Thanks for clarifying the situation for me. My main use of J is for what's called Explorative Data Analysis. This means that I construct J-expressions on the fly, putting data into temporary variables and, quite often, use plot to get a feel for my data. I.e., my use of plot is not in a carefully constructed and tested script. Getting length errors and such happens on a regular basis and only takes seconds to remedy. Crashing my J-session is several orders of magnitude more disruptive. I lose temporary data, I have to reload all reference data, and, worst of all, it disrupts my chain of thought. AFAIK, even the session log is gone, erasing any traces of the previous hour(s) of work. Now, I could make changes to how I work, but my work process has evolved to rely on my favorite stable array language (Decades ago I transitioned from IBM VS/APL to Sharp APL. Try to make Sharp APL crash if you can.) so I'll go back to 602.
On Wed, Oct 26, 2022 at 1:16 AM Norman Drinkwater <nrdrinkwa...@gmail.com> wrote: > Qt 6.2 for j904 on windows and macos. Qt 5.15 for j903 on those platforms. > > On Tue, Oct 25, 2022 at 4:34 PM Raul Miller <rauldmil...@gmail.com> wrote: > > > There's at least three different ways of thinking about this issue: > > > > (1) It's a bug, wait for a fix or work around it. But this doesn't > > conflict with falling back to an earlier version in the mean time. > > > > (2) Fall back to what you know works, which was j602, > > > > (3) Use what we know -- which is that this is a bug in qt6 to find a > > version which does not use qt 6 -- but that would probably mean > > downloading and inspecting intermediate versions (possibly using a > > binary search strategy). Or: i don't see any labels on the archives, > > detailing what versions of dependencies are present for each version. > > > > I'm not sure if this helps, > > > > -- > > Raul > > > > On Tue, Oct 25, 2022 at 5:21 PM Pablo Landherr <pablo.landh...@gmail.com > > > > wrote: > > > > > > As the discussion seems to have ended, it seems to me that the only > > remedy, > > > i.e. to not have J crash in case I give *plot* incorrect parameters, is > > to > > > revert to an older J-version as the problem was not present when I ran > > > J602. Can anyone advise on how far back I should go? 602 or would 701 > or > > J8 > > > series be ok? > > > > > > Kind regards, > > > Pablo Landherr > > > > > > On Thu, Oct 20, 2022 at 7:50 PM Raul Miller <rauldmil...@gmail.com> > > wrote: > > > > > > > On Thu, Oct 20, 2022 at 11:36 AM Henry Rich <henryhr...@gmail.com> > > wrote: > > > > > Yes, that's what I see in Windows. The crash dialog says > > > > > > > > > > Exception thrown at 0x00007FFE0355E2B4 (Qt6Widgets.dll) in jqt.exe: > > > > > 0xC0000005: Access violation writing location 0x0000000000000028. > > > > > > > > > > That's not something I know how to fix. > > > > > > > > I think that any qt paint events inside a qt paint event should > either > > > > be deferred till after the paint event finishes or discarded. > > > > > > > > Ideal perhaps would be to make one deferred attempt and then discard > > > > further attempts. > > > > > > > > Simplest would probably be to discard all nested paint events > (relying > > > > instead on eventual command line error reporting). > > > > > > > > -- > > > > Raul > > > > > ---------------------------------------------------------------------- > > > > For information about J forums see > http://www.jsoftware.com/forums.htm > > > > > > > ---------------------------------------------------------------------- > > > For information about J forums see http://www.jsoftware.com/forums.htm > > ---------------------------------------------------------------------- > > For information about J forums see http://www.jsoftware.com/forums.htm > > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
