Preserve header casing. Take two. 🎬

[tomchristie]

Closes #31

Based on #103, focusing on minimising benchmark regressions.

• Preserves HTTP header casing.
• event.headers is a sequence-like instance, containing two-tuples of bytes.
• event.headers.raw_items() is a list of two-tuples of (raw_name, value).

Three benchmark runs here against each case.

Running against current master...

7330.1 requests/sec
7357.9 requests/sec
7363.4 requests/sec
7310.5 requests/sec
7363.2 requests/sec
7354.3 requests/sec
7363.3 requests/sec

7321.4 requests/sec
7355.2 requests/sec
7353.7 requests/sec
7375.7 requests/sec
7352.6 requests/sec
7381.0 requests/sec
7410.0 requests/sec

7234.4 requests/sec
7274.0 requests/sec
7231.8 requests/sec
7272.5 requests/sec
7263.7 requests/sec
7268.9 requests/sec
7273.7 requests/sec

Running against this PR...

7022.7 requests/sec
7041.2 requests/sec
7052.2 requests/sec
7038.3 requests/sec
7038.5 requests/sec
7042.8 requests/sec
7053.2 requests/sec

7129.9 requests/sec
7145.6 requests/sec
7152.4 requests/sec
7158.3 requests/sec
7154.1 requests/sec
7116.6 requests/sec
7146.2 requests/sec

7080.6 requests/sec
7077.9 requests/sec
7100.3 requests/sec
7051.9 requests/sec
7091.7 requests/sec
7072.6 requests/sec
7056.3 requests/sec

[codecov]

Codecov Report

Merging #104 into master will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #104      +/-   ##
==========================================
+ Coverage   99.14%   99.15%   +0.01%     
==========================================
  Files          21       21              
  Lines         937      952      +15     
  Branches      173      175       +2     
==========================================
+ Hits          929      944      +15     
  Misses          7        7              
  Partials        1        1              

Impacted Files	Coverage Δ
h11/tests/test_connection.py	100.00% <ø> (ø)
h11/tests/test_events.py	100.00% <ø> (ø)
h11/tests/test_headers.py	100.00% <ø> (ø)
h11/_connection.py	100.00% <100.00%> (ø)
h11/_headers.py	100.00% <100.00%> (ø)
h11/_readers.py	100.00% <100.00%> (ø)
h11/_writers.py	88.73% <100.00%> (+0.16%)	⬆️
h11/tests/test_io.py	100.00% <100.00%> (ø)
h11/tests/helpers.py	100.00% <0.00%> (ø)
... and 1 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 1540843...5f77670. Read the comment docs.

[tomchristie]

Okay, I think this is about the best we can do here.
I've updated the docs to include the one extra point of API that's now exposed.

Unavoidably it does benchmark marginally slower that the current master branch.

Before...

7211.1 requests/sec
7300.6 requests/sec
7298.3 requests/sec
7271.7 requests/sec
7342.2 requests/sec
7310.1 requests/sec
7334.3 requests/sec

After...

7025.6 requests/sec
7094.9 requests/sec
7095.7 requests/sec
7094.9 requests/sec
7115.5 requests/sec
7053.5 requests/sec
7104.2 requests/sec

Averaging to: 7295 requests/sec -> 7083 requests/sec

What do we think about this folks? @njsmith, @pgjones?

(From my POV I'd really like to get this resolved - we're currently treating it as one of our HTTPX 1.0 blockers.)

🙏💚😇

[StephenBrown2]

Will there be output demonstrating the result?

[tomchristie]

Yup, see https://h11.readthedocs.io/en/latest/api.html - but I wasn't able to figure out the docs build locally just yet, so I've not seen the rendered docs myself.

[pquentin]

You can see the docs rendered for this pull request at https://h11--104.org.readthedocs.build/en/104/

[tomchristie]

Ah neat, thanks @pquentin.

Not 100% obvious what repr we'd want for the Headers there:

• Use <Headers([(b'host', b'example.org')])> to make it clear it's not (quite) a plain list.
• Use [(b'host', b'example.org')] to keep it looking simple.
• Use ((b'host', b'example.org'),) since it's a non-mutable sequence.
• Something else?

Some examples as currently rendered...

[pquentin]

@StephenBrown2 @pgjones Would one of you be able to review this fine work? I'm not qualified (and not part of python-hyper anyways) and @njsmith isn't currently available for OSS reviews (even on Trio)

[pgjones]

Minded to merge - happy to be corrected - will merge and release over the weekend if not.

[tomchristie]

@pgjones Fantastic - pretty excited to get this one resolved. 🤗
Presumably we'll bump to 0.11 for this release?