fix: Fix set_tbl_width_chars panicking with negative width

[lukemanley]

fixes #18386

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 79.30%. Comparing base (6a3fc59) to head (9d51089).
Report is 48 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #20906      +/-   ##
==========================================
- Coverage   79.30%   79.30%   -0.01%     
==========================================
  Files        1578     1578              
  Lines      224135   224138       +3     
  Branches     2574     2574              
==========================================
- Hits       177751   177748       -3     
- Misses      45793    45799       +6     
  Partials      591      591              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[alexander-beedie]

Should also be a similar check down in the Rust layer (on read)? 👍
(Since the caller could set the env var there too, or otherwise outside of Config).

[lukemanley]

Should also be a similar check down in the Rust layer (on read)? 👍 (Since the caller could set the env var there too, or otherwise outside of Config).

I actually updated the PR this to interpret -1 as no limit (on the rust side) as is done in other config settings. Let me know if you think is ok?

[codspeed-hq]

CodSpeed Performance Report

Merging #20906 will not alter performance

_{Comparing lukemanley:set-tbl-width-chars (9d51089) with main (6a3fc59)}

Summary

✅ 41 untouched benchmarks

[alexander-beedie]

I forget if we interpret any negative limit as "no limit" (which I always found a bit weird, as why would -7 mean "no limit" 😄), or if it's just -1? Either way, I agree that making things consistent seems like a good approach.

[lukemanley]

I forget if we interpret any negative limit as "no limit" (which I always found a bit weird, as why would -7 mean "no limit" 😄), or if it's just -1? Either way, I agree that making things consistent seems like a good approach.

it looks like other config settings interpret any negative value as "no limit", e.g. here is number of columns displayed:

import polars as pl

df = pl.DataFrame(
    {str(i): [i] for i in range(10)}
)

print(df)

with pl.Config(tbl_cols=-7):
    print(df)

shape: (1, 10)
┌─────┬─────┬─────┬─────┬───┬─────┬─────┬─────┬─────┐
│ 0   ┆ 1   ┆ 2   ┆ 3   ┆ … ┆ 6   ┆ 7   ┆ 8   ┆ 9   │
│ --- ┆ --- ┆ --- ┆ --- ┆   ┆ --- ┆ --- ┆ --- ┆ --- │
│ i64 ┆ i64 ┆ i64 ┆ i64 ┆   ┆ i64 ┆ i64 ┆ i64 ┆ i64 │
╞═════╪═════╪═════╪═════╪═══╪═════╪═════╪═════╪═════╡
│ 0   ┆ 1   ┆ 2   ┆ 3   ┆ … ┆ 6   ┆ 7   ┆ 8   ┆ 9   │
└─────┴─────┴─────┴─────┴───┴─────┴─────┴─────┴─────┘
>>> with pl.Config(tbl_cols=-7):
...     print(df)
... 
shape: (1, 10)
┌─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┐
│ 0   ┆ 1   ┆ 2   ┆ 3   ┆ 4   ┆ 5   ┆ 6   ┆ 7   ┆ 8   ┆ 9   │
│ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- │
│ i64 ┆ i64 ┆ i64 ┆ i64 ┆ i64 ┆ i64 ┆ i64 ┆ i64 ┆ i64 ┆ i64 │
╞═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╡
│ 0   ┆ 1   ┆ 2   ┆ 3   ┆ 4   ┆ 5   ┆ 6   ┆ 7   ┆ 8   ┆ 9   │
└─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┘

I just pushed an update to the docstring to make it consisent with other settings, e.g set_tbl_cols: https://docs.pola.rs/api/python/dev/reference/api/polars.Config.set_tbl_cols.html