If ioctl TIOCGWINSZ returns 0,0 try to use env vars instead. #893
Conversation
For some terminals (for example Emacs eshell/eterm) the ioctl with TIOCGWINSZ might falsely return 0 columns and 0 rows. If this happens we now try to use LINES and COLUMNS environment variables. Fixes crossterm-rs#891
- Added comment with rational for window size fallback mechanism. - Only unwrap once during parsing of env vars.
I don't expect the operation to be expensive. In addition it only happens when ioctl |
Did a quick benchmark. On a mac, this is ~100ns, so fast enough generally to not matter. If this turned out to be a hotspot (actually measured by a profiler), it would be easy to replace with a lazy static, but there's no need to do this upfront for now. |
|
@TimonPost this would probably be easy to merge |
|
Don't mean to be a PITA but is there anything I could do to increase chances of getting the PR merged? |
I've messaged @TimonPost about this on discord. For context I'm just an interested party - not a maintainer of crossterm. Perhaps if you have some time, it might be good to get some extra eyes on the other PRs and help out from the perspective of testing / reviewing code. Does that sound like something that would work for you? |
|
Personally do not have a presence if ppl writing cli applications find a very usefull use to this im fine with getting it in. |
For some terminals (for example Emacs eshell/eterm) the ioctl with TIOCGWINSZ might falsely return 0 columns and 0 rows. If this happens we now try to use LINES and COLUMNS environment variables.
Fixes #891