Rip panics trying to resolve spec "ntlk"

[notatallshaw]

Tested on both Windows and Linux

(.venv) PS C:\Users\damia\rip\rip> cargo r -- nltk
    Finished dev [unoptimized + debuginfo] target(s) in 0.29s
     Running `target\debug\rip.exe nltk`
2023-10-02T23:30:18.320971Z  INFO rip: cache directory: C:\Users\damia\AppData\Local\rattler/pypi
2023-10-02T23:30:18.399365Z  INFO rip: extracted the following environment markers from the system python interpreter:
Pep508EnvMakers {
    os_name: "nt",
    sys_platform: "win32",
    platform_machine: "AMD64",
    platform_python_implementation: "CPython",
    platform_release: "10",
    platform_system: "Windows",
    platform_version: "10.0.19045",
    python_version: "3.8",
    python_full_version: "3.8.10",
    implementation_name: "cpython",
    implementation_version: "3.8.10",
}
2023-10-02T23:30:18.402503Z  INFO rattler_installs_packages::resolve: collecting nltk
2023-10-02T23:30:18.402859Z  INFO rattler_installs_packages::http: executing request url=https://pypi.org/simple/nltk/ cache_mode=Default
thread 'main' panicked at 'failed to determine satisfiability of requires_python specifier:   × Can't use wildcard with GreaterThanEqual
', C:\Users\damia\rip\rip\crates\rattler_installs_packages\src\resolve.rs:209:26
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
error: process didn't exit successfully: `target\debug\rip.exe nltk` (exit code: 101)

[baszalmstra]

Ah yeah thats interesting. nltk 3.6 has the following required-python: >=3.5.*. But what does this mean?

3.5.* would mean >=3.5,<3.6
but
>=3.5 would allow larger than >=3.6.

Im ok with removing the panic but in that case should we use the restrictive bound (3.5.*) or the less restrictive bound (>=3.5)?

[baszalmstra]

I implemented a fix that uses the least restrictive bound in #41 . With that I am able to resolve nltk again.

[baszalmstra]

Ah yeah thats interesting. nltk 3.6 has the following required-python: >=3.5.*. But what does this mean?

3.5.* would mean >=3.5,<3.6
but
>=3.5 would allow larger than >=3.6.

Im ok with removing the panic but in that case should we use the restrictive bound (3.5.*) or the less restrictive bound (>=3.5)?

[baszalmstra]

I implemented a fix that uses the least restrictive bound in #41 . With that I am able to resolve nltk again.