39b1763730 Replace use of `ArgsManager` with `DatabaseOptions` (Kiminuo)
Pull request description:
Contributes to #21005.
The goal of this PR is to remove `gArgs` from database classes (i.e. `bdb.h` and `sqlite.h`) so that they can be tested without relying on `gArgs` in tests.
Notes:
* My goal is to enable unit-testing without relying on `gArgs` as much as possible. Global variables are hard to reason about which in turn makes it slightly harder to contribute to this codebase. When the compiler does the heavy lifting for us and allows us only to construct an object (or call a method) with valid parameters, we may also save some time in code reviews. The cost for this is passing an argument which is not for free but the cost is very miniscule compared to benefits, I think.
* GUI code is an exception because it seems fine to have `gArgs` there so I don't plan to make changes in `src/qt` folder, for example.
* My approach to removal of `gArgs` uses is moving from lower levels to upper ones and pass `ArgsManager` as an argument as needed. The approach is very similar to what #20158.
ACKs for top commit:
achow101:
ACK 39b1763730
ryanofsky:
Code review ACK 39b1763730. Just the two small ReadDatabaseArgs and Berkeley open changes that were discussed since the last review
Tree-SHA512: aa066b314db593e46c18698fe8cdd500f558b405dc04e4a9a3ff57b52b5b3a81a6cb090e0e661785d1d02c1bf18958c1f4cd715ff233aab63381e3f80960622d
fa5d2e678c Remove unused char serialize (MarcoFalke)
fa24493d63 Use spans of std::byte in serialize (MarcoFalke)
fa65bbf217 span: Add BytePtr helper (MarcoFalke)
Pull request description:
This changes the serialize code (`.read()` and `.write()` functions) to take a `Span` instead of a pointer and size. This is a breaking change for the serialize interface, so at no additional cost we can also switch to `std::byte` (instead of using `char`).
The benefits of using `Span`:
* Less verbose and less fragile code when passing an already existing `Span`(-like) object to or from serialization
The benefits of using `std::byte`:
* `std::byte` can't accidentally be mistaken for an integer
The goal here is to only change serialize to use spans of `std::byte`. If needed, `AsBytes`, `MakeUCharSpan`, ... can be used (temporarily) to pass spans of the right type.
Other changes that are included here:
* [#22167](https://github.com/bitcoin/bitcoin/pull/22167) (refactor: Remove char serialize by MarcoFalke)
* [#21906](https://github.com/bitcoin/bitcoin/pull/21906) (Preserve const in cast on CTransactionSignatureSerializer by promag)
ACKs for top commit:
laanwj:
Concept and code review ACK fa5d2e678c
sipa:
re-utACK fa5d2e678c
Tree-SHA512: 08ee9eced5fb777cedae593b11e33660bed9a3e1711a7451a87b835089a96c99ce0632918bb4666a4e859c4d020f88fb50f2dd734216b0c3d1a9a704967ece6f
There is no change in behavior. This just helps prepare for the
transition from boost::filesystem to std::filesystem by avoiding calls
to methods which will be unsafe after the transaction to std::filesystem
to due lack of a boost::filesystem::path::imbue equivalent and inability
to set a predictable locale.
Co-authored-by: Hennadii Stepanov <32963518+hebasto@users.noreply.github.com>
Co-authored-by: Kiminuo <kiminuo@protonmail.com>
Co-authored-by: MarcoFalke <falke.marco@gmail.com>
With this change, we get more fine-grained error messages if something
goes wrong in the course of communicating with the SQLite database. To
pick some random examples, the error codes SQLITE_IOERR_NOMEM,
SQLITE_IOERR_CORRUPTFS or SQLITE_IOERR_FSYNC are way more specific than just a
plain SQLITE_IOERR, and the corresponding error messages generated by
sqlite3_errstr() will hence give a better hint to the user (or also to the
developers, if an error report is sent) what the cause for a failure is.
Since we want tests to run quickly, and since tests do a lot more db
operations than expected we expect to see in actual usage, we disable
sqlite's syncing behavior to make db operations run much faster. This
syncing behavior is necessary for normal operation as it helps guarantee
that data won't become lost or corrupted, but in tests, we don't care
about that.
No change in behavior. Just remove a little bit of code, reduce macro usage,
remove duplicative functions, and make BDB and SQLite implementations more
consistent with each other.
This commit does not change to any code and behavior. It it is easily reviewed
with the --color-moved=dimmed_zebra git diff option.
Motivation for this change is to:
- Consolidate redundant functions
IsBDBFile /ExistsBerkeleyDatabase / SplitWalletPath, and
IsSQLiteFile / ExistsSQLiteDatabase in the next commits
- Detect SQLite wallets consistently regardless whether bitcoin is built with
SQLite support in the next commits
- Avoid attempting to open SQLite databases with the BDB library when bitcoin
is built without SQLite support in the next commits
If there is no terminating zero within the 16 magic bytes, the buffer would be
over-read in the std::string constructor. Fixed by using the "from buffer"
variant of the ctor (that also takes a size) rather than the "from c-string"
variant.
Rewrite uses the VACUUM command which does exactly what we want. A
specific advertised use case is to compact a database and ensure that
any deleted data is actually deleted.
sqlite3 recommends that sqlite3_initialize be called when the
application starts, and sqlite3_shutdown when it stops. Since we don't
always use sqlite3, we initialize it when a SQLiteDatabse is constructed
(calling sqlite3_initialize after initialized is a no-op). We call
sqlite3_shutdown when we see that there are no databases opened. The
number of open databases is tracked by an atomic g_dbs_open.
