This patch cleans up a style inconsistency.

Backspace key must generate the backspace character (\010) and
Delete key must generate the delete character (\0177). In
some systems the kernel configuration for erasing previous character
is \0177, so some programs (for example cat, ed, mail, ...), can not
understand the correct meaning of backspace. In this cases it is only
needed this command:

	stty erase 

This patch replaces current utf decoder with a new one, which is ~50
lines shorter and should be easier to understand. Parsing 5 and 6
sequences, if necessary, requires trivial modification of UTF_SIZ
constant and utfbyte, utfmask, utfmin, utfmax arrays.

These capabilities inform to programs how print in local printer
of the terminal.

This is very usefull in order to can select what is sent
to the plumber.

These new combinations generate the same behaviour (basically) of
vt102. It is a good way of communicating st with other programs.

[0] http://www.vt100.net/docs/vt102-ug/chapter2.html

This sequence is very useful because allows comunicate the content
of the terminal to another program.

This sequence print the current line. It is different to the
'printer on' sequence, where all the characters that arrive to the
terminal are printer. Here only the ascii characters are printed.

Add some note about the main reason why st is still that big.

This sequence control when the printer is enabled or disabled. This
sequence control the behaviour of the -o option.

Since we are using xmalloc, xrealloc ..., then it is not logical
call directly to strdup.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

The patch to add w3img support destroys our way to handle fps and so stop
wasting resources on fast scrolling. Due to w3img being a hack to display
images in an ugly way, is there no need to support this. Use some real way to
display images.

It is not defined when lateral effects of expressions are
calculated, so an expression as:

	--last >= gp && !(selected(last - gp, y)

is non portable.

The error condition was duplicated in two different switches.
This new version centralized the error handling in only one place.

If CONTROL + INSERT inserts a line then it seems natural CONTROL + DELETE
removes a line.

The AppMode and AsciiMode were inverted in the definition of this
combination.

Signed-off-by: Christoph Lohmann <20h@r-36.net>

Before this patch draw() calls drawregion which calls xdraws and then
updates whole window in one call thus overdrawing anything drawn by
w3mimgdisplay. After moving XCopyArea to xdraws it only updates the
regions which are being updated by XftDraw* functions. It may do a few
more calls to XCopyArea with this patch.

Signed-off-by: Amadeusz Sławiński <amade@asmblr.net>

for example
echo -e "\e[48;2;255;0;0m\e[38;2;0;0;255m test "
should render on red bg with blue fg

also now elinks works correctly when using 'truecolor' option
in preferences

Signed-off-by: Amadeusz Sławiński <amade@asmblr.net>

Sorry for another duplicated mail. I found the patch is malformed
significantly. I've been away from my laptop for a while, so I'm quite
unfamiliar with the settings on this system...

sbase did some interesting modifications to arg.h (basically it
was fixed an incorrect use of the _ namespace), and this commit
take this last version for st.

This fixes a bug that the parent tty gets resized whenever you launch
st through command line.

The problem was that ioctl was resizing cmdfd before it gets
initialized in ttynew. Since cmdfd is a global variable, its initial
value is 0, and consequently stdin was being resized.

Shitf modifier adds 12 to the function number of the function keys,
while Control adds 24, so Control + F4 generates F28

rmul means "exit underline mode", so a full reset of all
the attributes is not the correct way of exiting from
underline mode, because it is going to modify also another
attributes not related.

Since st is using now int32_t and uint32_t the inclusion of
stdint or inttype is mandatory, because in other case the
definition of these new types will not be known by the
compiler.

k0ga misktook applying patch of others. Sorry guys!!!!

This patch replaces long by int32_t. It saves
some memory on 64bit systems.

Signed-off-by: Christoph Lohmann <20h@r-36.net>

Conflicts:
	config.def.h
	st.c

Use correct type for Mousekey.b (XButtonEvent.button).

vt100 has support for two defined charset, G0 and G1. Each charset
can be defined, but in each moment is selected only one of both
charset. This is usually used selecting a national charset in G0
and graphic charset in G1, so you can switch between graphic
charset and text charset without losing the national charset
already defined.

st hasn't support for national charsets, because it is an utf8
based terminal emulator, but it has support for graphic
charset because it is heavily used, but it only supports G0,
without understanding G1 selection sequences, which causes some
programs in some moments can print some garbage in the screen.

This patch adds a fake support for multiple charset definitions,
where we only support graphic charset and us-ascii charset, but
we allow more of one charset definition.

This patch allow define G0 until G3 charsets, but only accepts
select G0 or G1, and it accepts some national charset definitions
but all of them are mapped to us-ascii.