This tool is a smaller, less featured version of resize from the xterm port.
It's very useful when accessing systems via serial console that don't run X (e.g. dev boxes).
Details
Details
- Reviewers
kan cem wblock - Group Reviewers
manpages Contributor Reviews (src) - Commits
- rS297897: Add a small tool, resizewin(1), to query terminal for window size
Diff Detail
Diff Detail
- Repository
- rS FreeBSD src repository - subversion
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
| usr.bin/resizewin/resizewin.c | ||
|---|---|---|
| 48 ↗ | (On Diff #10906) | __unused goes after parameter name, I think. |
Comment Actions
Pretty sure read gets EINTR and it exits.
Certainly if I test it on a distant system (so I have time to press something before the terminal replies) pressing ^C results in it printing "Unable to parse response".
Comment Actions
Default action for SIGTERM/SIGINT is to terminate process, not to interrupt read. So, this will not do what you think it will.
Comment Actions
Hm, isn't this surprising? Isn't the default SIGINT handler "Exit immediately?"
If we get a Terminate Program signal mid-run, do we leave the terminal in a usable state (echo, cursor in the right place, etc)?
Comment Actions
I just realised my previous statement about EINTR was wrong. The code disables ISIG which stops the terminal processing for signal key presses, i.e. ^C becomes just another character (specifically 0x03) so the process doesn't get a signal at all.
Comment Actions
I do like having something similar in our tree actually, thought this was committed already.
| usr.bin/resizewin/resizewin.1 | ||
|---|---|---|
| 35 ↗ | (On Diff #10906) | This whole sentence is fairly vague. This program only needs VT100 compatibility in terminal size reporting. Being more specific and possibly also citing the ANSI spec might help. (I suspect we are past the days where everyone knows what VT100 compatibility is). It is not clear what "update the kernel with it" actually means. |
| 40 ↗ | (On Diff #10906) | This is a really long sentence that says several things and could be split into a couple of sentences. .Nm is functionally similar to .Xr resize 1 from the .Xr xterm 1 distribution. However, .Nm only works on VT100 (ANSI?)-compatible terminals, and does not emit commands to set environment variables. Overall, this command is defined in terms of the resize(1) command. But that command or its man page might not be available to the reader. So again, more specifics on what this command does would be better. We probably don't need an examples section, but it would be nice to add a sentence or two about when and how this command would be used: "After a terminal window has been resized, running .Nm will update the kernel's..." |
| usr.bin/resizewin/resizewin.1 | ||
|---|---|---|
| 40 ↗ | (On Diff #10906) | I split it into two, replaced VT100 with VT100/ANSI and added when you'd use it. I'm not really sure what else to add for the "what it does" part without going and explaining how curses applications work out how big the terminal is and why it matters (which seems out of scope) Do you have something more specific I should add? |
| usr.bin/resizewin/resizewin.1 | ||
|---|---|---|
| 28 ↗ | (On Diff #14380) | This needs to be bumped before commit. |
| 35 ↗ | (On Diff #14380) | Still an aside, but easier to read with commas: Query the terminal emulator, which is assumed to be VT100/ANSI compatible, for |
| 37 ↗ | (On Diff #14380) | Use the command name, because "It" is vague: .Nm is functionally similar to .Xr resize 1 from the .Xr xterm 1 distribution. |
| 41 ↗ | (On Diff #14380) | Start new sentences on new lines. |
| 46 ↗ | (On Diff #14380) | s/will update/updates/ |
| 57 ↗ | (On Diff #14380) | "resize" again here is redundant. |
| 35 ↗ | (On Diff #10906) | Still needs a bit more to explain what "update the kernel" means. From IRC: "Perform the TIOCSWINSZ ioctl on the terminal file descriptor..." |