Changes to the handling of "position" and "positionAlign". #92
Conversation
|
This is a proposed alternative to #69 . I have taken a lot of your proposed text from there, but based it on the current spec and also applied it to positionAlign, since I maintain that it is an important and necessary addition to allow the different setting dimensions to be dealt with independently when authoring. I believe similar changes should be done for line and lineAlign. I also believe that we can re-introduce the automatic size changes, but I would like to see that done through introduction of an "auto" keyword for "size", see https://www.w3.org/Bugs/Public/show_bug.cgi?id=25660 . |
| @@ -4456,34 +4440,30 @@ | |||
| follows:</p> | |||
|
|
|||
| <dl class="switch"> | |||
| <!-- Q(Silvia): Why is "computed text position" dealt with again here??? The boxes are already | |||
| text positioned in step 8 and this should only be about percentage positioning the line position | |||
| direction. --> | |||
There was a problem hiding this comment.
Would be curious about feedback here...
There was a problem hiding this comment.
AFAICT this was the original issue in https://www.w3.org/Bugs/Public/show_bug.cgi?id=19178, namely that for snap-to-lines, all the previous positioning is ignored, instead positioning the bounding box of the text in both axes in a manner similar to CSS background-position. I think that bug should be reopened.
|
I would like this to be split into separate commits and branches, at least to this granularity:
Not necessarily in that order. Smaller commits makes reviewing easier and the resulting history easier to follow. |
|
OK, fair enough. I'll see if I can get these done today. |
Introduce both an "auto" value for "position" and "positionAlign" settings which depends on the "text alignment". Introduce a "computed text position" and "computed text position alignment" algorithm such that these can be set by script through VTTCue. This results in some simplifications in the rendering algorithm, in particular we now don't need positionSet and positionAlignSet variables any longer.
|
I'm gonna close this pull request, since we've pull out everything here into other pull requests. |
Introduce both an "auto" value for "position" and "positionAlign"
settings which depends on the "text alignment".
Introduce a "computed text position" and "computed text position
alignment" algorithm such that these can be set by script through VTTCue.
This results in some simplifications in the rendering algorithm, in
particular we now don't need positionSet and positionAlignSet variables
any longer.