The cache is used to speed up detection of console handles, but with the
original three handles I was only thinking of actual console handles,
not general file handles. Five handles implies: stdout, stderr,
CONOUT$, log file and output file.
In v1.70 I had saved the cursor position relative to the window, but
that means should the window scroll, the cursor would appear to be
restored to the wrong position. It will still be wrong should the
*buffer* be scrolled, but that can't really be helped (well, possibly
it could, but it would be more effort than it's worth).
Just copying the history from the source:
recognize the standard handle defines in WriteFile;
minor speed improvement by caching GetConsoleMode;
keep track of three handles (ostensibly stdout, stderr and a file);
test a DOS header exists before writing to e_oemid;
more flexible/robust handling of data directories;
files writing to the console will always succeed;
log: use API file functions and a custom printf;
add a blank line between processes;
set function name for MyWriteConsoleA;
scan imports from "kernel32" (without extension);
added dynamic environment variable CLICOLOR;
removed _hwrite (it's the same address as _lwrite);
join multibyte characters split across separate writes;
remove wcstok, avoiding potential interference with the host;
similarly, use a private heap instead of malloc.
Added support for CHT & CBT (move forward/backward by tabs), DECAWM (don't
wrap at EOL), CRM (display control characters, but still perform newline)
and REP (repeat last character, including BEL, BS, TAB, LF and CR).
It always bugged me that newline would add an unneeded blank line due to
wrap, but not enough to do anything about it. For some reason, adding CRM
got me to thinking about it, so I finally did it.
Stopped \e[K from erasing the first character of the next line.
Restore cursor visibility on unload.
Clear screen (\e[2J) will scroll in a new window the first time it's used,
or the window has scrolled, or the cursor is on the last line of the buffer.
Restore Cursor Position (\e[u) will recognise screen size changes and limit
itself to the new boundaries.
ANSICON_EXC can now be used to exclude an entire program (including children).
This is achieved by simply not specifying an extension: ANSICON_EXC=program.exe
will just ignore program.exe (its DLLs will still be hooked, as will its child-
ren), but ANSICON_EXC=program will not hook program at all (which also means
its children will not be hooked).
The various LoadLibrary hooks would only hook the DLL that was specified - any
DLLs it loaded would be missed. That has now been rectified. Similarly, a DLL
that is injected via CreateRemoteThread, using LoadLibraryA or LoadLibraryW as
its ThreadProc, will now be hooked.
-p uses CreateRemoteThread, determining kernel32.dll & LLW dynamically.
Loading via LoadLibrary will remember the current attributes, restoring them on
unload.
Tweaked log output (remove quotes around CreateProcess command line; add an
underscore to 64-bit addresses).
ansicon.exe will really output (to the console) strings as Unicode.
Fixed ansicon.exe, if installed, restoring the default attributes, not current.
ansicon.exe will start with ANSICON_DEF (if defined and -m not used).
for ANSI-LLW.exe.
Set the code page so ansicon.exe can display some strings properly.
Expand wildcards for -t.
VC6 can now compile the 32-bit version; use it for the release binaries.
Improvements to the VC makefile.
Describe the sequences in a bit more detail.
Added environment variable ANSICON_EXC to specify modules that should not be
hooked. This should work around the nvd3d9wrap.dll issue. Since it helps to
know what the modules are, logging is now always available, controlled by -l
or ANSICON_LOG. A side-effect caused debugstr.c to move to util.c.
GUI programs are once again not hooked, unless run by "ansicon" directly or in
the ANSICON_GUI environment variable. Since not hooking still leaves ANSICON in
the environment, created ANSICON_VER as a dynamic-only variable, which can also
serve as a version check.
Due to an email requesting a reverse video option, realised I always take the
current attributes as default. This means if you turned on reverse and ran a
program, it would take the reverse as its default. Created ANSICON_DEF variable
to explicitly set the default attribute, using the current if it doesn't exist.
The reverse video option is done via a "negative" attribute (e.g. "-m-f0" is
reversed black on white, meaning you'll get white on black, with foreground
sequences changing the background). (The difference from "\e[7m" is that it
won't be reset on "\e[m".)
A child program will inherit the parent's modes (but not shift); the parent will
read the child's modes on exit (but not unload). The exception is "ansicon",
which will always start with the default modes and leave the parent unchanged.
Improved the AutoRun entry, only running "ansicon" if ANSICON_VER doesn't exist.
The "ansicon" command is always first.
Stopped -u implying -p; return the program's exit code; don't restore the
original color when just using -p; output error messages to stderr.
