On Tue, May 17, 2016 at 10:22:55AM +1000, Dmitry Smirnov wrote: > Even though triggered by a recent change in mc, this is not a bug in mc, > rather a bug in the (ancient, unmaintained, Gtk2-based) libvte9 (vte-0.28). > > See the upstream mc bug, as well as the other mc bug linked from there for > explanation. > > Fix for libvte9 is at https://bugzilla.gnome.org/show_bug.cgi?id=729533#c6
Right, I see. I'll rebase that patch and report on vte then. Although from the look of things, it seems like it'd require taking over maintainership, and for that I'd need to clean up my TODO queue first. Which is a shame, as all terminals in Debian I'd call good currently use libvte9 (for the insanity of GTK3's vte, it's enough to look at the font selection dialog). However, when it comes to mc: terminal bugs resemble kernel bugs: even if we fix everything in unstable, plenty of people still have old terminals -- a vast majority of mc's users are people who ssh around a lot. And old terminals are plentiful -- I for one often use an ancient laptop running stable LXDE and N900 (osso-xterm, vte based). Thus, what would you say about a workaround: unconditional write(1, "\e[?2004l") on exit? That's ugly but not uncommon, I for one did a lot of similar hacks in kbtin or dgamelaunch. Meow! -- A tit a day keeps the vet away.
