implement request_to_curl function

[guillermo-bondonno]

Implement request_to_curl util function that generates a curl command string from a scrapy.Request object

[wRAR]

Cool!

[codecov]

Codecov Report

Merging #5892 (c6bb8c4) into master (d47c732) will increase coverage by 0.01%.
The diff coverage is 95.45%.

❗ Current head c6bb8c4 differs from pull request most recent head a042f06. Consider uploading reports for the commit a042f06 to get more accurate results

@@            Coverage Diff             @@
##           master    #5892      +/-   ##
==========================================
+ Coverage   88.82%   88.83%   +0.01%     
==========================================
  Files         162      162              
  Lines       11108    11129      +21     
  Branches     1805     1811       +6     
==========================================
+ Hits         9867     9887      +20     
  Misses        960      960              
- Partials      281      282       +1     

Impacted Files	Coverage Δ
scrapy/utils/request.py	97.61% <93.33%> (-0.58%)	⬇️
scrapy/linkextractors/lxmlhtml.py	95.55% <100.00%> (+0.20%)	⬆️

[wRAR]

Can you please also look into pylint failures?

[guillermo-bondonno]

Can you please also look into pylint failures?

@wRAR It's about using .format instead of f-strings. I think it's much cleaner to use format in this case instead of doing something like

cookies = f"--cookie '{'; '.join(f'{list(c.keys())[0]}={list(c.values())[0]}' for c in request.cookies)}'"

but if it can't be merged as it is, I'll find a way around it sure.
What do you think?

[wRAR]

I think this warning should indeed be disabled, per-line or globally. I'll wait for other opinions though.

[Gallaecio]

I don’t think format is a significantly better choice here, I would find something like this quite readable:

cookie = "; ".join(f"{k}={v}" for k, v in request.cookies.items())
cookies = f"--cookie '{cookie}'"

No strong opinion, though. I am OK with disabling the check in these lines, but I would keep it globally so that f-strings are at least considered in cases like this one.

[Gallaecio]

I think it would be ideal to cover this somewhere in the docs for discoverability, but at the same time I cannot find a good place in the existing docs for it, so +1 to merge as is.