160317 - Migrate TEC fix from bug 111728 to Chimera branch

Status: RESOLVED FIXED

(Camino Graveyard :: Page Layout, defect)

Description

[Pierre Baguis]

This bug report concerns Chimera v0.4.

Overview Description: Greek fonts are not displayed correctly in Chimera.

Steps to Reproduce: View any greek web page ( for example
http://www.ert3.gr/news/news.htm ).

Actual Results: Characters with accents, final "sigma" and the "pi" character
are not displayed in the same manner as the rest of the document. This rest of
the document, for some reason, appears in boldface slanted font. The spacing
between characters is also not correct. For comparison, try to view the same
page with Omniweb 4.1, which renders and displays correctly all the greek
characters and takes care of spacing too.

Expected Results: Display the greek pages like Omniweb 4.1.

Comment 1

[Greg K.]

I think you're encountering bug 111728 here. When that's fixed, it will need to
propagate to Chimera.

Comment 2

[Winnie Lam]

hmm... apparently the reporter for bug 150924 thinks that greek displays properly.

adding sfraser and ftang to cc list.

Comment 3

[Dimitris Tsarouhas]

The Same problem exists on mozilla 1.1 under MacOS X 10.2 too.
Can be reproduced when someone sees a greek web page (example http://www.helmug.gr )
when using any international 10.1 or 10.2 MacOS.

I think if ATSUI was fully supported on mozilla and chimera code, that maybe
none of these problems would appear.

Rainbow Computer has a piece of software "Greek Update" for MacOS X 10.1.4 that
has a pack of fonts and Greek Script on MacOS X. When Mozilla and chimera is
used under MacoS X with Greek Update (as local mac distributor calls it) greek
display works fine.

Omniweb 4.1 has correct greek character display.

Comment 4

[Kostas Papadimas]

Yep, Mozilla 1,1 under Macos 10,2 fails to render a number of Greek fonts.
Mozilla on MacosX  should render Greek pages without additional patches to
system. I think we need a separate bug for this.

Comment 5

[Dimitris Tsarouhas]

Attachment: display of greek in chimera 0.5

Bad display of greek in chimera 0.5. Problem probably because of use of TEC in
display engine of chimera. Compare this picture with the picture
omniweb4.1greek. omniweb has correct display of greek.

Comment 6

[Dimitris Tsarouhas]

Attachment: display of greek in omniweb browser (correct)

In comparison to chimera, omniweb uses the native text system of macos x and
shows greek characters correctly on MacOS 10.2 international.

Comment 7

[Greg K.]

Dimitris, what URL are those screenshots demonstrating?

Comment 8

[Dimitris Tsarouhas]

They are demonstrating

http://www.helmug.gr

at left...

It is the Hellenic Macintosh User Group web page
It uses postnuke and encoding Greek-ISO 8859-7
[e-mail]

Comment 9

[Greg K.]

Dimitris, your Mozilla screenshot is taken in 1.1. Post-1.1 builds with the fix
for bug 111728 look much better, spacingwise. Best possible display will arrive
with bug 121540.

Reassigning to Page Layout. Bug 111728's fix needs to make it's way over to the
Chimera branch if it can.

Comment 10

[Simon Fraser [no longer active]]

Taking to migrate TEC fix.

Comment 11

[Greg K.]

*** Bug 175006 has been marked as a duplicate of this bug. ***

Comment 12

[Niklas Dougherty]

Wait a minute... I haven't been using Chimera for a while, but 0.5 is
effectively behaving as bad as Mozilla rendering text. This wasn't always so.
Chimera 0.3 was using native ATSUI, but now seems to incorporate all the bad
hacks from Mozilla. What's the point of this? Then there is no need for Chimera,
since it doesn't really behave like a Cocoa app. People sensitive to good
rendering might as well continue using Omniweb.

The text below was rendered perfectly in Chimera 0.3, but now displays worse
than Mozilla, with Kanji spacing, wrong font, wrong weight and wrong size for
eth and thorn:

"Hróðstein ok Eilifír Áki ok Hákon reistu Þeir sueinar eftir sin faður kumbl
kenniligt eftir Kala dauðan þy mun góðs mans umgætit verða meðan stein lifir ok
stafir rúna"

Why this terrible regression?

Comment 13

[Greg K.]

Niklas, this bug is about migrating Mozilla's TEC fix from bug 111728 to the
Chimera tree. See comment 9 and comment 10.

Comment 14

[Simon Fraser [no longer active]]

> Why this terrible regression?

Because the ATSUI code was significantly slower.

Comment 15

[Niklas Dougherty]

> Because the ATSUI code was significantly slower.

Hence this bug. Hence no Greek. Hence no Latin Slavic. Hence no Icelandic. Hence
no Old English/Old Swedish/Old Danish. Hence no Cyrillic. Hence no Runic.

So ATSUI was removed because it is slower, and it will be reintroduced when
Mozilla eventually catches up using ATSUI, because then it will be... faster!?
Who is kidding who here? I heard Lynx is pretty darn fast!

Apple's philosophy with Mac OS X was to ship a slow system with full
functionality, and thereafter optimize for speed. That's a good strategy.

Chimera's philosphy, if any, seems to be to leave key functionality out until
"it gets faster". But that also means there is no need using Chimera.

Ah man, I thought this would be a killer browser.

Comment 16

[Greg K.]

Niklas, bug 157967 includes some insight from Simon on ATSUI and Gecko. You
might find it interesting, if not appeasing.

Comment 17

[Greg K.]

*** Bug 187942 has been marked as a duplicate of this bug. ***

Comment 18

[Chris Casciano]

how's this doing now that camino is on the trunk? To my eyes
http://www.helmug.gr looks pretty good at the moment.

(note: i got connection refused with the original URL)

Comment 19

[Simon Fraser [no longer active]]

fixed now we're on the trunk