Provide cell-pixel geometry in TIOCGWINSZ

[dankamongmen]

Abstract

Contour's implementation of the TIOCGWINSZ ioctl(2) doesn't fill in the following two optional fields:

               unsigned short ws_xpixel;   /* unused */
               unsigned short ws_ypixel;

Contour supports Sixel. Unfortunately, without this knowledge, it's difficult to use Sixel correctly for anything but trivial work.

Motivation

These fields ought be filled in, as they are by most terminals supporting bitmap graphics (all of which I am aware).

Specification

When TIOCGWINSZ is answered by Contour, the ws_xpixel field ought be filled in with the number of native pixels per cell width, and ws_ypixel ought be filled in with the number of native pixels per cell height.

When the cell-pixel geometry changes (for instance in response to a font change), the information ought be updated in subsequent ioctl(2)s, and ideally a SIGWINCH ought be delivered as it would upon a cell geometry change.

see dankamongmen/notcurses#1805

[christianparpart]

Oh.

Edit: should have been filled. I will be fixing that.

Edit2: also set x/ypixels at PTY construction and not just at resize events. (https://github.com/christianparpart/contour/blob/master/src/terminal/pty/UnixPty.cpp#L111)

[christianparpart]

Thanks @dankamongmen, I really appreciate your reports. I owe you a beer or two ;-)

[dankamongmen]

Thanks @dankamongmen, I really appreciate your reports. I owe you a beer or two ;-)

i'm strictly a krokodil / bathsalts / adrenochrome man, though i have been known to quaff some Sterno in my time

[dankamongmen]

btw i gave the wrong semantics here. i said they should have the cell-pixel geometry; it ought be the window geometry in pixels.

[dankamongmen]

confirmed, thanks much holmes!

 notcurses 2.3.4 by nick black et al on Contour 0.2.0-unreleased-master-96b8f89
  30 rows (22px) 130 cols (11px) (60.94KiB) 48B crend 256 colors+RGB
  compiled with gcc-10.2.1 20210110, 16B little-endian cells
  terminfo from ncurses 6.2.20201114
  avformat 58.45.100 avutil 56.51.100 swscale 5.7.100

 colors: 256 rgb: y ccc: y setaf: y setab: y
 sgr: y sgr0: y op: y fgop: y bgop: y
 rows: 30 cols: 130 rpx: 22 cpx: 11 (660x1430)
 sixel colorregs: 256

[dankamongmen]

  Width: 1446
  Height: 660

checks out (i've got a scrollbar which could easily be 16 pixels wide)!

[christianparpart]

Is this a problem? Just so i understand :)

…

On Tue, Jun 22, 2021, 06:58 Nick Black ***@***.***> wrote: Width: 1446 Height: 660 checks out (i've got a scrollbar which could easily be 16 pixels wide)! — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#241 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAN3O5UWLRGFHVOEI7HALDTUAJ7HANCNFSM47CA72TQ> .

[dankamongmen]

Is this a problem? Just so i understand :)

nope, just confirming via xwininfo that the rematerialized window pixel geometry was reasonable. it was! =]

[dankamongmen]

rows: 30 cols: 130 rpx: 22 cpx: 11 (660x1430) <----------- notcurses-info

Width: 1446 <--------------------- xwininfo
Height: 660

they're not an exact match, but i think the 16 extra pixels are probably due to the scrollbar.

[christianparpart]

Then I shall fix this

[dankamongmen]

no it's working! i must not be clear