* fix(ci): update workflows/examples.yml
* Removed archived `actions-rs` actions
* Upgraded `Swatinem/rust-cache`, `actions/checkout`, `actions/download-artifact`, `actions/upload-artifact`
* chore: deprecate `AnyKind`
* fix: remove dead references to DB features in `sqlx-core`
This reactivates a couple of tests that had stopped working.
* chore(ci): move sqlx-cli checks to their own workflow
* fix(ci): remove remaining `uses: actions-rs/cargo`
* fix warnings
* chore: add titles to sqlx-cli jobs
* fix warnings (2)
* fix error in pool example
* fix warnings (3)
* fix query_builder test
* fix: don't run CLI tests on Windows
* chore: upgrade `rust-toolchain` to 1.80
* fix(postgres): fix missing Tokio specialization in `PgCopyIn`
Caught by the new `unexpected_cfgs` lint.
* fix new warnings
The `spec_error` module implements the autoref specialization hack that's used in a single place in the derive macros. It wasn't meant to be a public API.
Given how arcane it must appear, I doubt anyone's tried using it outside of this project anyway.
* Fixed some rust docs intra-doc non functioning links
* Minor tweaks
* Added warning for MSSQL not being functional yet
* Fixed requested changes
* Readded missing time
* Aligned table
* chore: create regression test for #3344
* fix(derives): use a parameter name that's less likely to collide
* breaking(derives): emit `Error::ColumnDecode` when a `TryFrom` conversion fails in `FromRow`
Breaking because `#[sqlx(default)]` on an individual field or the struct itself would have previously suppressed the error. This doesn't seem like good behavior as it could result in some potentially very difficult bugs.
Instead of using `TryFrom` for these fields, just implement `From` and apply the default explicitly.
* fix: run `cargo fmt`
* fix: use correct field in `ColumnDecode`
* fix: made sure that the json feature is defined for postgres as well
* fixed the feature order
Co-authored-by: Austin Bonander <austin.bonander@gmail.com>
---------
Co-authored-by: Austin Bonander <austin.bonander@gmail.com>
* fix(postgres): don't panic if `M` or `C` Notice fields are not UTF-8
This has been observed with an old version of PostgreSQL (11.0.4)
running on Windows Server 2016 with windows-1252 encoding and French
locale.
This change replaces invalid UTF-8 fields with a default string, so the
other fields can still be read if they are valid.
* Revert "fix(postgres): don't panic if `M` or `C` Notice fields are not UTF-8"
This reverts commit 362ca98bbd.
* Check that Notice M and C fields are valid UTF-8
Otherwise, we return the invalid UTF-8 error to avoid panicking later.
* feat: add cube
* docs: cube docs
* docs: update readme
* fix: cube is now not feature flagged
* fix: formatting
* fix: typeo for PgCube vs Cube
* fix: correct types
* fix: postgres only types for cube
* fix: cube readme
* fix: dont unwrap cubes
* fix: typo on interval
* fix: zero volume cube array
* fix: return type
* fix: update tests
* fix: run with one test type
* fix: log bytes in error
* fix: typo in test
* fix: log bytes for failed length
* fix: string deser
* docs: remove cube from readme
* fix: int to float
* fix: trim floats
* fix: exttra test
* fix: type safe into vectors
* fix: improve error messages
* docs: remove comments
* fix: front load most important logic and const at start
* fix: extract constants
* fix: flags
* fix: avoid redundant buffer creation and use FromStr trait
* fix: handle serializing
* test: cube vec test
* fix: no array cube test
* fix: update test with array for cube
* fix: dont use try from for u8
* fix: conditionally remove padding
* fix: conditional trimming
* fix: idiomatic trimming
* fix: linting
* fix: remove whitespace
* fix: lower case array
* fix: spacing input
* test: one more vec test
* fix: trim square brackets in case they are using postgres spec page
* fix: result types
* fix: format
* fix: box error
* fix: the borrow produces a value
* fix: self serialise
* chore: merge main
* fix: borrow
* fix: clippy
---------
Co-authored-by: James Holman <james.holman@betashares.com.au>
The names for the SSL mode options were incorrect, added test for making
sure the built url can be parsed to ensure internal consistency at
least. Also changed "ssl-mode" to "sslmode" as it seems more standard
and sqlx can parse both.
