Fix issue #106, to make it work on a real DEC VT320

[sblendorio]

Fix issue #106

• Using A_REVERSE attribute (a reversed space), now the digit are visible also on a real hardware DEC VT320
• The blinking cursor is hidden at the start, and is re-enabled at the end of the execution. I used printf() with ANSI code because ncurses "curs_set" function did not work on the real hardware
• The small "glitches" of the date's window are now fixed throw 2 "box-drawing characters" (ACS_TTEE)
  

[anarcat]

would be nice to have comments inline in the code to explain where that magic come from.

also please provide a screenshot of what it will look like in modern terminal emulators, e.g. like xterm and gnome-terminal. i know it's backwards but most people will use ttyclock there...

[sblendorio]

Added comments to clarify "magic strings" and simplified some expression. Tomorrow I will test it on both real VT320 and "normal" Terminal window on Xwindow screen and will post screenshot here

[sblendorio]

Just tested it on both real DEC VT320 and terminal emulator under Xwindows after yesterday's commit.

These are the results:

It seems to be O.K.

[anarcat]

those seem a little ... terse. what is "on"? color? expand the variable.

why isn't this a #define too?

[sblendorio]

Ok, I'm going to use more significant names for "on/off", for example "pixel_on" and "pixel_off".
I've not used a #define because the aim is to calculate the value at the beginning of the function and use it immedialy in the inner for cycle. Since the #define is just a text replacement, it does not implement precalculated values.

[sblendorio]

@anarcat would you use different names than "pixel_on" / "pixel_off"? Can you suggest me? For example:

• colored_reverse
• transparent_normal

would be more descriptive?

[anarcat]

i frankly have no idea what's going on here, so i'll trust your judgement blindly.

[sblendorio]

I will explain:

pixel_on is a bitmask made by:

• COLOR_PAIR(2), defined by "init_color" here https://github.com/xorg62/tty-clock/blob/master/ttyclock.c#L540
• A_REVERSE is the attribute that defines reverse characters (reversed space is a big "pixel")

pixel_off is anothter bitmask made by:

• COLOR_PAIR(0) which means "not visible"
• A_NORMAL is "not reverse" attribute

so I use "pixel_on" to draw the "block characters" which are reversed spaces, supported by both xterm and DECterm, and "pixel_off" to draw the "empty space".

is it good to merge it? ;)

[anarcat]

Just tested it on both real DEC VT320 and terminal emulator under Xwindows after yesterday's commit.

These are the results:

It seems to be O.K.

awesome, thanks, looks good.

[anarcat]

On 2022-02-08 01:32:49, Francesco Sblendorio wrote: @sblendorio commented on this pull request. > @@ -218,6 +218,8 @@ void draw_number(int n, int x, int y) { int i, sy = y; + unsigned int on = COLOR_PAIR(2) | A_REVERSE; + unsigned int off = COLOR_PAIR(0) | A_NORMAL; I will explain: pixel_on is a bitmask made by: - COLOR_PAIR(2), defined by "init_color" here https://github.com/xorg62/tty-clock/blob/master/ttyclock.c#L540 - A_REVERSE is the attribute that defines reverse characters (reversed space is a big "pixel") pixel_off is anothter bitmask made by: - COLOR_PAIR(0) which means "not visible" - A_NORMAL is "not reverse" attribute

those explanations would be tremendously more useful as comments in the source code. ;)

so I use "pixel_on" to draw the "block characters" which are reversed spaces, supported by both xterm and DECterm, and "pixel_off" to draw the "empty space".

then those are good names.

…

-- Perl is "some assembly required". Python is "batteries included". PHP is "kitchen sink, but it’s from Canada and both faucets are labeled C". - Alex Munroe, PHP: a fractal of bad design

[sblendorio]

@anarcat just added the comment for "pixel_on" and "pixel_off"

[anarcat]

awesome, LGTM.

[sblendorio]

awesome, LGTM.

no merge?

[anarcat]

On 2022-02-11 01:05:41, Francesco Sblendorio wrote: > awesome, LGTM. no merge?

i'm actually not the owner of this repository. typically, i approve MRs, then let them sit a while to give them a chance to see changes, then i merge if nothing happens. -- In a racist society it is not enough to be non-racist, we must be antiracist. - Angela Davis

[dirtslayer]

i ended up having to comment out the lines 70, 71
// init_pair(0 ...
// init_pair(1,
so that i can use gnome-terminal profiles to pick the colors.
seems like a command line option to disable colors and use terminal profile might be handy

[anarcat]

hmm... why do we mess with colors here at all then?

[dirtslayer]

to support backward compatibility we could leave the current color options, but adding an additional command line option, perhaps --no-color or -nc could disable the color setting and use the terminals 3 colors ( background, text, highlight) whatever they are called.