First of all, odd-height character cell sizes are rare.
Most text video modes produce an even-height character cell.
Secondly, it is something that users can fix themselves.
Thirdly, this package has a relatively small (but dedicated!)
following.  Those three things together explain why no-one
has reported it before this.  I suspect that a number of
users have experimented with the package, found that certain
modes did not work, and just decided not to use them.  They
never bothered to investigate why they didn't work.  I myself
have been using svgatextmode for years and just this week
finally figured out why certain modes didn't work.

But you want proof?  Well, I thought it would be obvious.
a ten-pixel-high font simply won't fit in a nine-pixel high
character cell.  It's physically impossible.  But OK, you
want proof.  I'll give you a test case.  Using the FontSelect
statements as they are, select the 80x43x8 text mode.

stm 80x43x8

This mode produces 473 scan lines and an 8x11 character cell
size.  STM is going to try to use a 12-pixel-high font in this
character cell.  Press Enter a bunch of times if necessary,
to fill the screen with shell prompts.  Now carefully count
the number of lines on the screen.  Do you count 43?  I don't.
I count 39.  If you start typing a command at the shell prompt,
you will discover that you can't even see what you are typing.
The command is being typed below the physical screen.
Now issue

setfont lat1u-10

Now you can see all 43 lines of the logical screen.  You may see
about 4 lines of garbage after that, but at least you see
all 43 legitimate lines and you can see the command that you
are typing.  That's why I don't use character cell sizes that
don't have a font of the exact height to go with them.
But if you are going to support the odd-height cell sizes at all,
it's better to support them this way than the other way.
Maybe it's better to not support the odd sizes at all unless
you're going to ship some odd-height fonts to use with them.



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to