W3C home > Mailing lists > Public > www-style@w3.org > February 2012

RE: [css3-flexbox] in multipage multiline vertical flexbox, what does 1st page 2nd column contain?

From: Alex Mogilevsky <alexmog@microsoft.com>
Date: Thu, 16 Feb 2012 05:08:40 +0000
To: Tab Atkins Jr. <jackalmage@gmail.com>, Daniel Holbert <dholbert@mozilla.com>
CC: "www-style@w3.org list" <www-style@w3.org>
Message-ID: <D51C9E849DDD0D4EA38C2E539856928412E5BDE9@TK5EX14MBXC214.redmond.corp.microsoft.com>
± From: Tab Atkins Jr. [mailto:jackalmage@gmail.com] 
± Sent: Tuesday, February 14, 2012 6:08 PM
± On Tue, Feb 14, 2012 at 6:02 PM, Daniel Holbert <dholbert@mozilla.com> 
± wrote:
± > One more pagination question:
± >
± > Suppose we have flex items A, B, C, D, E, F, G in a vertical multiline 
± > flexbox, with a specified-height that fits 4 items vertically. 
± > Un-paginated, it looks like this:
± >
± >     A  E
± >     B  F
± >     C  G
± >     D
± >
± > Now, if we paginate this into pages tall enough for 2 items, which of 
± > the following should we expect?
± >
± > (1)  A  C
± >     B  D
± >     ----
± >     E  G
± >     F
± >
± > I think (1) is correct, since that's what I'd expect with text and 
± > IIRC multiline flexbox is supposed to behave like text, roughly.  I 
± > thought it was worth clarifying, though, because I don't think the 
± > pagination spec-text makes the answer to this question clear yet.
± Based on what fantasai told me of her and Alex's discussion (I wasn't 
± there for it ^_^), yes, (1) is correct.  First, linebreak based on 
± available space on the page, then continue with fresh lines on the next 
± page.

Yes (1) is correct.

± The only weird bit comes if, for example, A is taller than the remaining 
± space on the page.  It'll still get taken for the first line, but it'll 
± overflow to the next page as well.  The solution that fantasai outlined 
± was that the set of lines on the next page would start just below where A 
± ended, in essence making the first set of lines as long as A (though still 
± doing linebreaking for the second line on the first page based on the 
± "normal" length).

There is no good solution for that. I think it should simply overflow and do what overflow does.

In other words, if each portion of paginated flexbox would be replaced with a regular block of the same and with children of same size and position (maybe modeled as floats or positioned or tabled - whatever is the least weird), whatever overflow behavior is defined in that case - clip, overlap or whatever - should apply to flexbox too.

And BTW if that kind of overflow is detected flexbox should first make sure it is at top of the page.


Received on Thursday, 16 February 2012 05:09:14 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:12 UTC