DL->HDU: Could you please takeover?

What kind of Linux do you use ? What input method is being used ?

OS: Redhat 8.0, input method for XFree86 below was used.

Input method for XFree86 & GTK2 from:
http://groups.yahoo.com/group/tamilinix/files/keyboard/
XFree86: tamil_kmap_xfree86.tgz  
GTK2: gtk2im-tamil99.tgz  

You can also get it from http://www.tamillinux.org -> Download ->
Keyboard Drivers -> tamil_kmap_xfree86.
(http://www.tamillinux.org/modules.php?op=modload&name=Downloads&file=index&req=getit&lid=5)

Tamil Unicode fonts:
code2000.ttf, latha.ttf (from WinXP) or GPL'd fonts from
http://groups.yahoo.com/group/tamilinix/files/unicode/fonts/tamil_opentype_fonts.tgz
 

How the keyboard layout is, is explained inside the tar-files.

What about the localization stuff, when will it be available?

Thanks
Manmathan

no idea about localizations

I have now tried to do some more testing and found out that Tamil
Unicode input method works fine (with some minor bad behaviors).

If tried to open a HTML containing tamil unicode text. in oowriter and
I found that it displayed it very well. Then I changed to "View ->
Print Layout" mode and some glyphs got wrongly displayed.

So I have now concluded that thing are only wrong when you are in
"Print-layout mode". The problem is not in input method but in
displaying correctly when in "Print-alike-view".

Try typing the keys "kOjh" while in Tamil unicode X-input mode and you
can see how this looks in different view modes.

The bad behaviors i talked about above is when you start a word by
typing "o", "p", "i" "h" og any other "adjoining-vowels". With meny of
the GPL'd fonts you get a lots of spacing-symbols.. I would prefer to
get nothing at all or just a single space.

The first time I tested for some time ago I seemed to get many thing
wrong, but it doesn't seem to be anyway.

I hope this can help in solving the issue.

Thanks
Manmathan

Since my printer is not working right now, I tried to print to a
postscript file and could se the same errors in the ps file..

Please try the tamil unicode sample page at:
http://www.murasu.com/unicode/sample.html

Isn't it possible to fix this issue before 1.1 gets rolled out?

Hi Philip!

Only the 3 "Two-part dependent vowel signs" (0x0BCA, 0x0BCA, and
0x0BCC) as mentioned in page 3 here
http://www.unicode.org/charts/PDF/U0B80.pdf cause rendering (correct
spacing?) problems.

Excerpt from the PDF: "These two-part dependent vowel signs have glyph
pieces which stand on both sides of the consonant. These vowel signs
follow the consonant in logical order, and should be handled as a unit
for most processing."

If you need more information please ask.

I hope this will be a quick fix.

Thank you
Manmathan

http://ooo.ximian.com/lxr/source/gsl/psprint/source/fontsubset/sft.h#567
says:
Maps a Unicode (UCS-2) string to a glyph array. Returns the number of
glyphs in the array which for TrueType fonts is always the same as the
number of input characters.

I am not sure if this will be true for the Tamil Truetype (opentype?)
fonts such as latha.ttf from mircrosoft, because of the "Two-part
dependent vowel signs". But how can that be that rendering in
HTML-mode or non-Onlinemode works fine?

In Online-View the critical step of adjusting the display text to the
printer is not necessary, so the problem seems to be caused by a bug
in this step. Will investigate...

pl->hdu: since you volunteered and really are the expert on this i'll
send this to you.

Dear Herbert,

I am very much interested in fixing this bug and would very much like
to help in doing so. Can you point to me where to look for the error.
Possible files/classes/methods to look into would help me a lot. 
Is my comment above (From manmathan 2003-06-08 09:19 PDT ) irrelevant? 

Or should I just keep my hands tight and wait for you to fix it and
stop spamming you?

Rgds,
Manmathan

Targetd corrected.

Won't be able to fix this for 1.1 -> 2.0.

> I am very much interested in fixing this bug and would very much like
> to help in doing so. Can you point to me where to look for the error.

The locatation for printer adjustment on the UNX platform is in
vcl/source/gdi/sallayout.cxx in the GenericSalLayout::AdjustLayout()
and especially GenericSalLayout::ApplyDXArray. If "glyph fallback" is
involved then the MultiSalLayout::AdjustLayout() in the same file is
important. When MultiSalLayout::AdjustLayout() is called this is a
sign that "glyph fallback" is active. Just knowing if this is the case
would help.

In this layer the script+language dependent processing is already
done, so a change there will affect everybody, please be very careful.
Starting next week when I get back after my hospital stay + recovery I
will be able to look closer into this.

Moving issue from l10 to gsl.

.

According to the OpenOffice.org roadmap
(http://tools.openoffice.org/releases) this issue was retargeted to
"OOo Later".

The malayalam Rendering bug 15883, Bangali issue 11061 and this bug are the same.
This bug will effect all the Indian Languages.........

*** Issue 10476 has been marked as a duplicate of this issue. ***

Created attachment 21512 [details]
a sample document (using the Latha font)

.