While asciidoc is probably technically more flexible, some sites
(such as pkg.go.dev) do not support it well. For the basic docs
we have here, the extra richness from asciidoc is not needed, and
we would prefer not to sacrifice the broader support that MarkDown
enjoys.
While here the text was cleaned up, made consistent, with grammar
and spelling fixes.
The Tutorial was revised somewhat to provide more accurate guidance
for readers.
For most terminals this also is done with 'sgr0', but on some it
appears that we need to explicitly use the 'oc' capability. (This
is the documented behavior in terminfo.)
fixes#449 Lost keyboard input after suspend on Windows 10 PowerShell
fixes#148 Make tcell usable with any io.Reader and io.Writer
This introduces a new Tty interface so that applications can supply
their own implementation. This should facilitate work for applications
that wish to provide e.g. a webasm version of the terminal, or that need
to use different kinds of file plumbing.
I published this yesterday on github. It's a tmux plugin that scrapes
the active tmux pane for a magic wormhole code, then downloads and
optionally opens the file tied to the code. I wrote this to make it
easier to get pcaps from termshark into Wireshark. This plugin uses
gowid which itself depends wholly on tcell to manipulate the
terminal. Link: https://github.com/gcla/tmux-wormhole
Suspend call currently exits ca mode. So after the first suspend/resume
the program does not use the alternate screen anymore. Therefore Resume
needs to enter ca mode again.
fixes#194 Starting multiple screen sessions (lost key event)
You can test this by using the mouse demo, which now supports pressing
CTRL-Z. This does not actually suspend the demo, but starts a subshell
which takes over. After the subshell is exited, the demo takes control
of the screen back. This can be done multiple times, and it is possible
to start multiple "nested" iterations of the demo this way.
At some point in the future this is going to be replaced with
a simpler channel design, and the current blocking approach
is both non-idiomatic, and fragile. Don't use it.
This adds optional MouseFlags that can be used to adjust what is
tracked for mouse reporting (leaving the other modes to be handled
by the terminal.) This should work on all XTerm style terminals,
but on Windows we have no way to be selective here.
This does sort the main problem of screen.Fini(), but really we
want to use a separate Pause() and Resume() function, because
screen.Fini() is not meant for reuse in this way.
Note that one side effect of this change is that applications which
redirect stdin and stdout and expect us to just use /dev/tty instead
are going to break -- we are now using stdin and stdout like nearly
every other screen oriented application.
dependabot[bot]