Path to this page:
Subject: CVS commit: pkgsrc/x11/x3270
From: Olaf Seibert
Date: 2023-03-29 15:27:16
Message id: 20230329132716.D3574F9F4@cvs.NetBSD.org
Log Message:
x11/x3270: update to version 4.2ga9.
Known issues
Resizing or maximizing the wc3270 window makes bad things happen.
Please do not do this.
Improvements
4.2ga9
Modified the ./configure script to be able to find the macOS
installations of OpenSSL from Homebrew and MacPorts.
4.2ga8
Added the reshort option to the wc3270 Session Wizard to re-create
all desktop shortcuts. This makes deploying the fix for FAQ Why is the
wc3270 window messed up on Windows 11? simpler.
Bug fixes
4.2ga9
x3270 mouse selections are much more predictable. Double- and
triple-clicks no longer move the cursor.
x3270 no longer crashes when it is reconfigured (model, font, etc.)
while the floating pop-up keypad is visible.
The security indicator in x3270 no longer disappears when it is
indicating unverified state and the window is reconfigured.
The ./configure option to give an alternate location for the OpenSSL
library (--with-openssl=) works correctly again. (Thanks to Sam Mearns
for the bug report.)
4.2ga8
When pasting in NVT mode and the text wraps to the next line, the
cursor no longer ends up in the wrong location.
Extended the source tarball generation script mktar.sh to expand
submodules, so the Expat library builds properly for Windows.
Modified the Session Wizard so that wc3270 doesn't end up running
inside Windows Terminal. (See the FAQ Why is the wc3270 window messed up
on Windows 11? for details.)
The x3270a script no longer causes a spurious Invalid (empty)
hostname error pop-up when the display's DPI is other than 96. (Bug
report by Devonian Teuchter is here.)
4.2ga7
The wc3270 Session Wizard no longer exits silently when it
encounters an error.
x3270 will no longer change fonts unnecessarily when the host code
page is changed. (bug report)
The background color is now propagated correctly by the RA order.
(courtesy of Sven Schnelle)
In a resource file, a resource value can now end with a
properly-quoted backslash, without the backslash being interpreted as a
line continuation.
The Transfer() action, when invoked from within a file read by the
Source() action, now blocks properly -- it returns once the file
transfer is complete.
If the Source() action is aborted, it will no longer leave the
keyboard locked.
Typeahead is now processed correctly when the host has negotiated
TN3270E Contention Resolution and the unlockDelay resource is set to
false. (bug report)
Files: