clundro

and we should use this header template. ``` //===----------------------------------------------------------------------===// // // BusTub // // rid.h // // Identification: src/include/common/rid.h // // Copyright (c) 2015-2019, Carnegie Mellon University Database Group //...

I guess the format_header py can insert some place and then execute the commands like 'make project2'

so just use an `int64` value as the ValueType for BPlusTree?

I think it doesn't cause the situation. The macro INTERNAL_PAGE_SIZE is defined in [b_plus_internal_tree.h](https://sourcegraph.com/github.com/cmu-db/bustub@dfa6cd4e82ef42eb111b889604cbf280771b7850/-/blob/src/include/storage/page/b_plus_tree_internal_page.h). ```cpp // valuetype for internalNode should be page id_t template class BPlusTreeInternalPage; template class BPlusTreeInternalPage; ```...

``` [bustub-private/src/storage/index/b_plus_tree.cpp:22:BPlusTree] INFO - Value_size:8 PageId_size:4 INTERNAL_PAGE_SIZE:254 [bustub-private/src/storage/page/b_plus_tree_leaf_page.cpp:38:Init] INFO - Value_size:8 Mapping_size:16 LEAF_PAGE_SIZE:254 [bustub-private/src/storage/page/b_plus_tree_internal_page.cpp:37:Init] INFO - Value_size:4 Mapping_size:12 INTERNAL_PAGE_SIZE:339 ``` I add `print` statements and find that `INTERNAL_PAGE_SIZE` is right...

I think it's good to throw a warning before an exit.

the dependency on parking_lot's impl in webkit is a bit heavy . perhaps we can consider folly (https://sourcegraph.com/github.com/facebook/folly/-/blob/folly/synchronization/Lock.h).

Emmm, it seems the result is related to your implementation. The origin doc is right for me.

I will test it in the latest branch after a few days. I forgot the origin commit.

ok, i delete unnecessary changes.