loss record 3 of 4 ==20327== at 0x4C2E0EF: operator new(unsigned long) (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==20327== by 0x5A44D8: BitFunnel::Shard::CreateNewActiveSlice() (Shard.cpp:132) ==20327==...

This is using the data from chunk1 of http://bitfunnel.org/wikipedia-as-test-corpus-for-bitfunnel/ and the code as of a740476e7955452c2dd6396367c2e8fc341e995d. statistics and a termtable were regenerated and the config files from the post weren't used....

``` Thread 1 exited. ==21366== Mismatched free() / delete / delete [] ==21366== at 0x4C2F24B: operator delete(void*) (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==21366== by 0x5B9D16: BitFunnel::DocumentCache::~DocumentCache() (DocumentCache.cpp:45) ==21366== by 0x5B9D88: BitFunnel::DocumentCache::~DocumentCache() (DocumentCache.cpp:37) ==21366==...

What are the consequences of a Term::Hash colliding with a Fact?

2

Obviously, this means that a fact could incorrectly get asserted, but what are the specific negative consequences of that?

AbstractRowEnumerator::PadWithDuplicateRowIds ``` { // All the rows for this rank are invalid rows. // Need to replace it with the match-all row id. planRows.PhysicalRow(shard, currentId) = m_matchAllTermRowIds[shard]; } ``` There's...

Shard has this design note: ``` // DESIGN NOTE: using pointers, rather than embedded instances to avoid // initializer order dependencies in constructor list. ``` But this note isn't actually...

We should use/create `emplace_back`.

### Quick summary When following the instructions in https://github.com/Automattic/themes/blob/trunk/CONTRIBUTING.md, `npm install` errors out if `python` / `python3` is `python3.12`. I believe this is because `distutils` was removed in `3.12`. Changing...

Fix some typos.