Comment shebang line before passing to tokenizer

[hdwalters]

Fixes #435.

[Mte90]

Wondering if we can implement a test for this.
Also @Ph0enixKM what do you think of this change?

[Ph0enixKM]

Ohhh yeah.. good point!

[Ph0enixKM]

This fix seems to be good enough

[Ph0enixKM]

@Mte90 I think that this was my fault. I forgot that this would alter the size of the output file.

[Ph0enixKM]

Actually... can we replace shebang with an empty line? The comments can now persist in the generated bash file: #403

[hdwalters]

We could, but wouldn't it be better to show that there was a shebang in the input file, for debugging purposes?

[Ph0enixKM]

Okay let's leave it as a comment then

[Ph0enixKM]

Just a small tweak so that we output a cleaner bash code

[hdwalters]

Wondering if we can implement a test for this.

I'm not sure that's going to work without changes to the test framework, as it does not currently appear to capture standard error:

hwalters@Wintermute ~/git/amber (comment-shebang-line) 
$ cat src/tests/validity/snippet_no_shebang.ab 
#!/usr/bin/env amber

// Output
//  ERROR  Function 'upper' does not exist
// at src/tests/validity/snippet_no_shebang.ab:12:6
// 
// 11| import { lower } from "std/text"
// 12| echo upper("shouty")
// 13| 

import { lower } from "std/text"
echo upper("shouty")

hwalters@Wintermute ~/git/amber (comment-shebang-line) 
$ amber src/tests/validity/snippet_no_shebang.ab 
 ERROR  Function 'upper' does not exist
at src/tests/validity/snippet_no_shebang.ab:12:6

11| import { lower } from "std/text"
12| echo upper("shouty")
13| 

hwalters@Wintermute ~/git/amber (comment-shebang-line) 
$ cargo test tests::validity::test_validity_src_tests_validity_snippet_no_shebang_ab
    Finished `test` profile [optimized + debuginfo] target(s) in 0.11s
     Running unittests src/main.rs (target/debug/deps/amber-67b01322e90afde0)

running 1 test
test tests::validity::test_validity_src_tests_validity_snippet_no_shebang_ab ... FAILED

failures:

---- tests::validity::test_validity_src_tests_validity_snippet_no_shebang_ab stdout ----
thread 'tests::validity::test_validity_src_tests_validity_snippet_no_shebang_ab' panicked at src/tests/mod.rs:22:21:
ERROR: Function 'upper' does not exist
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace


failures:
    tests::validity::test_validity_src_tests_validity_snippet_no_shebang_ab

test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 191 filtered out; finished in 0.03s

error: test failed, to rerun pass `--bin amber`

[Mte90]

I'm not sure that's going to work without changes to the test framework, as it does not currently appear to capture standard error:

Good point, we need something for that. What do you think @Ph0enixKM @b1ek

[Ph0enixKM]

Let's not bother testing formatting of errors just yet. Unless someone wants to implement them ofc. There are some more important things to be done now

[Ph0enixKM]

Looks good! Thanks for your help 🙏

[Ph0enixKM]

@Mte90 that can be done with how we already validate the tests in stdlib and validity. We'd just check if the tests are failing and if so - with what error name.

I think that we should create error codes for different errors. This way we could just check which error code was triggered instead of checking the error message which could change anytime. I'll create issue for this