Please enable JavaScript to view this site.

A-Shell Development History

1134.9.1

XTREE bug fixes: fix problem with column 1 disappearing due to changes in 1134.8.

1134.9.2

XTREE enhancements: The following column attributes can now be set as defaults by assigning them to the zero column: Fold=, Dspmin=, Dspwid=

1134.8.1

XTREE refinement: Advance Coldef clause ScrollMode=Tree may now be associated with column 0 to make it the default for all columns (rather than having to do it individually for each column containing combos).

1134.8.2

XTREE bug fix: Some column 0 attributes were getting reset when there were multiple complete column 0 coldef statements.

1134.8.3

XTREE bug fix: left/right arrow keys in split trees (when not in edit mode, i.e., in order to scroll horizontally) were causing A-Shell to crash.

1134.7.1

XGETARG bug fix: conversion from numeric types (in the caller) to S or X (in the receiving function/procedure/sbx) were allowing space for a trailing null. Thus passing the numeric value 78 to a routine expecting an S,2 parameter was resulting in "7" instead of "78". This was a pretty obscure case, and probably suggests questionable application design, but nonetheless was an A-Shell bug (since in A-Shell/Basic, strings do not need trailing nulls, and can hold numeric values, thus a two-byte string should be able to store the string representation of the number 99).

1134.7.2

SBXINP.SBX 2.1(7) restores a bug fix originally included in 2.1(4), but lost in 2.1(6) - bug was causing it to treat certain forms as if they were character acceptance lists.

1134.7.3

Support dropped for TER14.DLL (prior version of module underlying XTEXT). It was supercededby TER15.DLL serveral months ago, but A-Shell was still supporting the older version if the newer one wasn't present. (The cleanup is preparation for an eventual upgrade to TER16.DLL.)

1134.6.1

INFLD Combo bug fix: ||F and |F were both causing combo boxes to act like "fast" fields (i.e. exiting immediately upon clicking on an item in the drop-down list). This effect should be reserved for TYPE F alone.

1134.6.2

Installer fixes and improvements: (1) ersatz.ini, prtxls.ini and ashelp.mdf no longer get overwritten when updating; (2) Help menu links updated to reference online HTML docs; (3) Default miame.ini adjustments (was set up more for ATE than A-Shell).

1134.5.1

VUE bug fix (3.2(290)) - The autosave operation was sometimes causing some or all of the last line of the file to be duplicated over the current line.

1134.5.2

It is now possible to Disable ATSYNC using the Windows Registry.

1134.5.3

It is now possible to Restore Profiles with ATECFX that have previously been virtualized to HKCU registry entries under Vista.

1134.4.1

XPUTARG / XGETARG bug fix: variable sizes larger than 64K were being truncated to the mod-64K value when passed back and forth between functions, procedures, and SBX routines.

1134.3.1

INMEMO bug fix: whenever a character is entered into the last position of the current line, INMEMO waits for the next character to decide whether or how to wrap the current line. This was working correctly except in the case where the deciding character was an ESC, in which case the cursor was advanced to the start of the next row, but the ESC was otherwise dropped. The problem wasn't serious, except in the case of MMO_SIL mode where the ESC was the last character in the TEXT string, leaving INMEMO waiting for keyboard input but without any visual clues. The first ESC now causes INMEMO exit, as it should have all along.

1134.3.2

GUI scaling bug fix: Controls within certain nested control frames (like children of a TAB control) were not being scaled up when the dialog was maximized. (Introduced in 1134.0).

1134.3.3

AUI_CONTROL enhancement: when a dialog is created with a non- standard background color (see 1133.1), child static and groupbox controls that omit the BGC parameter (or which set it to the default of -2 or NUL_BGC) now assume the parent dialog background by default (eliminating the need to explicitly specify the dialog's BGC parameter when creating each child. Note that this does not apply to TPRINT/DPRINT, since they effectively set the BGC to -1 to use the "current" BGC as set by TAB(-3,x).

1134.3.4

XTREE enhancement: split and non-split trees now work better together. Previously, only one kind of tree could be operational active at one time, making it difficult to click between them.

1134.3.5

XTREE enhancement: the range of XTREE control IDs has been extended from 0-9 to 0-31. In other words, you can now have 32 trees defined at once. (Believe it or not, the previous limit of 10 was cramping some programmers' style, although in some cases, the issue wasn't so much the need to have more than 10 XTREEs in existence at the same time, but the desire to use hard-coded reference numbers for them.)

1134.2.1

GUI bug fix: close loophole in which AUI_WINDOW, and possibly other GUI query operations would return incorrect results, or even crash due to divide-by-zero, if called before any controls were created. (Problem may have only started with 1134.0)

1134.1.1

XTREE enhancement: new flag values in XTR'EXPANDLEVEL field may be used to save and restore the expand/collapse status of every row.

1134.0.1

Fix INMEMO bug: Control+E in MMO_DPG mode led to vertical scrolling anomalies.

1134.0.2

Compiler bug fix (425): Use of sizeof(var) within a block of code that was being skipped due to conditional compilation was nevertheless complaining if the var wasn't previously compiled.

1134.0.3

ISAM-A fix/refinement: Remove 2GB limit on IDX files. (Previously only ISAM DAT files could exceed 2GB.) Also, don't rewrite the IDX header during UPDATE'RECORD if keys don't change (minor optimization in most cases; major optimization for EFS mirroring.) New library version 6.11.

1134.0.4

GUI refinement: Maximized dialogs no longer cause their child dialogs to be affected in size. (Previously, a normal dialog launched from within a maximized dialog was likely to be scaled way up.)

1134.0.5

XTREE refinement: It is not recommended that you include binary fields in the data array, even if those fields are not referenced in the COLDEF, because certain byte values can get misinterpreted as part of the ATE command packaging. That said, people do it anyway, and this patch fixes a problem with some those binary values causing the data load to end prematurely. (Note that the fix only applies on the ATE server side, not the workstation side.)