Tags: view all tags

Known Issues

Table Widget Editor

should only accept table source inputs (allowed me to enter a file name) * Fixed *, there was no editting on the fields at all. Made AP & Process name required fields.
hide font options, as they don't do anything (or use them as a default for the text on the widget) * Fixed * Now they do something
data source does not display in input sometimes, hard to reproduce. Try start in .DKL/PLO DRAG 'N DROP, Opt Int, then Opt 94 to .DKL/PLO SHELL SELECTED FILE,then A/C .DKL/PLO DETAILS SELECTED FILE, then IE, double click widget. Or Run the input from design: .DKL/PLO DRAG ’N DROP, Drag some files, Close the input, then: - Start in .DKL/PLO DRAG 'N DROP, Opt Int, then Opt 94 to .DKL/PLO SHELL SELECTED FILE,then goto A/C .DKL/PLO DETAILS SELECTED FILE, then Image Editor, double click widget. - I can't replicate this (pete)
Invocation Type s/b required field. ** Fixed **
Scan on App Id in Widget Editor for Tables gives 'Return encountered outside of subroutine' error. * Fixed **
Incorrect process name displayed at Image level if you edit the Table Source process from here. * Fixed *
Check online help
There is no spec to enable/disable multi-select
There is no option to require that at least one record be selected by the user at runtime.
There is no way to specify a child constraint for a table source process. This means that all of the records in the PCF must be read each time the table is populated.
There is no way to prevent a table widget from being repopulated when returning from an optional child process.
There is no ability to specify GUI options for column fields.
There should be separate event points for Post Read, Pre-Select, and Pre-Display
There should be an option intercept event point with the ability to intercept table events such as click, double-click, etc.
There is no option to specify which column the table should be sorted on when initially displayed.
There is no option to define column headings and their GUI attributes. Column heading text is taken from the data dictionary definition of the field.
There is no option to specify that a field (column) should perform a data lookup. Consequently, there is no option to have a data lookup populate columns for fields from other files.

Table Source Editor

hitting enter in key entry mode does not bring up list. Read First is set, I can't see what the problem is. Child Constraints maybe? * Fixed ***
App Id 0LA should display as --- * Fixed ***
Can't Browse and then Navigate to hits or drill down to error location if there are compiler errors - * Fixed *. Also added Options 96, 97, 98, 99 and 'Rules' to Item level.
link from DD does not work * Fixed ***
link from DD is not highlighted if table sources exist * Fixed ***
when adding fields from other files, warning message refers to 'Post User Selection' EP. * Fixed ***
Check online help
Accepts Invalid field names, accepts invalid occurrence numbers, Does not automatically check app --- if field not found in current app, Scan for field names does not work. * Fixed * There were edits in there, but they were conditioned on --- ERRORS = 0, so once one error occurred, no further editing was done. Created new sub ITEM (EDIT FLD) (TABLESOURCE), based on ITEM (EDIT FLD), which validates field names, occurrences, field types (GROUP fields not allowed), Domains, and checks 0LA if not found in current ap. Replaced all editing with this new routine.
Changes made may or may be not logged in SCCS, needs to be tested. Probably shouldn't fix this unless the Design Transfer problem is also fixed. Will likely have a ripple effect in SCCS re: closing tasks, etc. * Fixed * Was logging at the PROCESS level, but not the ITEM level. Tested locks, intent locks, color coding, task closing & design xfer.
Can't change mask if field is not in the current application - * Fixed *
Can't run this process via Opt 99. In theory, you should be able to run this process by itself to test your ILF code in SOP, EOP, Post PCF Read, etc. The 0AD INVOC PROC process has no way to invoke a TABLESRC process type, since there is no ILF verb for TABLESRC (as there is for INPUT, OUTPUT, QUERY, etc). Or, to put it another way, we need a TABLESRC ILF command. This will affect the x-ref processes, so they will need to be made aware of this new verb. * Can you do an Opt-97 to build the EM? You can't run a FILE process directly. I'm not sure running a Table Source process standalone has any value. But I could be wrong (pete) * Yes, you can use Opt 97, I added all those helper options to the input. You can always run the input that uses the Table Source to step thru the code, so maybe this could be deferred for later. I was just thinking of consistency, Opt 99 always let's me run a process, even queries (Jean).

Table Widget Runtime

double click option difficult to trigger * fixed * Now using desktop double-click value.
if a Size is set, data is truncated to that size, even if the user expands the column. ** fixed **
Filling dead space - right now the last column is expanded –if necessary- to create a row that fills the total with of the table, but that sometimes gives trouble in resizing the last column. ** Fixed **
Wide Columns appear to be truncated. Try to display a 512 character field. ** Fixed **
If there are compile errors in the Table Source, they are not displayed. Instead the default display for the Table Widget is shown. (Fixed, now showing EM compiler errors)
It seems that rewriting an ITEM (HTM VIEWER) to a table doesn’t work anymore. This is the code in OI:

SET --- WIDGET NAME = *TBL INF02

READ --- WIDGET HOLD 1 FT 0 BY WIDGET NAME

T SET --- WIDGET WIDGET TYPE = TABLE

T SET --- WIDGET D SRC AP = OTO

T SET --- WIDGET D SRC INV TYPE = SUBPROCESS

T SET --- WIDGET D SRC NAME = *TBL INF02

T REWRITE --- WIDGET FAIL 0
If a field has a size override >80 characters, the table will not display. Fields larger than 80 will display properly if there is no size override.
Token fields do not work properly, doesn't display the token text, I get some other character, or the table doesn't display at all.
Does not consistently execute Pre-Display, etc, after option set, ie, the table widget sets option to USER 1, Opt Int executes and then Pre-Display should re-execute (since --- OPTION is not cleared), but it doesn't always do that. See ROLE SECURITY MAINTENANCE - ACL for replication.
Same circumstances as above, table data is not reloaded.
The table widget is only suitable for displaying a relatively small number of records. Otherwise, the time needed to download the records to the client can get excessive. Records can only be loaded at a rate of about 2,000 records per second. Consequently, the table widget probably cannot be used as a record selector for larger records sets. For example, the time needed to load about 30,000 records is about 15 seconds.
The table widget is always repopulated when returning from an optional child process. This can cause long pauses if the number of records included in the table is more than a couple of thousand.
The mouse is the only way to navigate within a table widget. pgup, pgdn, cursor up, cursor dn, etc cannot be used.
Columns sort based on the displayed value in a column, not the field values. For example, dates displayed as mm/dd/yy will not sort in date order.
At runtime, a table widget never seems to keep the focus. When you select a record, the focus returns to the field which had the focus before you clicked to select a record.
At runtime, you cannot click on a row to select it unless all modifiable fields on the screen have satisfied the hard edits which are enforced by the engine. For example, an invalid character in a numeric field on the same screen will prevent you from selecting a record.

Drag & Drop Editor

There should be a flag on the --- WIDGET spec to indicate if the field can accept a Drop action. This should be available for the following Widget types: button, label, table widget, html viewer, and Window Background. Currently only works for Widgets attached to a field on the image. * Fixed * There is now --- WIDGET DROP TARGET that marks a widget as able to receive a drop action. It also looks at --- WIDGET FI CHOOSE MODE to see what kind of Drop should be supported. --- WIDGET FI CHOOSE MODE has values of ANY, FILE, FOLDER, TREE where ANY is a single file or folder name. Many more widget types now support being a drop target, but still require a field on the image.

Drag & Drop Runtime

There are times when it would be nice for the ADC Window to gain focus when an object is dropped on it. * Fixed * Client window now gains focus on a drop.
When dragging files over an Appx screen show where you can drop your files (change border or something of the widgets that can accept a drop action). More of a 'nice to have', since the mouse pointer changes when the mouse passes over a drop target. * Fixed *
Provide a runtime API to copy the --- DROP* files to memory versions of those files (MDROP* ?) to make it easier for a designer to save the contents of --- DROP.
Files dropped remain in the temporary folder forever. They should probably be cleared when the --- DROP* files are initialized. Alternatively, provide a method for the designer to specify a location for dropped files, and the designer will manually clear the directory at login. Is this 'Fixed', now that there is a setting for dropCachePath?
Create unique filenames for the temp file. Now the subject seems to be the name of the tempfile. This gives problems in dropping more items with the same subject. * Fixed ***
Possible bug? When files are dragged from Thunderbird & dropped, the temp file seems to be erased almost immediately (there is no opportunity for the Appx process to get it). * Fixed ***

New Event Points Editor (Tab In/Out/Value Changed)

Turning on/off any of the three turns them all off/on. The same field name is in all three positions. ** Fixed **
They only show up on the Alpha field GA, should be for any tabbable/modifiable field.
check online help

New Event Points Runtime (Tab In/Out/Value Changed)

If you have 2 fields beside each other, one with Tab Out and one with Tab In, only one event fires when you tab from one to the other. The Tab Out event on the first field does not fire, only the Tab In event on the next field fires. Probably needs to fire 2 events, the Tab Out (with --- CURSOR AT FIELD pointing to the field the user is leaving), then again for the Tab In. * Fixed * Now fire the Opt Intercept EPs for Tab Out, then Tab In, setting all the CURSOR AT... PDFs appropriately.
Two fields on an image. Both with tab in/out set to yes, when you switch from the desktop client to another program, then switch back to the client. Then tab to the next field, no option is fired. Update: Now the Tab Out event is fired, but then the cursor is put back in the field I just tabbed out of. If I tab out again, then the Tab Out/Tab In events are fired. * Fixed ***
Link tab in/tab out/value changed to a data lookup. This way it is easy to give already existing screens a more direct interaction with the user without redesigning the whole screen. A designer should tell though that the data lookup should be performed per field. If any of the tab in/out/value changed flags are set and there's a DLU, perform the DLU at the same time as the event is fired? * Fixed ***
The default value of valueChangedTimer can be set to 300 instead of 1000. Maybe we need some more testing to be done on that to find out what a nice default can be. * Fixed ***
When a Tab Out event fires, the --- CURSOR AT FIELD refers to the field the cursor is at now. It should refer to the field that triggered the Tab Out, not the next field it tabbed into. Or to put it another way, the Tab Out event should fire before the cursor moves to the next field. ** Fixed **
If value changed is set but the user tabs out of the field before valueChangedTimer has expired, the value changed event does not fire. If the user has changed the field, the value changed event should fire when they tab out. If Tab Out is also set, it should fire next. ** Fixed **
Setting the flags via ILF on non RAW TEXT widgets does not cause the events to fire. Probably should fire for all modifiable/tabable widgets.
It also seems like performing a shift-tab (going back to the previous field on the screen) does not work when your cursor is in a textfield with multiple rows.

Create Subroutine via ILF

0AD STMT (IO) (OPEN) receives --- AI and PASSes it to 0AD SE ALLOC, but SE ALLOC does not receive it, so no value is returned from SE ALLOC, and therefore no value is returned to STMT (IO) (OPEN)'s caller. * Fixed * - removed RECEIVE --- AI and PASS --- AI, removed PASS from .UTIL SUBR CREATE. The routines just rely on 0AD SEGEN SE POINTER directly.
0AD STMT (IO) (CLOSE) receives --- AI and PASSes it to 0AD SE FREE, but SE FREE does not receive it, so no value is returned from SE FREE. * Fixed * - removed RECEIVE --- AI and PASS --- AI, removed PASS from .UTIL SUBR CLOSE. The routines just rely on 0AD SEGEN SE POINTER directly.
The .UTIL SUBR routines do not follow standards - * Fixed * updated routines to API standards, added additional edits, etc.
You can call .UTIL SUBR CREATE more than once without calling .UTIL SUBR CLOSE in between. This will probably cause problems. * Fixed *
You can call the .UTIL SUBR WRITE and .UTIL SUBR CLOSE without calling .UTIL SUBR CREATE first. *Fixed*
.UTIL SUBR DELETE does not remove audit history - * Fixed * added call to 0AD PROCESS (REF INT) (DEL) to duplicate processing done when Sub deleted in 0AD.

Better Error/Warning Control Runtime

In client preferences, the messageComboRows shows a default of 0, but it seems if there are >8 messages, a scrollbar is added. Default should show as 8? ** Fixed **
Transitory messages (like process name during Em compile, restructuring, etc) no longer appear in the status bar.
If >1 ERROR message, color coding is lost (does not appear in Red). Check Warnings/Messages for the same problem.
The Control-M display no longer color codes the messages (It used to in an earlier version, I think).

Better HTML Widget

VLC option does not work - * Fixed once I properly installed VLC *
line breaks not honored in javascript viewer * Fixed * seems to work for me now.
Need proper Widget attributes vs @SHTML= * Fixed * Now have CODE VIEWER, HTML EDITOR, FLASH PLAYER, MEDIA PLAYER, WEB BROWSER widget types on Text fields. Also, CODE VIEWER looks at --- WIDGET CONTENT TYPE to see what syntax highlighting to do. WEB BROWSER looks at --- WIDGET SHOW DECORATE to see if it should show it's Web Browser controls or not.
Does this replace the old HTML Viewer? - No, the old HTML viewer is still available.
Praclox Web Page not rendered correctly. Compare Appx client version to web page in IE, notice Praclox banner at top of page missing.
HTML Editor is not able to load a resource which references a CKEditor Config file. The resource must reference a file which is in the format of a compiled CKEditor Config file and must conform to the format of a compiled config file.
There are problems if you attempt to use the HTML Editor widget on "small" text fields.
- If the field has three rows or less, the toolbar of the editor will completely overlay the field making it impossible to edit the field.
- If the field has fewer than 22 rows, you will not be able to use the Spellchecker option because the dialog that pops up is clipped at the top and bottom rendering it unusable.
- If the field has fewer than 13 rows, you will not be able to use the Find or the Find and Replace options because the dialog that pops up is clipped at the top and bottom rendering it unusable.
- The link dialog also get clipped. I haven't tired the table dialog but I suspect that it also gets clipped.
When the editor window has the focus, none of the APPX keys work. END, CANCEL, OPTION, TAB, etc. don't work. The toolbar buttons do still work so you can click on ENTER and CANCEL. To get out of the editor field you must use the mouse to click on the next field. END, CANCEL, TAB, OPTION, and perhaps some of the other APPX keys need to be honored even when editing text. TAB has no meaning within HTML except when used within a <pre> tag so I definitely would like to see TAB be available to TAB out of the editor field to the next field.
APPX should store HTML text without modification. e.g. Don't translate LF to paragraph characters. APPX should also pass stored text to the editor without modification. The purpose of the HTML editor is to create and edit HTML text. APPX should not modify the text that is created by the editor.

Add Mode No Longer Clears Screen

Cursor is moved to bottom of display instead of inserting a record at current location - Need to look into this more. Some screens adding at the bottom seems like the expected thing. Others not so much.
Some scrolling inputs in 0AD (Fields, End User Selections, Designer Selections, others?) attempt to position the cursor just after the file name. They expect the screen to be cleared, so now they position the cursor on the wrong row.

Client Window Sizing & Scaling

resizeMainWindowPane can be set to "true" to turn on the option to automatically resize the client window to fit the current frame/image size.
resizeMainWindowPane need to be moved out of the Experimental section of the Client Preferences
The other functionality described in feature 44 has not been implemented

Option Numbers in Design Mode

Option numbers are now displayed on buttons on the image editor if Show Option Numbers is turned on. But, there is no toolbar button or pulldown menu option to turn it on or off. It must be turned on or off by using the right-click menu. Better than nothing but not obvious.

Other

Can't xref table source process types * Fixed * You can x-ref to see where they are used, browse them to see what objects they use, and all x-refs know how to look at Table Source processes and show the hits.
Can't print tech doc for table source types - * Fixed * Formatting is similar to INPUT, but no frame/image detail, just list of Items.
Can't Design Transfer Table Source processes - * Fixed *
Need .WIDGET routine to REFRESH Table Widget via WIDGET VISIBLE INIT, maybe?. * Fixed * Now have field --- WIDGET RELOAD DATASRC as a widget field that if set to Y in Option Intercept will cause the Table widget to reload it's content even though it may not go to Pre-Display. There is also a new subroutine called .WIDGET RELOAD DATASRC that if passed a Widget Name will set the field with much less ILF code needed. You can also pass it two widget names and it will BEG AT and END AT on the two names and set all the widgets it finds.
Drag & Drop libraries are not installed during client install
Mac (and probably Linux) clients do not run