Going by last night's devel rev, VR was close to working as expected when I used the toggle command, at least with the particular layout I was using when I tried it. The branch for this PR, though, was a mess so I'm looking forward to your new approach. Hopefully I will be able to adapt it to VR3 too, since this part of the code started out the same.
On Friday, May 10, 2024 at 5:58:50 AM UTC-4 Edward K. Ream wrote: > This Engineering Notebook post discusses issues with the nested_splitter > (NS) plugin uncovered two days ago while working on #3892 > <https://github.com/leo-editor/leo-editor/issues/3892>: Add VR commands > to cover/uncover the body pane. I've been considering what to do ever since. > > *tl;dr*: See the Summary. > > *Background* > > PR #3893 <https://github.com/leo-editor/leo-editor/pull/3893> is a > thorough revision of the VR plugin. This PR leaves the VR3 plugin unchanged. > > Late-stage testing revealed that Leo's caches were not restoring the ratio > between the body and VR panes. A* lengthy *investigation showed that the > caching code was not to blame. Instead, the culprit is the > *LeoQtFrame.ratio* property, aka *c.frame.ratio*. > > The PR adds the VR pane in a way that breaks the c.frame.ratio property. > After several days of noodling, I suspect this property is not well-defined! > > *Remedies* > > First, I'll replace the ratio properties with explicit calls to two new > methods, say *c.frame.compute_ratio* and *c.frame.compute_secondary_ratio*. > There is no need to obscure where the computations happen! > > Second, I'll attempt to fix the computations. This work will likely > involve changes to the nested_splitter plugin. For example, the primary and > secondary splitters should have well-defined names. Computing the ratios > should depend on the names of various splitters, not their containment > relationships. Why? Because the NS plugin can change those relationships! > > *What I won't do* > > I don't want to change the PR to accommodate the NS plugin. That would > have the tail wag the dog. Instead, > c.frame.compute_ratio and c.frame.compute_secondary_ratio should produce > correct results in all situations. > > *Summary* > > Leo's Qt gui defines two (hard-to-find!!) and buggy properties: > *c.frame.ratio* and *c.frame.secondary_ratio*. I suspect these properties > are not well-defined! > > For clarity, Leo's Qt code will use two new methods instead: > *c.frame.compute_ratio* and *c.frame.compute_secondary_ratio*. These > methods may use enhancements to the NestedSplitter class. All aspects of > this work are experimental. > > The existing properties must remain for compatibility. However, each > property will print a deprecation warning the first time scripts call them. > > All your questions and comments are welcome. > > Edward > -- You received this message because you are subscribed to the Google Groups "leo-editor" group. To unsubscribe from this group and stop receiving emails from it, send an email to leo-editor+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/leo-editor/00acaa74-c1f9-454a-a505-d893a2ad3848n%40googlegroups.com.
