34176 - Text in articles does not wrap correctly (table-layout: fixed)

Status: RESOLVED FIXED

(Core :: Layout: Tables, defect, P3)

Description

[chutzpah]

The text on http://www.geeknews.net does not wrap at the edge of the screen, it
goes off to the side of the scree fazr enough that it cannot even be all seen by
scrolling over. I am not sure whether this is a bug in the table engine, but I'm
guessing it probably is.

Comment 1

[R.K.Aa.]

i see this too in build ID 2000033113
When the page has finished loading and you move the scrollbar, it immediately
resize the right main table (grey backgr.), and the last character in each line
is half hidden - one can't scroll to see it.

Comment 2

[Cyril Bortolato]

Attachment: Simplified testcase

Comment 3

[Cyril Bortolato]

Marking bug confirmed.
The attached testcase shows a text that does not wrap at the end of the
line, and for which no horizontal scrollbar is available. As a result,
the hidden text is unaccessible.

Comment 4

[R.K.Aa.]

I looked at this page in NC4.7 - it's worse than it looks: There's supposed to
be a third table on that page, about the size of the left table. That one
completely vanishes since what is really the middle table takes all space on the
right side ++

Comment 5

[John Morrison]

Adding 'table-layout: fixed' to summary for searching purposes

Comment 6

[John Morrison]

*** Bug 35223 has been marked as a duplicate of this bug. ***

Comment 7

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

The problem dosn't seem to be that the text is wrapped wrong but rather that the
outer table dosn't expand to contain the inner table.

The outer table seems to check the width of the inner table before the inner
table gets width is calculated.

Simplifyed the testcase a bit further and added borders to illuminate the problem...

Comment 8

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Attachment: further simplifyed testcase

Comment 9

[John Morrison]

*** Bug 42207 has been marked as a duplicate of this bug. ***

Comment 10

[Bernd]

Attachment: patch

Comment 11

[Bernd]

Attachment: patch wihtout assertion

Comment 12

[Bernd]

The proposed patch fixes both testcases and the duplicates.
the ideas behind are:
transmit the minimum table size correctly.
don't compute remaining width and overallocation if the table size is unconstrained
in case of overspecified percent columns reduce them proportionaly.

The patch breakes the table regression tests!
especially 10269-1 and 10269-2 due to the correct implementation of
overAllocated, further 7243, due to the overAllocated percent reduction.

Comment 13

[Bernd]

*** Bug 27244 has been marked as a duplicate of this bug. ***

Comment 14

[Bernd]

*** Bug 53170 has been marked as a duplicate of this bug. ***

Comment 15

[karnaze (gone)]

r=karnaze, assumming the regression tests are ok. At some point the fixed layout
strategy needs to share code with the basic strategy (and this fix will be
moot). Until that happens however, this fix should be checked in.

Comment 16

[Chris Waterson]

Is it worth taking some time to factor the code now?

Comment 17

[Bernd]

My impression is that Chris has now more important items in his buglist (that 
was my reason to fix it), so I would propose to check it in, as it fixes the 
prominent geeknews.net. I would volunter to finish and to test Chris new code. 
And once this is done we could change. Otherwise my pessimistic estimate is, 
that we have the problems for another month (or 2??) and who knows when the next 
PDT phase will start....

Comment 18

[Chris Waterson]

ok, sr=waterson

Comment 19

[karnaze (gone)]

The patch is in.

Comment 20

[Amarendra Hanumanula]

QA contact update