[TESTCASE] khtml and table widths

[Ladislav Strojil]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thursday 18 July 2002 18:26, Vadim Plessky wrote:
> On Thursday 18 July 2002 7:55 pm, Dirk Mueller wrote:
> |  On Don, 18 Jul 2002, Michael Ritzert wrote:
> |  > is KHTML currently conforming to the CSS2 table layout algorithm as
> |  > described in
> |  > http://www.w3.org/TR/REC-CSS2/tables.html#auto-table-layout ? Or isn't
> |  > this the issue at hand here?
> |
> |  Not an issue. This is about malformed (i.e. overconstrained) tables,
> | like those from KMail, which set 100% on every column. when there is more
> | than
>
> So, shouldn't KMail be fixed?
> It's part of KDE, and what we can expect from the real-world sites, if we
> can't fix own program?..

Not really. KMail does not really want 100% for each column.
The table has two columns, the first one has no width speficied, the second 
one claims 100%. This seems perfectly reasonable to me. 

The algo really should be:
a) If columns with widths claim less then whatever percentage of the table is 
left after layouting those columns without width, give them exactly that and 
scale those without width specified to fill in the rest.
If there is no column with width specified, scale UP to occupy the space 
required.
This means that a column with size would never be scaled UP if the table 
contains a column with unspecified width.

b) If they claim more than is left, multiply by a coeficient to maintain 
relative sizes.

Would this work? 
Looking at Dirk's commits, this should not be to hard to achieve. I'll have a 
look at it and see if I can modify it myself.

Cheers,
Lada
- -- 
    ~       Ladislav Strojil, MFF UK
  ' v '               
 //   \\              
/(     )\    Powered by Penguin.
  ^ ' ^
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE9Nu/dTnMZPFqq4nwRAs16AKCiwU7boCBTPRrkJUAVs6sTiTIzpwCgyvxo
x+ywPUiq8W6m+93kKlkgwDE=
=ovMx
-----END PGP SIGNATURE-----