[Fix] shrinks window before connecting

[4383]

Close #821

For further details please see:

• https://lore.kernel.org/all/20231103092706.6rw2ehuigxfdvhlc@lion.mk-sys.cz/
• https://bugzilla.suse.com/show_bug.cgi?id=1216858
• tests fail with kernel 6.6 #821

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (3f8c9e4) 56% compared to head (8029f27) 56%.

Additional details and impacted files

@@          Coverage Diff          @@
##           master   #905   +/-   ##
=====================================
  Coverage      56%    56%           
=====================================
  Files          89     89           
  Lines        9722   9722           
  Branches     1810   1810           
=====================================
  Hits         5462   5462           
  Misses       3885   3885           
  Partials      375    375           

Flag	Coverage Δ
ipv6	23% <ø> (ø)
py310asyncio	53% <ø> (ø)
py310epolls	53% <ø> (+<1%)	⬆️
py310poll	53% <ø> (-1%)	⬇️
py310selects	53% <ø> (ø)
py311asyncio	53% <ø> (-1%)	⬇️
py311epolls	53% <ø> (-1%)	⬇️
py312asyncio	50% <ø> (ø)
py312epolls	51% <ø> (+<1%)	⬆️
py37asyncio	50% <ø> (-1%)	⬇️
py37epolls	51% <ø> (ø)
py38asyncio	51% <ø> (-1%)	⬇️
py38epolls	53% <ø> (+<1%)	⬆️
py38openssl	52% <ø> (+<1%)	⬆️
py38poll	53% <ø> (-1%)	⬇️
py38selects	53% <ø> (ø)
py39asyncio	51% <ø> (ø)
py39dnspython1	51% <ø> (+<1%)	⬆️
py39epolls	53% <ø> (ø)
py39poll	53% <ø> (ø)
py39selects	53% <ø> (+<1%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[4383]

@jirislaby & @mkubecek: Please, let me know if that patch looks ok for you and if it fix your issue. Thanks.

[jirislaby]

@4383 Note that the original bug was worked around by 5f39d55 in the meantime. If I revert that (I see the failure again) and apply this, it works too.

[jirislaby]

And also the time is from:

real    0m1.442s
user    0m0.168s
sys     0m0.037s

to

real    0m0.214s
user    0m0.153s
sys     0m0.061s

[4383]

@jirislaby: Cool. Thanks for your feedbacks

[itamarst]

I am on kernel 6.6.6. master runs the test in 1.3 seconds, the PR runs the test in 0.1 seconds.