✨ add Windows support

[AucaCoyan]

Hi! I like this app so much that I'm pull requesting the windows support ❤️

It compiles in windows, and works on Debian 💯

[AucaCoyan]

would you approve the workflow to see if I made a mistake somewhere? 🙏🏼

[AucaCoyan]

I merged master and solved the issue with lifetimes.
Sadly, ran the tests and had a failure in test merge_002

running 18 tests
test branch_003 ... ok
test branch_002 ... ok
test branch_001 ... ok
test branch_005 ... ok
test branch_004 ... ok
test merge_002 ... FAILED
test merge_003 ... ok
test merge_001 ... ok
test complex_001 ... ok
test orphan_001 ... ok
test merge_004 ... ok
test orphan_002 ... ok
test merge_005 ... ok
test stash_003 ... ok
test stash_001 ... ok
test stash_002 ... ok
test stash_004 ... ok
test straight_001 ... ok

failures:

---- merge_002 stdout ----
*   9280933 Merge branch '40'
|\
| * 2298534 041
| * f4022e8 032
| * 248648a 031
|/
| * cf3313d 022
| * 3be6c3d 021
|/
| * b053aba 012
| * 61077e9 011
|/
* b30f89c 001
thread 'merge_002' panicked at tests\graph.rs:1166:9:
Image dimensions are different. expected: (600, 500), actual: (550, 450)
Image dimensions are different. expected: (550, 500), actual: (550, 450)
stack backtrace:
   0: std::panicking::begin_panic_handler
             at /rustc/9fc6b43126469e3858e2fe86cafb4f0fd5068869\library/std\src\panicking.rs:665
   1: core::panicking::panic_fmt
             at /rustc/9fc6b43126469e3858e2fe86cafb4f0fd5068869\library/core\src\panicking.rs:76
   2: core::panicking::panic_display<alloc::string::String>
             at C:\Users\aucac\.rustup\toolchains\stable-x86_64-pc-windows-msvc\lib\rustlib\src\rust\library\core\src\panicking.rs:269
   3: graph::assert_graph_images::panic_cold_display<alloc::string::String>
             at C:\Users\aucac\.rustup\toolchains\stable-x86_64-pc-windows-msvc\lib\rustlib\src\rust\library\core\src\panic.rs:100
   4: graph::assert_graph_images
             at .\tests\graph.rs:1166
   5: graph::merge_002
             at .\tests\graph.rs:467
   6: graph::merge_002::closure$0
             at .\tests\graph.rs:424
   7: core::ops::function::FnOnce::call_once<graph::merge_002::closure_env$0,tuple$<> >
             at C:\Users\aucac\.rustup\toolchains\stable-x86_64-pc-windows-msvc\lib\rustlib\src\rust\library\core\src\ops\function.rs:250
   8: core::ops::function::FnOnce::call_once
             at /rustc/9fc6b43126469e3858e2fe86cafb4f0fd5068869\library/core\src\ops\function.rs:250
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.


failures:
    merge_002

test result: FAILED. 17 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 42.28s

error: test failed, to rerun pass `--test graph`

It just fails on windows, I ran this branch on Debian and all the tests passed successfully

[JereKaplas]

The test fails because of

.env("HOME", "/dev/null") in tests\graph.rs

I think the idea is to prevent git from loading the system config. A better way might be with:

.env("GIT_CONFIG_GLOBAL", "/dev/null")

Even though /dev/null isn't valid on Windows, the failure to find the HOME directory results in a failure, where as not finding the GIT_CONFIG_GLOBAL is handled more gracefully and the missing global config is just ignored.

[AucaCoyan]

Thank you! With your help I could change the line and pass the tests 🧪 .

I'm okay to merge this as is, but I'm not currently interested in the project, in my day to day use I tig which implents git blame, and see details of the commits (it's a different product, it doesn't have nice graphics like serie 😢).

So I'm fine both with merging this as somebody can continue working or closing and have it archived 😊. Cheers!