Alpine > Common#3632
Conversation
Signed-off-by: apostasie <spam_blackhole@farcloser.world>
Signed-off-by: apostasie <spam_blackhole@farcloser.world>
Signed-off-by: apostasie <spam_blackhole@farcloser.world>
| // that does not support inf/infinity. | ||
| // This constant is provided as a mean for tests to express the intention of sleep infinity without having to | ||
| // worry about that and get windows compatibility. | ||
| Infinity = "3600" |
There was a problem hiding this comment.
It's an infinitely finite number, right? :P
There was a problem hiding this comment.
I can make this an infinite number, but I am pretty sure you are going to complain about the size of the PR 😂
There was a problem hiding this comment.
Maybe just use "18446744073709551615" (or "4294967295" if it overflows)?
It should be also moved to *_windows.go
There was a problem hiding this comment.
I wonder if this is worth the effort.
For all intents and purposes, these containers never live more than a few seconds.
Furthermore, having two distinct behaviors for linux (infinity) and windows (4294967295) is inviting trouble.
Finally, this is just a workaround for a situation that hopefully will resolve eventually.
wdyt?
| ) | ||
|
|
||
| const ( | ||
| // It seems that at this moment, the busybox on windows image we are using has an outdated version of sleep |
There was a problem hiding this comment.
busybox should have been supporting inf since 2019 🤔
mirror/busybox@edca770
There was a problem hiding this comment.
Pretty sure the k8s folks are using https://github.com/rmyorston/busybox-w32
We could get our own rolled-up with some effort.
I do not have a windows rig here though... touching anything windows for me means I have to code blind and hope the CI will be happy :s.
There was a problem hiding this comment.
Looks like everybody is still on 1.29-2 - untouched in 6 years... k8s image tag is likely lying for windows.
2 participants
On top of #3631
More changes to allow better compatibility with windows test (move from AlpineImage to CommonImage where appropriate).
Second commit fixes an image test that did not match the description.