1637.4.2

EIR

INPUT CSV #CH, ARY() refinement: an empty element is now added to the end of the array if the input data ends with a comma. So for example, the data line { one,two, } will return ARY() with an extent of 3, with the last field empty. Note that it is still the case that an empty line will return an array with a zero extent.

1637.4.1

Fix

XPUTARG: putting a null string ("") to an uninitialized dynamic variable was generating a spurious error 3. Problem was introduced in 1635.1.

1637.3.2

Fix

Compiler edit 860: compiler was failing to flag a syntax error for WRITECD and WRITETD statements lacking a comma or semicolon between each field. For example:

WRITECD #CH, A, B  C

This was mostly likely to occur accidentally when the statement was broken on to multiple lines. The result was that all the fields would be output, but there would be a missing delimiter in the output line matching the location of the missing comma in the input line.

Note that probably this should be applied to all forms of PRINT statements, but a space between PRINT arguments has long been treated equivalently to a semicolon, which might actually be the understanding and desired behavior for some programmers, so we're going to leave it that way for the traditional PRINT statements.

1637.3.1

Fix

INPUT CSV: null fields were being ignored when inputting into a base array. For example, if the data contained { field1,,field3 }, then the statement { input csv #ch, ary() } was ignoring the null second field and loading the third field (field3) into ary(2).

1637.2

Fix

INPUT CSV: when inputting into an X type variable, quoted fields were ending up with a spurious quote character on the end. This was mainly an issue when inputting into a dynamic array reference (see 1629.0)since it requires use of the X type; otherwise string or numeric types would mostly likely be used.

1637.1

Fix

XTREE: The RGBfgbg foreground color was not properly overriding the default red foreground for negative values in numeric columns.

1637.0.6

Fix

KILL.LIT 2.2B(126) works around a problem in which a loophole in MX_GETENV (see above) might cause it to hang.

1637.0.5

Fix

Tracing: certain traces were causing interference with other operations returning string values. For example, the MALLOC trace was causing MX_GETENV to return the terminal device instead of the correct value.

1637.0.4

Fix

MX_GETENV: when retrieving the entire list of environment variables into a dynamic string, the last several characters were being truncated.

1637.0.3

New

A new GDI printing direction option called Cursor-Relative Coordinates allows the x,y coordinates in all //GDI directives to be expressed relative to the current cursor position.

1637.0.2

Fix

XTREE: columns with cformat "BC" (see 1636.0) were in some cases causing subsequent columns to display with unwanted/unexpected colors.

1637.0.1

EIR

XTREE enhancement: footer text (XTR'SHOWFOOTER) may now include embedded CRLF line breaks in order to show multiple lines (same idea as for headers). For example: coldef += "1~15~Header::> Footer1" + chr(13)+chr(10)+"Footer2#~~"