823: wrong data populating right-shifted columns after F10 (top/bottom split)
Posted: 14.07.2020, 18:49
And now here's the second problem, newly born with 823's attempt at fixing the right-shift creep of column dividers in the lower pane from repeated F10 toggles.
With 820-822, repeated toggles via F10 would see the main divider between TREE and DETAILS panes progressively shift further and further right with each pair of F10 toggles. This no longer happens. However there still is an initial right-shift just once of the divider from the first F10. After that, each pair of F10 toggles keeps that main divider in the lower panel right-shifted exactly where it was the first F10... in the wrong place, but no longer progressively shifting further and further to the right with each pair of F10.
Turns out I can actually drag the main divider left, so that the TREE panes in the upper and lower panels now line up vertically. And now when I close and re-open the program sure enough the main divider location has "stuck" just where I placed it. This is how things worked with 8221. But it's not how the very first F10 worked with 823 by default (which saw one increment of right-shift creep), although it now seems to be "locked" properly after I manually corrected the main divider location.
But what is new with 823 is that when F10 is used repeatedly to go from single-panel to dual-panel (in top/bottom mode), the data that populates the columns in the lower panel is incorrect. If I change the selected folder (in the TREE pane) the columns now populate properly. And if I return to the original selected folder the columns again re-populate properly. But if I F10 to single-panel mode and then F10 again to dual-panel mode, once again the wrong data populates the columns for whatever folder is currently selected in the lower TREE when the F10 produces dual-panel mode.
But the right-shift creep now seems gone, as a result of my manual correction left-shift. Once that was done the lower panel TREE no longer grows in width each pair of F10, nor is it at the wrong location as it was when that very first F10 after installing 823 took place.
(1) So, here is the initial state of things, in single-panel mode. I have widened the window so that there is no horizontal scrollbar. These columns are too wide for my liking, but for the purpose of this demonstration I'm not going to touch them (since the new narrower locations are not remembered anyway with 823).
(2) Now I press F10 to get dual-panel mode, top/bottom. Note that the TREE in the lower pane is too wide, i.e. that the main divider has shifted right from where it really should be. It will remain "locked" at this incorrect location indefinitely as I repeatedly toggle F10, and will only change (and remain) at a new left-shifted proper location when I eventually drag it left. Once it's done it will stay there. But the initial F10 with the window widened so that no horizontal scrollbar is presented and with each column width at whatever size FCXE decided to make it... the bottom TREE is initially the wrong size. And consequently while there isn't a horizontal scrollbar in the upper pane, there is one in the lower pane.
(3) Now I F10 to get back to single-panel mode.
(4) And now I F10 again to get back to dual-panel mode. Note that once again the lower panel shows the one-unit increment of right-shift creep for the main divider (i.e. TREE is too wide by one incremental unit).
But the real defect now is that the column data is incorrect. The wrong information is used to populate the column cells.
(5) Still in split-screen dual-panel mode, I now change the selected folder in the lower TREE and voila the correct data now populates the column cells.
(6) Now F10 again to return to single-panel mode.
(7) And F10 again to return to dual-panel split-screen mode. Once again, incorrect data populating the columns of the selected folder in the lower TREE when this F10 is performed.
(8) And once again select a different folder in the originally incorrect lower panel, and now the column data correctly populates the column cells.
With 820-822, repeated toggles via F10 would see the main divider between TREE and DETAILS panes progressively shift further and further right with each pair of F10 toggles. This no longer happens. However there still is an initial right-shift just once of the divider from the first F10. After that, each pair of F10 toggles keeps that main divider in the lower panel right-shifted exactly where it was the first F10... in the wrong place, but no longer progressively shifting further and further to the right with each pair of F10.
Turns out I can actually drag the main divider left, so that the TREE panes in the upper and lower panels now line up vertically. And now when I close and re-open the program sure enough the main divider location has "stuck" just where I placed it. This is how things worked with 8221. But it's not how the very first F10 worked with 823 by default (which saw one increment of right-shift creep), although it now seems to be "locked" properly after I manually corrected the main divider location.
But what is new with 823 is that when F10 is used repeatedly to go from single-panel to dual-panel (in top/bottom mode), the data that populates the columns in the lower panel is incorrect. If I change the selected folder (in the TREE pane) the columns now populate properly. And if I return to the original selected folder the columns again re-populate properly. But if I F10 to single-panel mode and then F10 again to dual-panel mode, once again the wrong data populates the columns for whatever folder is currently selected in the lower TREE when the F10 produces dual-panel mode.
But the right-shift creep now seems gone, as a result of my manual correction left-shift. Once that was done the lower panel TREE no longer grows in width each pair of F10, nor is it at the wrong location as it was when that very first F10 after installing 823 took place.
(1) So, here is the initial state of things, in single-panel mode. I have widened the window so that there is no horizontal scrollbar. These columns are too wide for my liking, but for the purpose of this demonstration I'm not going to touch them (since the new narrower locations are not remembered anyway with 823).
(2) Now I press F10 to get dual-panel mode, top/bottom. Note that the TREE in the lower pane is too wide, i.e. that the main divider has shifted right from where it really should be. It will remain "locked" at this incorrect location indefinitely as I repeatedly toggle F10, and will only change (and remain) at a new left-shifted proper location when I eventually drag it left. Once it's done it will stay there. But the initial F10 with the window widened so that no horizontal scrollbar is presented and with each column width at whatever size FCXE decided to make it... the bottom TREE is initially the wrong size. And consequently while there isn't a horizontal scrollbar in the upper pane, there is one in the lower pane.
(3) Now I F10 to get back to single-panel mode.
(4) And now I F10 again to get back to dual-panel mode. Note that once again the lower panel shows the one-unit increment of right-shift creep for the main divider (i.e. TREE is too wide by one incremental unit).
But the real defect now is that the column data is incorrect. The wrong information is used to populate the column cells.
(5) Still in split-screen dual-panel mode, I now change the selected folder in the lower TREE and voila the correct data now populates the column cells.
(6) Now F10 again to return to single-panel mode.
(7) And F10 again to return to dual-panel split-screen mode. Once again, incorrect data populating the columns of the selected folder in the lower TREE when this F10 is performed.
(8) And once again select a different folder in the originally incorrect lower panel, and now the column data correctly populates the column cells.