Tweak mode-line faces

[ryanartecona]

This PR

1. removes :box from mode-line faces
2. adds mode-line-inactive overline/underline colors

The reason for (1) is that the :box property isn't applied the same way to xpm images or non-text spaces the same way as it is applied to normal text within the mode-line. Or something like that. This resulted in weird underline/overline discontinuities in the presence of xpm images or other non-text stuff in the mode-line when high-contrast-mode-line was nil (see below). Powerline uses xpm images as separators.

Appropriate underline/overline properties are now relied upon exclusively, and IIUC, were responsible for the lines seen before this PR anyway in normal (text) mode-lines. To my knowledge, this shouldn't affect normal mode-lines.

The reason for (2) is to specify appropriate underlines/overlines in support of (1). Colors shouldn't change except that the inactive mode-line now gets an overline of s-line and an underline of base01 when high-contrast-mode-line is t, where before the overline and underline would both be base02 in that case, and difficult to see (see below).

Before:

• with high contrast:
• without high contrast:

After:

• with high contrast:
• without high contrast:

[thomasf]

I'll have a look at this soon, I'll have to verify it with a couple of more things before merging.

[ryanartecona]

In case it helps focus on things, what most concerned me were (a) whether removing :box breaks the face in certain environments that depend on it (I am new to emacs and only use it in a standalone Mac app), and (b) the header-line face, which I didn't touch, also uses a :box but with a line-width of 2, so I was unsure if/how my changes shpould be translated to that face. I also don't use a header-line (does anyone?).

I tested neither (a) nor (b).

On Apr 1, 2015, at 3:57 PM, Thomas Frössman notifications@github.com wrote:

I'll have a look at this soon, I'll have to verify it with a couple of more things before merging.

—
Reply to this email directly or view it on GitHub.

[c256]

Header-lines are used by a few modes, including Info, ruler, some tabbed-buffer emulation, and eww. I'm willing to bet that you actually do use a header line now and then and just don't notice it.

Have you tried your change with emacs in a terminal?

[thomasf]

IIRC I had to do something specific for mode/header lines to work well together.. I think It has something todo with avoiding double lines between the header/mode or something.. Not saying that this patch breaks anything, just adding information for the moment...

Here's how the header/mode lines looks without this patch:

[ryanartecona]

@c256 re header-line: You're probably right. Do you happen to know if any of those try putting images in the header-line? If so, I can try to install if necessary and try it out; if not, it may be ok to leave the header-line face as-is as far as this PR is concerned.

re non-gui emacs: I just tried it in iTerm with xterm-256colors on, both before and after this PR's changes. It looks unaffected to me—the mode-line just gets a textual underline and no overline instead of graphical underline/overline.

[ryanartecona]

@c256 scratch that, I didn't do it properly before. This change adds an underline to both active and inactive mode-line faces, where they had no underline before, when high-contrast-mode-line is on; it does not affect the mode-line faces without high-contrast on; they had textual underlines already, and still do.

Should the mode-line underline only be set if emacs is loaded in a GUI environment? And if so, what function/variable reports that, so I can update the PR?

[ryanartecona]

Any update on this? Anyone care whether an underline gets added to active/inactive high-contrast modelines in a terminal? Any related changed need to happen on the header line?

[thomasf]

I never got to this becase of all the other smalle *-line issues that popped up and it seemd that it was for the best to resolve those first.. I will try to include to check the things I want to and merge this the next time I work on solarized.

[ryanartecona]

@thomasf works for me! I agree with you. No rush here, just didn't want it to be forgotten :)