Fix getting a read timeout for logs/attach with a tty and slow output

[segevfiner]

Fixes #931

P.S. There is another possible instance of getting an unexpected timeout in attach_socket since you return the socket from requests/urllib3 to the user, which will have a timeout set.

@shin-

[akki]

Would any of the maintainers like to review this PR?

What would help in resolving #931?

[TomasTomecek]

Looks good!

[akki]

@TomasTomecek Since you created #931 and you find this PR to be its solution for you to merge it? Or can you tell what else we can do for it to get into master?

[akki]

We are affected in Airflow because of this issue. Although there are ways to hack around it but we would love it if we could fix this properly.

[TomasTomecek]

@TomasTomecek Since you created #931 and you find this PR to be its solution for you to merge it? Or can you tell what else we can do for it to get into master?

Find a maintainer of this repository and work with the person on a review for this PR. I'm not a maintainer, only a contributor.

Since @shin- already assigned this PR to himself, hopefully you'd get a maintainer review from him soon.

[CodingJonas]

@shin-, I hope it is okay to ping you.

This MR is open for over two years now, what is blocking it? I would be happy to help in any way to get this merged.

[HaloKo4]

@shin- is there a chance you can review?

[akki]

This PR is near its 3 years anniversary without any activity from any repo maintainers. So I hope it is okay to ping @ulyssessouza @aiordache to get their attention as they seem to be the most active maintainers of this repo right now.

[michaelfresco]

This PR would be so cool for Airflow and Docker Swarm. I hope this PR can be approved. Small ping to @ulyssessouza, @aiordache, @thaJeztah

[HaloKo4]

hey @aiordache @ulyssessouza

could you please get some attention to this PR? :)
3+ years for 2 lines for code is not reasonable.
I don't know why Docker ignores this PR - you are making you customers life really difficult.

[HaloKo4]

@ulyssessouza @aiordache @thaJeztah
Kindly ask for a comment, any comment.
What else are we to do to get attention from the repo maintainers? :(

[thaJeztah]

I'm not a maintainer for this repository (and not very proficient on Python code), but perhaps one of the others can have a look @ulyssessouza @aiordache

Looks like this PR needs a rebase though (I see there's a merge conflict)

[HaloKo4]

Looks like this PR needs a rebase though (I see there's a merge conflict)

It's 2 lines of code. Assuming this PR is accepted I have no problem to open a new PR with it or the maintainers can take over. This is really small modification. If not accepted kindly explain how maintainers suggest to solve the problem.

[thaJeztah]

It's 2 lines of code.

I understand it's a small change, and I see a similar approach is taken in other parts of this file, but it still needs to be reviewed. A single line change does not make a change less risky.

As said; I'm not a Python coder, so I defer review to others mentioned. It would be good if @segevfiner could rebase this PR and resolve the merge conflict, otherwise it can't be merged.

[ulyssessouza]

LGTM. Just needs a rebase.

[segevfiner]

Rebased.

[HaloKo4]

@aiordache when should we expect 5.0.3 to be released?

[aiordache]

@HaloKo4 Sorry for the delay, we'll have 5.0.3 out by Oct. 8th