SCT Issues / RiverWare 6.4 / December 2013
... with Phil's notes, edit 12-19-2013 (a).
Issues from David's 12-5-2013 e-mail: Summary of SCT issues.
See also release test plan:
R:\doc\sct\TestingPlan\SCT_TestingPlan.docx
(I) Issues addressed for BOTH RiverWare 6.4.2 and 6.5:
Fixed / Stat | |||
1 | Double clicking cell starts edit, doesn't toggle details (Gnats 5422, part). [6.4.1]. | ||
2 | 12-06-2013 | SCT opens with very wide column for arrows (Gnats 5422, part). | |
Analysis: We have been able to reproduce this only in Release builds, and it is intermittent. It might be occurring only the first time an SCT is shown within a RiverWare session. (We're not all seeing exactly the same behavior) -- SEE SCREENSHOT. | |||
3 | 12-11-2013 | Every time you run a DMI it takes you back to the top of the SCT | |
Note: It turns out that this was occuring only when the Input DMI changed a series slot's time range. | |||
4 | 12-09-2013 | Doesn't put the red cross hatching on slots that have dispatching disabled - It does work in Horizontal Time, but does not work when in Vertical Time. | |
Analysis: This seems to depend on the particular chosen Windows desktop "personalization" and some other user-specific configuration we haven't yet identified. On David's Windows 7 machine, the description above is accurate for all but the simplest "Windows Classic" personalization. Phil's Windows 7 machine always shows SOME ornamentation ... SEE SCREENSHOT ... the column header "missing-crosshatch" problem is only with the selected state, i.e. when any of the cells in the slot column are selected. On the selected columns, at least a red frame shows up. Given the use of the Qt4 model/view architecture, there is no "reasonable" programmatic way of altering this appearance. We provide the "background brush" (pattern and color) for column header cells, and its up to Qt4 to do what it does. Qt4 attempts to reconcile certain provisions with the style theme of the workspace, to make different applications have similar appearance. This necessarily prevents us from having full control of display features. Solution: In the Vertical Time axis orientations, we will instead crosshatch the actual data cells in columns corresponding to slots on object for which dispatching is disabled. |
|||
5 | 12-09-2013 | The “go to” menu puts the reservoir you choose at the bottom of the screen instead of the top | |
Analysis: The QAbstractItemView::scrollTo() method is turning out to not be sufficiently precise. The placement parameters are only "hints". The current algorithm places the desired row or column in the middle of the display. We need to implement this function ourselves at the QAbstractScrollArea level. | |||
6a | 12-10-2013 | Arrow keys don't work to move laterally [when completing an incell edit by pressing an arrow key]. | |
Comment: What is being requested is somewhat strange behavior (though it's true that the old SCT did do it). During an incell edit, the right and left arrow keys normally move the editor text cursor to a different place within the text being edited. We need to subclass QLineEdit to capture and specially handle the arrow keys. | |||
6b | In Horz Time mode, the Tab doesn't work after editing cells; it takes you to next slot but at the incorrect timestep. | ||
7 | 12-10-2013 | In Vert Time, the Column Headers: Include Scale and Units and Include Aggregation Summary Function don't seem to be working correctly. It seems like you can show both or neither but not one or the other unless you do very specific clicking sequence. This may just be an Apply issue as it seems to work when you click OK. | |
Analysis: The problem is that it's strangely difficult to cause a QHeaderView to update. There had already been a substantial amount of code introduced to force this update. It's still not working. Note that, in this bug scenario, an update can be manually forced (after clicking Apply in the SCT configuration dialog) by clicking in a different cell in the SCT series data table. I'm still trying to figure out how to do that programmatically! The obvious ways of doing that -- and several non-obvious ways of attempting to do it -- are not working. The chosen solution was to programmatically, momentarily hide and re-show the SctDialog (only when configuration settings specific to column headers have changed). This is a tad drastic, but this sort of SCT configuration change isn't made very often. |
|||
8 | 12-11-2013 | Summary config tab: Show Timestep Flag Colors. The first three options are not working. Only "Always" works. | |
subsequently discovered problems ... | |||
9 | 12-13-2013 | When you have notes on series slots that are displayed on the SCT, there is an icon shown. Previously, you got a tooltip with the text. There is no tooltip any more. | |
10 | 12-16-2013 | SCT lock status was preventing data editing. (Reported by Steve Setzer on 12-16-2013). This has been addressed for 6.4.2 and 6.5. |
(II) Lower level issues, mostly to be addressed only for RiverWare 6.5:
Fixed / Stat | For Vers |
||
A. | 12-12-2013 12-13-2013 |
6.4.2 | Previously noted by TVA, the close SCT confirmation is not useful to then and would like it to not happen. Perhaps it could be avoided if the SCT is locked. |
This has been addressed by refraining from showing the "OK to close SCT?" confirmation dialog if the SCT is locked AND had never been unlocked since last loading or saving the SCT from or to an SCT file. | |||
B. | 12-13-2013 |
6.4.2 | Lock syntax/notation is odd. Tooltip says "Lock SCT Config" regardless of the state. It should say "The SCT Configuration is locked, click to unlock" and vice-versa when unlocked. |
C. | 12-13-2013 | 6.4.2 | Aggregation config is enabled when SCT is locked. Documentation says this is not allowed. There are lots of things enabled when the SCT is locked. Perhaps we need to re-think what the lock does? |
The SCT Configuration toggle button (in the upper-left of the SCT Dialog) now effects also these controls: (1) The controls on all of the tabs of the SCT Configuration Dialog, and that dialog's "OK", "Apply" and "Reset" buttons. (2) Config >> Defaults >> Apply Default Settings (3) Aggregation >> Timestep Aggregation Config ... (4) Aggregation >> Set Summary Functions ... |
|||
D. | 12-11-2013 12-12-2013 |
6.5 | General Config tab: "Display Slots as Integer Indexed" is not clear what it means. Reword this. |
This checkbox has been replaced with a pair of radio buttons (revised 12-12-2013 -- SCREENSHOT): Series indices: (o) Timesteps (standard) (o) Integers |
|||
E. | 6.5 | Horz Time config: "Text labels in Slot Divider Rows" - Need verb to be consistent. Reword to "Show Text labels in Slot Divider Rows". | |
F. | 12-11-2013 | 6.5 | On the Scalar and Other Slots tab, why do we have row numbers? That is not a common way to do it. |
[Response from Phil]: These are sortable listviews. We're using this column to restore the custom order, i.e. by clicking in the row number column header which itself has only a sort indicator, not header text (the column needs to stay narrow). We can remove (actually: hide) this column; we would lose its sort indicator display. But the column isn't actually needed for restoring the custom order; when sorted by a different column, a (Custom Order) "Restore" button is shown below the list. | |||
G. | 12-17-2013 | 6.5 | When aggregating, the user should have some control over where the datetimes are placed in the detail cell. Currently in Horz Time, they are right justified. I would think Left justified would be easier to see, esp for long slot names. |
The SCT's aggregated, horizontal timestep view now supports an alignment option for detail row header text. That row header text (for detail rows) used to be just right aligned. Now it can be "left aligned" -- but this is really implemented as right aligned within a smaller computed rectangle based on ALL of the detail rows' header text. See screenshots http://cadswes2.colorado.edu/~philw/2013/SCT/Dec/images/DetRowHeaderAlign-Bake1-Anim.gif http://cadswes2.colorado.edu/~philw/2013/SCT/Dec/images/DetRowHeaderAlign-Bake1-Left.gif |
|||
H. | 12-18-2013 | 6.5 |
On the toolbar, the enabled flag buttons look like they are pressed in. We should have a better way to show the enabled vs disabled buttons. |
http://cadswes2.colorado.edu/~philw/2013/SCT/Dec/images/SctFlagButtons1.gif http://cadswes2.colorado.edu/~philw/2013/SCT/Dec/images/SctFlagButtons1x300.gif |
--- (end) ---