On Fri, Feb 02, 2018 at 11:47:34PM +0000, Guillaume MM wrote: > Le 01/02/2018 à 22:23, Scott Kostyshak a écrit : > > On Thu, Feb 01, 2018 at 07:19:03PM +0000, Guillaume MM wrote: > > > Le 01/02/2018 à 18:21, Scott Kostyshak a écrit : > > > > On Thu, Feb 01, 2018 at 02:00:43PM +0000, Jürgen Spitzmüller wrote: > > > > > Am Donnerstag, den 01.02.2018, 11:40 +0100 schrieb Jürgen Spitzmüller: > > > > > > Am Donnerstag, den 01.02.2018, 11:22 +0100 schrieb Jean-Marc > > > > > > Lasgouttes: > > > > > > > This makes much sense. You could probably move the new line to the > > > > > > > other > > > > > > > GuiWorkArea constructor... > > > > > > Yes, this seems to work as well. > > > > > I have now pushed this fix to master. > > > > Thanks for the fix, Jürgen. Let's double-check with Guillaume and port > > > > to 2.3.x. Guillaume does this patch make sense to you? > > > > > > > > Scott > > > > > > Dear Scott, this looks good to me. > > Thanks for taking a look. > > > > > I would not have proposed it for 2.3, > > I'm pretty sure I know what you mean but I want to make 100% sure: > > > > You mean the new compressor in the first place (i.e. commit 789617b8), > > right? Conditional on us keeping that commit for 2.3.0, you do agree > > that it makes sense to backport the follow-up 5ce36019, right? > > Yes, you need Jürgen's fix.
OK. > > The new compressor is in 2.3.0rc2. If we do not receive any other > > reports that seem linked to it, do you agree that we should keep the > > compressor for 2.3.0? Or do you have reason to think there is a > > significant chance of a nasty bug that we might not encounter while > > testing that could be serious? > The source of uncertainty is that it changes the behaviour on all platforms > in a non-specific way. Just make sure it is well tested on all platforms. To > test it, put the system in heavy load and check that LyX behaves nicely when > overwhelmed with keyboard input, not necessarily with autorepeat. Ah this is good to know. I might ask on the lyx-users list for testers of rc2 on Mac and Win if they can test. Is the following a reasonable idea? ---------- LyX 2.3.0 will have a more responsive way of handling keyboard input. Could you please help us test this new method? The following is one way: 1. Go to a large Wikipedia article (e.g. [1]), select all, and copy. 2. Go to LyX and paste. 3. Immediately after step 2 (i.e. while LyX is pasting), start typing a sentence as fast as you can. It is normal that during step 3, LyX will not be responsive for a few seconds while it pastes. What we want to know is whether the sentence you typed in 3 shows up correctly (i.e. without dropped characters) after LyX becomes responsive again. ---------- Can you think of any other specific instructions we could give to testers? Thanks, Scott [1] https://en.wikipedia.org/wiki/List_of_compositions_by_Franz_Schubert
signature.asc
Description: PGP signature