2000-03-04 10:02:11 +00:00
|
|
|
|
|
|
|
Random remarks about the ncurses support for GoldEd+
|
2000-03-24 22:45:59 +00:00
|
|
|
by Jacobo Tarrio, 2:348/105.93@fidonet <jtarrio@iname.com>
|
2000-03-04 10:02:11 +00:00
|
|
|
|
2000-07-09 00:19:24 +00:00
|
|
|
This version of GoldEd+ bears the fifth public release of the ncurses
|
2000-03-04 10:02:11 +00:00
|
|
|
implementation of its video code; it's still considered beta, so expect some
|
|
|
|
strange behaviour while it is being completed.
|
|
|
|
|
|
|
|
It is being done in order to allow any Unix clone with ncurses installed
|
|
|
|
(and, probably, any implementation of curses) to run GoldEd+, independently
|
|
|
|
of the character set used by the system and of the terminal type; GoldEd+
|
|
|
|
will run fine from the console with a cyrillic character set, from a xterm
|
|
|
|
with the iso-8859-1 character set, or from a monochrome VT-100 dumb
|
|
|
|
terminal. It might even be possible in the near future to use GoldEd+
|
|
|
|
compiled for DOS, Win32 or OS/2, and linked with PDCurses.
|
|
|
|
|
|
|
|
This file tries to list some already known bugs along with other nasty
|
|
|
|
"features" of the code. Expect most of these bugs to disappear and a lot of
|
|
|
|
fresh new ones to show up in the next public release of the ncurses support
|
2000-03-10 20:46:26 +00:00
|
|
|
:-) However, I do not always succeed, so, old, already existing bugs are
|
2000-03-04 10:02:11 +00:00
|
|
|
marked with '!', whilst new ones are marked with '*'.
|
|
|
|
|
|
|
|
If you find any bug not listed here, please report to Jacobo Tarrio,
|
2000-03-24 22:45:59 +00:00
|
|
|
2:348/105.93@fidonet <jtarrio@iname.com> ASAP to get it fixed soon, or,
|
2000-03-10 20:46:26 +00:00
|
|
|
even better, fix it yourself and send me the patch :-)
|
|
|
|
|
2000-07-09 00:19:24 +00:00
|
|
|
=== 2000-07-09: Fifth public release
|
|
|
|
Known bugs:
|
|
|
|
|
|
|
|
* None 8-)
|
|
|
|
|
|
|
|
Notes:
|
|
|
|
|
|
|
|
* Changed function key management to work as expected in Debian Potato. That's
|
|
|
|
all I can do. I will do no more. Reason: function keys are not portable -
|
|
|
|
repeat - they are not portable.
|
|
|
|
|
|
|
|
|
2000-04-19 21:27:07 +00:00
|
|
|
=== 2000-04-19: Fourth public release
|
2000-03-10 20:46:26 +00:00
|
|
|
Known bugs:
|
|
|
|
|
|
|
|
! Messages from the program viewed before entering full-screen are not
|
|
|
|
displayed properly
|
2000-03-24 22:45:59 +00:00
|
|
|
|
|
|
|
Notes:
|
2000-03-10 20:46:26 +00:00
|
|
|
|
2000-03-24 22:45:59 +00:00
|
|
|
* I'm downgrading this one from bug to note: "Screen may become corrupted if
|
|
|
|
someone other than GoldEd+ writes on the terminal". As it happens with
|
|
|
|
other platforms too, I'm considering it no longer a ncurses-specific bug.
|
2000-04-19 21:27:07 +00:00
|
|
|
* Trouble with your function keys? I'm sorry but I cannot help it. Every
|
|
|
|
kind of terminal manages function keys differently, so they cannot be
|
|
|
|
relied upon. Well, there IS a way to get function keys et.al. reliably and
|
|
|
|
all, but that would make GoldED+ unportable, bloated and fragile, and
|
|
|
|
neither I nor you want it, do you? Workaround: map these functions to
|
|
|
|
regular keys.
|
2000-03-04 10:02:11 +00:00
|
|
|
|
2000-07-09 00:19:24 +00:00
|
|
|
|
2000-03-04 10:02:11 +00:00
|
|
|
=== 2000-02-17: Third public release
|
|
|
|
Known bugs:
|
|
|
|
|
|
|
|
! Statusline displays a couple of odd characters (after GoldEd+'s name and
|
|
|
|
before the current time)
|
|
|
|
! Screen may become corrupted if someone other than GoldEd+ writes on the
|
|
|
|
terminal
|
|
|
|
! Messages from the program viewed before entering full-screen are not
|
|
|
|
displayed properly
|
|
|
|
* The cursor grows very big every 30 seconds when editing a message
|
|
|
|
(possible reincarnation of a similar bug in former versions)
|
|
|
|
* The splash screen looks ugly, but not as much as before :)
|
|
|
|
|
|
|
|
Notes:
|
|
|
|
|
|
|
|
* ESCDELAY is now set by default to 50ms. If it is too short for you, you
|
|
|
|
can set the ESCDELAY environment variable to override it (under ncurses).
|
|
|
|
It must hold an integer value representing the delay in milliseconds.
|
|
|
|
Quite useful over wide area network links or on very-very-i-really-mean-it
|
|
|
|
slow terminals. Undocumented in ncurses 4, but functional. Documented
|
|
|
|
in ncurses 5 and above (I believe).
|
|
|
|
|
|
|
|
|
|
|
|
=== 2000-02-05: Second public release
|
|
|
|
Known bugs:
|
|
|
|
|
|
|
|
! The splash screen looks a bit ugly :-)
|
|
|
|
! The cursor disappears every 30 seconds when editing a message
|
|
|
|
* Statusline displays a couple of odd characters (after GoldEd+'s name and
|
|
|
|
before the current time)
|
|
|
|
* Screen may become corrupted if someone other than GoldEd+ writes on the
|
|
|
|
terminal
|
|
|
|
* Messages from the program viewed before entering full-screen are not
|
|
|
|
displayed properly
|
|
|
|
|
|
|
|
Notes:
|
|
|
|
|
|
|
|
* I have "ment" the escape key affair (see below) by using some sort of
|
|
|
|
ncurses version detection and the so. If this misfeature (for I'm sure it
|
|
|
|
is going to give more trouble than good) annoys you, you can change the
|
|
|
|
delay after escape by modifying the "ESCDELAY = 50;" assignment at the
|
|
|
|
file 'goldlib/gall/gkbdbase.cpp', or by commenting it out.
|
|
|
|
|
|
|
|
Anyway, you should perhaps be able, with this default setting, to "press"
|
|
|
|
ALT on terminals that don't have this, by pressing ESC and the desired key
|
|
|
|
very quickly :-)
|
|
|
|
|
|
|
|
|
|
|
|
=== 2000-01-25: First public release
|
|
|
|
Known bugs:
|
|
|
|
|
|
|
|
* Tab key does not work properly
|
|
|
|
* Boxes are seen fine (at least with Latin-1) but lines are still drawn
|
|
|
|
with hyphens (-), plus signs (+), vertical lines (|) and the so.
|
|
|
|
* The splash screen looks a bit ugly :-)
|
|
|
|
* The box in the screensaver leaves traces when moving
|
|
|
|
* The cursor disappears every 30 seconds when editing a message
|
|
|
|
|
|
|
|
Notes:
|
|
|
|
|
|
|
|
* When pressing the escape key, you must wait for 1 second for GoldEd+ to
|
|
|
|
take action. This is due to the way most terminals send function keys and
|
|
|
|
keypad presses; they send ESC followed by some characters, and curses
|
|
|
|
waits for these characters to arrive. There is a way to shorten the
|
|
|
|
timeout to, say, 30ms, which is suitable for a 300-baud terminal and still
|
|
|
|
too fast for humans to notice, but it would make GoldEd+ dependant of a
|
|
|
|
particular implementation of curses, and of a particular version of that
|
|
|
|
implementation; also, it would make GoldEd+ unusable over a network or
|
|
|
|
with some dumb terminals or terminal emulators; this way you can press
|
|
|
|
ESC followed immediately by an A to "press" ALT-A.
|
|
|
|
|