elaborate on use-case/file format

[raylu]

I noticed that coverage lcov is quite slow because it needs to parse all the source code into ASTs to figure out the number of statements (LF)

I thought this would solve my problem, but it appears it tries to parse .coverage as some kind of text-based line-delimited format. AFAICT, coveragepy switched from marshal to pickle to JSON to sqlite. I don't understand what format this is parsing

[yining]

hi @raylu

this project was to address "a very specific problem in one of my workflows" where I had to use an old version of coveragepy that does not support outputting lcov format. It does by parsing the .coverage file generated by coveragepy and generating in the corresponding lcov format.

If you are using a relative new coveragepy, it's most likely not what you have been looking for. I am sorry for the confusion.

[raylu]

right, I saw that. and I checked out coveragepy 6.2, but that still generates a sqlite .coverage file, not something like

coverage2lcov/src/lib.rs

Line 10 in 84a9ad8

/// `"src/lib/config.rs 45 11 76% 37-45, 60, 73"`

[yining]

IIRC, the version I used was 4.5.x... anyway, try coverage report --show-missing