Height tests are broken

[BryanCrotaz]

I've fixed some tests, but those where the grid is supposed to fit the parent's available space are broken.

I'm not sure I like the algorithm you've used in calculateHeight which works out how the grid fits in its parent. The method you've used wouldn't update if the parent size changed.

Is there any reason why we couldn't set height: 100% instead?

[shaunc]

Hmm... I guess that if we are in "take up space in parent mode" we could use height=100%... however, body still needs explicit height I think, so need to listen to resize of parent....

[shaunc]

glanced at this ... didn't figure out how they broke yet though -- some changes I don't understand without a bit more time to think.

[BryanCrotaz]

They have always been broken

Bryan Crotaz
Silver Curve

On 14 Sep 2015, at 07:26, Shaun Cutts notifications@github.com wrote:

glanced at this ... didn't figure out how they broke yet though -- some
changes I don't understand without a bit more time to think.

—
Reply to this email directly or view it on GitHub
#25 (comment).

[shaunc]

well --- were working for me when I pushed them, but aren't working now. Of course could have been always broken for you on windows/IE

[shaunc]

(for that matter, tests used to work in firefox... :) ... hmm... I haven't figured that out yet either)

[BryanCrotaz]

I made some changes to the way the asserts were done last week which fixed
a number of them

Bryan Crotaz
Silver Curve

On 14 Sep 2015, at 07:40, Shaun Cutts notifications@github.com wrote:

(for that matter, tests used to work in firefox... :) ... hmm... I haven't
figured that out yet either)

—
Reply to this email directly or view it on GitHub
#25 (comment).

[shaunc]

It used to be that element heights (header/body/footer) were set explicitly. They seem not to be anymore, but the code depends on this, I think.

[shaunc]

For instance:

_bodyShellHeight() {
  // calclulate the height of content except contents of body
  var headerStyle = this._actualHeaderStyle();
  var bodyStyle = this._actualBodyStyle();
  var footerStyle = this._actualFooterStyle();
  var _N = val=> val == null ? 0 : parseFloat(val);
  var shellHeight = 0;
  shellHeight += _N(headerStyle.offsetHeight) - _N(headerStyle.bottomBorderWidth);
  shellHeight += Math.max(
    _N(headerStyle.bottomBorderWidth), _N(bodyStyle.topBorderWidth));
  shellHeight += Math.max(
    _N(footerStyle.topBorderWidth), _N(bodyStyle.borderBottomWidth));
  shellHeight += _N(footerStyle.offsetHeight) - _N(footerStyle.topBorderWidth);
  return shellHeight;    
},

The idea was that user might want to use css on header/body/footer -- and we should respect if so -- so I read computed heights. However, now the header is 0 height and content within it is visible because of overflow. I'm not sure how we want to resolve.

[shaunc]

Do you know why the header & footer are set to 0 height? I could go down to look at cell height. What I haven't figured out is how the body gets its height currently.

[shaunc]

In chrome seems to figure out height of header despite css. But gets confused in different ways in both safari and firefox... I'll try to look more carefully at it after I get some sleep. It would be helpful for me to understand how you think it should work. I thought the idea was to read the height of "rest" (excluding body) and calculate body height from that.

But how is height of "rest" being determined? It would seem that in chrome (and mostly in safari, though there seems to be some other glitch) it suffices that the header cells have content of given height.

In firefox for me, the "filler" element in header gets rendered to height 400px and this pushes out calculated height of header itself. Hmm...

Going to bed, now -- ... if you could write a note on how the heights of header/footer are supposed to be set that might help me tomorrow.

[BryanCrotaz]

table layout is used. If height is zero the row auto sizes to content.
Remaining rows split remaining space evenly.

Bryan Crotaz
Silver Curve

On 14 Sep 2015, at 08:02, Shaun Cutts notifications@github.com wrote:

Do you know why the header & footer are set to 0 height? I could go down to
look at cell height. What I haven't figured out is how the body gets its
height currently.

—
Reply to this email directly or view it on GitHub
#25 (comment).

[shaunc]

(came back up after brushing teeth :) ) hmm.. so what is now happening in firefox is that header (for scroll example) or header & footer (for overview table) are expanding to split rest of space (ie all of space) equally regardless of their content.