fix incorrect first Tid during index scan which using bitmap index

[gongxun0928]

Change logs

the _bitmap_findnexttids function has issues. It incorrectly skips data in some cases

The BMIterateResultstructure can store a maximum of 16 * 1024 = 16,384 TIDs, and in each iteration, we can return at most 16,384 TIDs. However, the BMBatchWordsstructure can store 3,968 words, each of which can be compressed (as specifiedby BMBatchWords hwords; see details in src/backend/access/bitmap/README). The maximum number of TIDs that can be stored in one word is equal to BM_MAX_TUPLES_PER_PAGE, which is an integer multiple of 64 and greater than 16,384.

This PR aims to fix the bug in this case. If words->cwords[0] is compressed and its length exceeds 16,384, the TIDs greater than 16,384 in cwords[0] will be skipped during access.

the first call: param words: { firstTid = 1; cwords[0] = 0x80000200 hwords[0] = 0xc0000000 }

result: { nextTid = 1; lastScanWordNo = 0; nextTids = (0, repeat 16384 times); numOfTids = 0; nextTidLoc 1; } After the first call: word: { firstTid = 1; cwords[0] = 0x80000100 hwords[0] = 0xc0000000 } result: { nextTid = 16385; lastScanWordNo = 0; nextTids = (1, 2, 3,..., 16384); numOfTids = 16384; nextTidLoc = 1; }

the second call: word: { firstTid = 1; cwords[0] = 0x80000100 hwords[0] = 0xc0000000 } result: { nextTid = 16385; lastScanWordNo = 0; nextTids = (1, 2, 3,..., 16384); numOfTids = 16384; nextTidLoc = 16385; }

words->firstTid is still equal to 1, but cwords[0] has already decremented the number of TIDs returned during the first call (0x100 * 64 = 16,384). At this point, firstTid does not match the actual data in cwords, and it is less than result->NextTid. The _bitmap_catchup_to_next_tid function will adjust firstTid, and simultaneously decrement cwords[0] by 16,384, resulting in TIDs 16,385 to 32,768 being skipped.

We need to ensure that firstTid is in the correct position. Therefore, every time we modify cwords, whether iterating by bit within a single word or skipping an entire word, we must synchronize the update of firstTid.

Why are the changes needed?

Describe why the changes are necessary.

Does this PR introduce any user-facing change?

If yes, please clarify the previous behavior and the change this PR proposes.

How was this patch tested?

Please detail how the changes were tested, including manual tests and any relevant unit or integration tests.

Contributor's Checklist

Here are some reminders and checklists before/when submitting your pull request, please check them:

• [ ] Make sure your Pull Request has a clear title and commit message. You can take git-commit template as a reference.
• [ ] Sign the Contributor License Agreement as prompted for your first-time contribution(One-time setup).
• [ ] Learn the coding contribution guide, including our code conventions, workflow and more.
• [ ] List your communication in the GitHub Issues or Discussions (if has or needed).
• [ ] Document changes.
• [ ] Add tests for the change
• [ ] Pass make installcheck
• [ ] Pass make -C src/test installcheck-cbdb-parallel
• [ ] Feel free to request cloudberrydb/dev team for review and approval when your PR is ready🥳

[gfphoenix78]

Can we produce some cases that hit this issue?

[my-ship-it]

Please fix CI failures？