NEXT-37516 - Changed import error handling for database deadlock error

[CR0YD]

To counter errors like database deadlocks happening during an import of a chunk I reorganized the code in the sync_chunk method so that we can use the new -t --try-count argument of the sync command to specify the number of tries a request gets when it returns a processable error.
Errors that should be caught and processed can be added to the match statement in the function.
New error structures can now be added in the SwError enum.
New error codes can be added as constants like ERROR_CODE_DEADLOCK in the SwError enum.

[LarsKemper]

Nice job 💪 Only the failing test 😅

[github-actions]

Summary of the total line code coverage for the whole codebase

Total lines	Covered	Skipped	% (pr)	% (main)
2022	1132	890	55.98	57.03

Summary of each file (click to expand)

File	Total lines	Covered	Skipped	%
src/api/filter.rs	171	168	3	98.25
src/api/mod.rs	369	62	307	16.80
src/cli.rs	45	31	14	68.89
src/config_file.rs	76	56	20	73.68
src/data/export.rs	136	0	136	0.00
src/data/import.rs	162	0	162	0.00
src/data/transform/mod.rs	334	259	75	77.54
src/data/transform/script.rs	292	258	34	88.36
src/data/validate.rs	298	298	0	100.00
src/main.rs	139	0	139	0.00

More details (click to expand)

Download full HTML report

You can download the full HTML report here: click to download
Hint: You need to extract it locally and open the index.html, there you can see which lines are not covered in each file.

You can also generate these reports locally

For that, you need to install cargo-llvm-cov, then you can run:

cargo llvm-cov --all-features --no-fail-fast --open

Hint: There are also other ways to see code coverage in Rust. For example with RustRover, you can execute tests with coverage generation directly in the IDE.

Remember

Your tests should be meaningful and not just be written to raise the coverage.
Coverage is just a tool to detect forgotten code paths you may want to think about, not your instructor to write tests

[MalteJanz]

LGTM 👍