GNOME Bugzilla – Bug 736196
Please replace current alternate screen scroll patch with the new one, making this scroll permanent
Last modified: 2014-09-06 23:18:50 UTC
I forward this request from an Ubuntu bug, reported by Monsta (also author of the patch). Applying his patch upstream would finally fix it for all distros/terminals based on libvte9. So please have a look at it: https://bugs.launchpad.net/bugs/1340687 BTW: I can confirm that it is still a problem in libvte9 0.28.2-5ubnuntu1 in my up-to-date Linux Mint 17
I don't understand your request. The Ubuntu bugreport is about backporting something from newest vte to their old libvte9 0.28. 0.28 is old, abandoned, no longer maintained by the Gnome team. Distros might still backport some patches, it's not our business. The proper place for this request is the downstream Ubuntu bug that you linked. Feel free to submit a similar downstream bug report to Mint too. By the way, it's high time everybody ported their app to newest vte so that 0.28 can be retired for good.
Thanks for clarification. I hoped that there is still some support for past Gnome versions. So I close this report. We will continue to fix this in the Gnome 2 successor "MATE": https://github.com/mate-desktop/mate-terminal/issues/2
Oliver, are you a part of the Mate team, or in close touch with them? VTE has received hundreds of patches since 0.28, including fixes for crashes, intermittent display corruptions, permanent data corruptions, incorrect terminal behaviors etc. The alternate scrolling patch is nowhere near the most important ones. If anyone insists on using gtk 2.x / vte 0.28 and not upgrading to newer versions, probably they should step up and maintain a fork of vte 0.28. They should review all the patches that occurred in vte since then, and backport all the important ones. Think about maybe ~100 patches to backport that would IMO make sense and make that a better terminal emulator with significantly fewer known issues. In my opinion, backporting one particular relatively insignificant patch while keeping many more serious bugs unresolved doesn't make too much sense.
I used Ubuntu and Gnome since ~10 years. I still learn a lot by helping to fix bugs wherever it suits me most. I had to leave Ubuntu and Gnome since both decided to drop many features that grew to perfection over the years. With Mint and MATE I found loyal successors with all these features I need. I always report bugs as upstream as possible. But as you guessed: vte 0.28 has been forked already by MATE. So gnome.org was a bit "too upstream", sorry for that. BTW: Until end of 2013 I stuck to Ubuntu 11.04 with libvte 0.27.9! But I never saw crashes, display or data corruption. And mate-terminal in the newest Mint 17 seems very stable too. I only miss the alternate screen scroll.
(In reply to comment #4) > But I never saw crashes, display or data corruption. "I" is the key here. Most of the bugs are rare enough that probably one particular person never hits them, but some people from the whole Mate user base do suffer from them. Maybe you never saw crashes, display or data corruptions (and I forgot to mention the terrible slowness for example), but they are there, and if you looked up the bugreports you'd be able to reproduce them with vte 0.28. That's why I firmly believe that if the Mate team indeed decided to fork vte 0.28, they shouldn't just apply patches on a one-off basis as users request them; instead, they should proactively look at vte master's git log and bugzilla entries, and automatically backport at least all the important bugfixes (but preferably the improvements too).
(In reply to comment #4) > BTW: Until end of 2013 I stuck to Ubuntu 11.04 with libvte 0.27.9! But I never > saw crashes, display or data corruption. Also note that for the most well-known long-standing display corruption problem (bug 542087), Ubuntu (probably 11.04 already) has shipped a patch to fix that, which only recently got included in mainstream vte. Makes me wonder if it's there in Mate's fork... (sure I could go ahead and figure out, but I won't, sorry :))
I was wrong: libvte9 0.28 was NOT forked by MATE yet! That may be the reason why nobody has picked up the new scroll patch of Monsta. Before he made a patch for mate-terminal to fix this. After your comments here I think this was the better approach, at least for MATE. Thanks for your time!