<, > now compare real parts

This was the implicit behaviour in MATLAB, but not in Octave

funDef.txt

@@ -546,7 +546,8 @@ Z:	2	4	3		1	1	1		true	true	true	true	if ischar(in{1}), in{1} = double(in{1}); en
 X;	1	1	1		1	1	1		true	true	true	true	out{1} = acos(in{1});	inverse cosine (radians)	\matlab+acos+
 Y;	1	1	1		1	1	1		true	true	true	true	out{1} = asin(in{1});	inverse sine (radians)	\matlab+asin+
 Z;	2	2	2		1	1	1		true	true	true	true	out{1} = bsxfun(@atan2, in{1}, in{2});	four quadrant inverse tangent (radians; element-wise, singleton expansion)	\matlab+atan2+, element-wise with singleton expansion
-<	1	2	2	1	1	1	1		true	true	true	true	if numel(in)==1, in{2} = in{1}.'; end	is less than? (element-wise, singleton expansion)	\matlab+<+ (\matlab+lt+), element-wise with singleton expansion. If $1$ input: a second input is used given by the first transposed
+<	1	2	2	1	1	1	1		true	true	true	true	if numel(in)==1, in{2} = in{1}.'; end	is less than? (element-wise, singleton expansion)	\matlab+<+ (\matlab+lt+), element-wise with singleton expansion. For complex values, the real parts are compared. If $1$ input: a second input is used given by the first transposed
+													for k = 1:2, if ~isreal(in{k}), in{k} = real(in{k}); end; end; clear k
 													out{1} = bsxfun(@lt, in{1}, in{2});
 X<	1	3	1		1	2	1	[false true]	true	true	true	true	if numel(in)==2, out{1} = bsxfun(@min, in{1}, in{2});	minimum	\matlab+min+. If $2$ inputs: element-wise with singleton expansion. With more than $2$ inputs, the second is replaced by \matl+[]+. This function does not support flags \matlab+'omitnan'+ or \matlab+'includenan'+. Output is \matlab+char+ if there is $1$ input and it is \matlab+char+, or if there are $2$ inputs and both are \matlab+char+, or if there are $3$ inputs, second is empty and third is \matlab+char+. \sa \matl+X>+, \matl+Xl+
 													elseif numel(in)==1, [out{:}] = min(in{:});
@@ -561,7 +562,8 @@ Z<
 X=	0	inf	2	numel(STACK)	1	1	1		true	true	true	true	if numel(in)>=2, out{1} = isequal(in{:}); else out{1} = true; end	true if arrays are numerically equal	\matlab+isequal+. Works for any number of inputs
 Y=	2	2	2		1	1	1		true	true	true	true	if isnumeric(in{1}), in{1} = char(in{1}); end; if isnumeric(in{2}), in{2} = char(in{2}); end; out{1} = strcmp(in{:});	compare strings	\matlab+strcmp+. If first or second inputs are numeric they are converted to char
 Z=
->	1	2	2	1	1	1	1		true	true	true	true	if numel(in)==1, in{2} = in{1}.'; end	is greater than? (element-wise, singleton expansion)	\matlab+>+ (\matlab+gt+), element-wise with singleton expansion. If $1$ input: a second input is used given by the first transposed
+>	1	2	2	1	1	1	1		true	true	true	true	if numel(in)==1, in{2} = in{1}.'; end	is greater than? (element-wise, singleton expansion)	\matlab+>+ (\matlab+gt+), element-wise with singleton expansion. For complex values, the real parts are compared. If $1$ input: a second input is used given by the first transposed
+													for k = 1:2, if ~isreal(in{k}), in{k} = real(in{k}); end; end; clear k
 													out{1} = bsxfun(@gt, in{1}, in{2});
 X>	1	3	1		1	2	1	[false true]	true	true	true	true	if numel(in)==2, out{1} = bsxfun(@max, in{1}, in{2});	maximum	\matlab+max+. If $2$ inputs: element-wise with singleton expansion. With more than $2$ inputs, the second is replaced by \matl+[]+. This function does not support flags \matlab+'omitnan'+ or \matlab+'includenan'+. Output is \matlab+char+ if there is $1$ input and it is \matlab+char+, or if there are $2$ inputs and both are \matlab+char+, or if there are $3$ inputs, second is empty and third is \matlab+char+. \sa \matl+X<+, \matl+Xl+
 													elseif numel(in)==1, [out{:}] = max(in{:});

spec/funDefTable/funDefTable.tex

@@ -70,13 +70,13 @@
 \matl{X;} & 1 & 1 & \matlab+acos+ \\
 \matl{Y;} & 1 & 1 & \matlab+asin+ \\
 \matl{Z;} & 2 & 1 & \matlab+atan2+, element-wise with singleton expansion \\
-\matl{<} & 1--2 (2 / 1) & 1 & \matlab+<+ (\matlab+lt+), element-wise with singleton expansion. If $1$ input: a second input is used given by the first transposed \\
+\matl{<} & 1--2 (2 / 1) & 1 & \matlab+<+ (\matlab+lt+), element-wise with singleton expansion. For complex values, the real parts are compared. If $1$ input: a second input is used given by the first transposed \\
 \matl{X<} & 1--3 (1) & 1--2 (1 / 2nd) & \matlab+min+. If $2$ inputs: element-wise with singleton expansion. With more than $2$ inputs, the second is replaced by \matl{[]}. This function does not support flags \matlab+'omitnan'+ or \matlab+'includenan'+. Output is \matlab+char+ if there is $1$ input and it is \matlab+char+, or if there are $2$ inputs and both are \matlab+char+, or if there are $3$ inputs, second is empty and third is \matlab+char+. \sa \matl{X>}, \matl{Xl} \\
 \matl{Y<} & 1--3 (1 / 2) & 1 & \matlab+cummin+. Output is \matlab+char+ if first input is. \sa \matl{Y>} \\
 \matl{=} & 1--2 (2 / 1) & 1 & \matlab+==+ (\matlab+eq+), element-wise with singleton expansion. If $1$ input: a second input is used given by the first transposed \\
 \matl{X=} & 0-- (2 / $^\ddagger$) & 1 & \matlab+isequal+. Works for any number of inputs \\
 \matl{Y=} & 2 & 1 & \matlab+strcmp+. If first or second inputs are numeric they are converted to char \\
-\matl{>} & 1--2 (2 / 1) & 1 & \matlab+>+ (\matlab+gt+), element-wise with singleton expansion. If $1$ input: a second input is used given by the first transposed \\
+\matl{>} & 1--2 (2 / 1) & 1 & \matlab+>+ (\matlab+gt+), element-wise with singleton expansion. For complex values, the real parts are compared. If $1$ input: a second input is used given by the first transposed \\
 \matl{X>} & 1--3 (1) & 1--2 (1 / 2nd) & \matlab+max+. If $2$ inputs: element-wise with singleton expansion. With more than $2$ inputs, the second is replaced by \matl{[]}. This function does not support flags \matlab+'omitnan'+ or \matlab+'includenan'+. Output is \matlab+char+ if there is $1$ input and it is \matlab+char+, or if there are $2$ inputs and both are \matlab+char+, or if there are $3$ inputs, second is empty and third is \matlab+char+. \sa \matl{X<}, \matl{Xl} \\
 \matl{Y>} & 1--3 (1 / 2) & 1 & \matlab+cummax+. Output is \matlab+char+ if first input is. \sa \matl{Y<} \\
 \matl{Y?} & 0 & 1 & answer why. Sort of \\