Wedge

rev 2306
 3 files changed, 3 insertions(+), 3 deletions(-), 1006 bytes

+ Wess can now also handle $context variables. Dunno why I forgot about this one, but... (Subs-Cache.php)

! Fixed a minor bug in quick edit causing an XML error popup to show up. (JSModify.php)

! Last nail in the header font size bug's coffin. Pretty much. It's not really a bug, you see, it's just a series of commits I'm always postponing, and the codebase ends up being desynchronized. (index.css)

rev 2305 -- part one of my CSS fixes and updates. Including the header font size fix, ah ah!
 10 files changed, 81 insertions(+), 65 deletions(-), 3.64 KiB

* Many changes to the color converter in Wess, to make it more predictable. (Class-CSS.php)

  - Renamed 'stronger' to 'strength', 'darker' and 'lighter' to just 'luma', and '(de)saturated' to just 'saturation'.
  - Added an alias to 'luma', i.e. 'brightness', if you find it easier to grasp.
  - Regarding the above (and 'channels'), you may now 'force' a destination value by not using a sign.
  - If you still want to apply a value relative to the existing one, add a + or - sign before the number.
  - So, basically, darker($color, 10%) is now luma($color, -10%), and saturated($color, 10%) is now saturation($color, +10%)
  - As always, luma($color, -.1) applied to a color with a luma of 50% (.5) results in .4; luma($color, -10%) gets you .45, of course.

* These changes were applied to... (common.css, index.css, index.member.css, sections.css, Wilde/extra.css, Warm/extra.css, Wuthering/extra.css, Wine/common.css, Wine/extra.css)

! Fixed header border radius. (Wilde/extra.css)

! Fixed header font size. A long requested one... (Wine/extra.css, Warm/extra.css, Wuthering/extra.css)

* Better header font. My goal was mainly to remove Georgia from the list, because the latest Chrome Mobile will now substitute a serif font for it, instead of falling back to sans-serif as requested. (Wilde/extra.css)

So, in case you didn't know, the Wess engine for CSS pre-processing allows you to apply several functions to color values.

For instance, darker(#fff, 10%) would return something like #eee (not exactly that value, but you get the point.)

Yesterday, I was looking into taking rgb values without an alpha channel, and apply a channel to them. To do that, I did alpha(rgb(99,99,99), .1), hoping for it to return rgba(99,99,99, .1)... Of course, it didn't do that. It return rgb(99,99,99), because the '.1' was understood was Wess as "add .1 (10%) of the alpha channel's currently value to the variable". Given that the current value was 1 (100%), the final value would be just the same, as it's clamped to 1.

So, I looked into my code a bit, and yes, I never documented it, and yes I always meant to, and no I don't have much time for documenting my stuff these days.
It's still a pretty powerful system, but what you need to know is: you can never 'force' a value with it, you can only influence an existing value. For instance, to return rgba(99,99,99, .1), I should instead do alpha(rgb(99,99,99), -.9) or alpha(rgb(99,99,99), -90%).
That makes sense. But what when you don't know the original value, and you just want to set a default alpha value..?

I figured out that thing: if you add a sign before the number, it should be a relative value to apply, and with no sign (- or +), it should be an absolute value.
I then proceeded to rewrite my code, and am done now, but I'm at a loss when it comes to 'understanding' it all, and most importantly, helping others understanding it.

I'm going to try, but please tell me if there's a better way to do this all, or if I forgot anything.

alpha(rgba(9,9,9, .5),  33%) => rgba(9,9,9,  .33) (forced, so: 33% = 0.33)
alpha(rgba(9,9,9, .5), +33%) => rgba(9,9,9, .665) (0.5 + (0.5*33%) = 0.665)
alpha(rgba(9,9,9, .5), -33%) => rgba(9,9,9, .335) (0.5 - (0.5*33%) = 0.335

alpha(rgba(9,9,9, .5),  .33) => rgba(9,9,9, .33) (forced, so: .33)
alpha(rgba(9,9,9, .5), +.33) => rgba(9,9,9, .83) (0.5 + 0.33 = 0.83)
alpha(rgba(9,9,9, .5), -.33) => rgba(9,9,9, .17) (0.5 - 0.33 = 0.17)

So, basically, you can do any operations you like this way, but of course I'm concerned about the fact that percentage values aren't the same as floating point values when using signs. Is it too complicated? Or does it, on the contrary, make sense..?
Or maybe, even better, I just need to point people to this topic, so that it can make sense to them...? :P

Yep. Preview fails. :(

rev 2302
 53 files changed, 4 insertions(+), 31.98 KiB

+ Added, to the attic, my Cyna smiley set, originally designed for my Cyna and Noisen websites. Since it's a mash-up between my own roxxor skillz and other smiley sets, I won't be distributing it alongside Wedge, but still, it's nice to have a copy around. (other/images/cyna/*)

rev 2301
 5 files changed, 102 insertions(+), 68 deletions(-), 3.51 KiB

+ Support for >=, <= and !== in variable tests. The first two are really new, the latter is just a synonym of !=, since Wess always does loose matching anyway. (Class-System.php)

+ Finally added support for mixing variable and constant tests, e.g. you can do @if (admin || ($status == "admin")) && !ie; it is important to note this convention: if you're using a literal string that's the same as a Wedge system constant, you need to surround it in quotes (as in the example above). Otherwise, Wess will consider it to be the value of that constant, instead of the string itself. (Class-CSS.php, Class-System.php)

+ Added support for a few globals to Wess tests: $settings, $theme and $txt. I guess the latter is less useful, and I'm not going to add support for $txt like I did for JS files, but you could still want to use it for hacks (e.g. holding non-translation data in a variable). Also, $theme already has a similarly named variable, but it relies on logic: use $theme to get the theme's URL, and $theme&#91;'var'] to get a theme variable. (Subs-Cache.php)

+ Support for '@else if' in addition to @elseif in Wess. (Class-CSS.php)

! Fixed a couple of bugs with brackets use in Wess. Mostly, the line of CSS below failed, because Wess expected the first bracket to represent a surrounding bracket, rather than the first part in a long series of tests. As a result, it created an infinite loop, which I'll also be accounting for from now on (albeit in a crude way), because I don't want people to come with me with support requests because they forgot to close a bracket somewhere. (Class-CSS.php)

* Avoid redundant keywords in CSS filenames: between 'admin', 'member' and 'mUSERID', there can be only one. Thus, only the most specific of the three will remain. mID > admin > member. (Subs-Cache.php)

! Very, very bad commenazi. (index.css)

! Don't output #junk definitions if you can't see debug info in the first place. It saves about 5 lines of CSS, and it required two days of work and all the changes above. You can definitely say I'm an optimization freak. But seriously, the variable rewrite had been in my to-do list for too long. (index.css)

@ For the curious, the CSS line in question is:

$view_junk = $settings&#91;'db_show_debug_who_log']
@if $view_junk == "any" || (admin && $view_junk == "admin") || (member && ($view_junk == "regular" || $view_junk == "mod"))
...

And forgot to log this:

* Small color tweaks to make headers and linktree a bit brighter. (Wilde/extra.css)

rev 2299
 5 files changed, 32 insertions(+), 16 deletions(-), 1.95 KiB

+ Thought privacy now shows list type icons, instead of the generic list icon. (Security.php, script.js)

+ Ported the PrivacySelector code to PHP, so that it can also be hardcoded into the page HTML. Although I'm not sure I'll use that eventually (my plan is to go fully JS), it's still nice to have for now. (Security.php)

+ Topic privacy also gets list type icons, of course. (Post.template.php)

! Fixed thought privacy showing, in edit mode, the 'Admin' icon when selecting the 'Members' privacy type. (script.js)

! Translation error. (Login.french.php)

- Removed some unused code. (Post.php)