986.9.1 |
INFLD: Setting PARENTID to ">fname" (i.e. without a parent name or number prior to the ">") was causing it to double the row. |
986.9.2 |
XTREE: The limit on the number of RGBfg and RGBbg color definitions has been increased from 16 to 26 (from A-P to A-Z). |
986.9.3 |
XTREE: Masks were not being applied to data loaded from the ANSARY. (This was only an issue when the application pre-loads the ANSARY data, rather than just clearing it to spaces and letting XTREE load it, but that's not uncommon.) |
986.9.4 |
Compiler fix: Use of the CHAIN statement within a Function or Procedure was reporting the wrong error message. It was saying "Illegal IF construct" but it should have said "Illegal operation inside function." |
986.9.5 |
XTREE: Fix cell editing problem in which certain editing characters (e.g. backspace, delete) would stop working after passing through a read-only editable cell. |
986.8.1 |
Editable cells in XTREE with conditional validation were always acting as if they had been changed if there was a mask defined. |
|
|
986.7.1 |
XTREE: When exiting an editable tree from selection mode (as opposed to editing mode) the returned XTR'ROW and XTR'TARGETROW are now set to the row # (just as ANSWER would have been for a non-editable tree) and XTR'XCOL / XTR'TARGETCOL are set to 0. On re-entry, this combination will select the row but not start out in editing mode. Previously this was somewhat inconsistent, making it difficult for an application that cared to either detect, or control, whether the XTREE enters/exits in selection mode rather than editing mode. |
986.6.1 |
Fix various XTREE bugs with XTF_DRAGDROP when used in conjunction with editable cells or without XTF_MSEL. |
986.5.1 |
XTREE: On entering an editable tree with targetrow/targetcol=1,1 (or xrow/xcol=1,1), if the real column 1 was hidden, the edit was starting in the last column of the row, instead of the first. (This wasn't exactly a bug, since you are supposed to specify the real column number (based on the order of the column definitions) not the display column number. In the case where this is a re-entry and the previous exit was caused by a navigational key, then the initial cell would be determined by moving in the same direction until we hit an editable cell. On an initial load, there is no navigational history to guide us the app tries to put the edit focus on an invisible column, but defaulting to the first visible column seems to make more sense. |
986.4.1 |
EVENTWAIT wasn't treating checkboxes created by INFLD as INFLD controls, and thus wasn't exiting when the focus was put on one (assuming EVW_EDIT flag set). |
986.3.1 |
Increase maximum TEXTRECTANGLE paragraph length from 1K to 16K. |
986.3.2 |
HOST.LIT now supports the ability to call a subroutine when exiting A-Shell. |
986.3.3 |
SHLEXC.LIT 1.0(102) now works under all platforms. Under UNIX, requires ATE or at least ZTERM. (ZTERM doesn't support the wait flag though.) Help message updated to be more specific to the platform. |
986.3.4 |
Mouse click fix: Clean up yet another issue related the 985.3 and subsequent fixes to prevent click actions on logically suspended dialogs. This one was causing click events on the active modal dialog itself to be ignored (noticed in LEO and perhaps when using context menus on a dialog). |
986.1.1 |
FTP wildcard transfer fix: The test for an empty directory was not counting files with the hidden, read-only, system, or archive bits set. This was leading it to sometimes think the entire directory was empty even when it was full of files. |
986.0.1 |
Mouse click fix: Clean up problems with being able to click on controls that are not in the active modal dialog (that were introduced in 985.3). The way it is now, if there is a modal dialog, you should only be able to click on controls that are either in that modal dialog, or in any modeless dialog, or on the main window menu. Prior to 985.3 there were basically no limitations on this, and between 985.3 and 985.5, there were too many restrictions. |
986.0.2 |
AUI_MENU enhancement: When deleting a top level menu, the submenu items beneath it are now automatically deleted. This isn't really a bug fix because the documentation always said that you needed to delete the submenu items manually, but failing that, programs would exhibit a bug where running the program a second time would result in a top level menu that didn't work. |
986.0.3 |
Tab Control grid bug fix: In some cases, controls within a Tab control within an ALTPOS dialog were being spaced using the main window grid rather than the ALTPOS grid. |
986.0.4 |
AUI_IMAGE bug fix: Images within ALTPOS dialogs were not being sized using the ALTPOS grid. |
986.0.5 |
Fix problem with dialog scaling getting messed up after returning from nested maximized dialogs. |
986.0.6 |
FTP bug fix: Attempting to transfer all files (using * wildcard) from a PC directory which contained no files was causing A-Shell to hang. This remains a bug in the underlying wodFTPDLX.DLL module, but ASHW32 and ATE now avoid it by pre-testing for an empty source directory and just skipping the entire operation if there is nothing to transfer. |
986.0.7 |
FTP bug fix: Wildcard transfers from the PC to the host now fold the filenames to lower case. It is debatable whether this is a bug fix, feature, or mere change, but since files stored with upper case names are difficult to access from A-Shell/UNIX, folding them to lower case seems like a nice thing to do. (Note that the problem would normally only have occurred when transferring files not created under A-Shell/Windows, since it creates files in lower case.) |
986.0.8 |
|
986.0.9 |
ZTXFER wildcards are now supported if the client is ATE. |
986.0.10 |
|
986.0.11 |
TWAIN scanning support module (VICTW32.DLL) now included in the standard release package. |
986.0.12 |
ATE.EXE: Connection Properties dialog now forces itself on top of everything. This hopefully eliminates the problem of it getting hidden beneath other windows and giving the appearance that the main ATE window was hung. |
986.0.13 |
MX_GDICALC: Fix a problem in which the vertical spacing for the advance calculations was more affected by the default font of the printer than by the font specified in the MX_GDICALC call, resulting in the box either being too big or too small for the specified text. |
