Created attachment 22551 [details]
Set of patches from OpenOfficeTLE 1.1.2 (against 1.1.3)

Got it. Will write an issue for each one that has not been submitted.

Item 10 notes.

OOo 1.1 hard-codes a default font size of 10 points or 12 points in a number of
places. For most Thai fonts 10 points or 12 points is unreadably small.  Thai
characters have four vertical levels (base, 1 diacritic below and two diacritics
above).  When Thai characters are mixed with Latin characters, the relative
sizes of Thai and Latin characters are usually chosen so that the bottom of
below diacritics is lower than the bottom of Latin descenders, and the top of
the second level Thai diacritics is higher than the top of Latin ascenders. The
(not unreasonable) normal convention for Thai fonts is that in an N point font
the distance from the bottom of the below-diacritic to the top of the second
above-diacritic is N points.  Thai typesetting would typically use a 16 point
default font size.  Thai fonts typically contain Latin characters as well: the
Latin characters in a 16 point Thai font would be about the same size as the
characters in a normal 12-point Western font.  

Some newer fonts don't follow this convention.  Instead they size of a font is
based on the size of the Latin characters.  In this case, the font will either
(a) have a much larger line-spacing than usual or (b) will have to use a
unusually vertically compact design for the Thai diacritics. Loma does (a) I
believe. Tahoma does (b). (a) is not good if you use the font for English alone.
(b) is not good because the characters look ugly. A better solution would to use
an OpenType BASE table to specify a different line-spacing for different
scripts.  I don't know of any currently available font that does this, but I
expect to see them in the future.  In particular work is ongoing on adding Thai
characters to Bitstream Vera using this approach.

OfficeTLE's solution to this is just to hard-code 16 points instead of 12 points
as the default font size.  OfficeTLE also changes the default fonts for Western
use to be Thai fonts (which contain Latin characters for which 16 points is a
reasonable default point size).  This is not so good even for Thai users,
because the Latin characters in Thai fonts are usually not very good quality.

I'm not exactly sure what the right solution is. I think part of it is probably
to allow VCL.xcu to optionally specify a font size for each default font (eg

<value>Norasi:16;Bistream Vera Sans:12;Tahoma:12</value>

Item 21. On Windows with m74 with Thai system locale, in Calc in the Format 
Cells dialog, if I set Category to Currency, Language to Thai, and then click 
the down arrow on Format, and scroll through the list of currencies, I see 
many entries for Indic and Middle eastern (bidi) scripts with messed up 
formatting.

James, I get the same behavior even when the locale is set to English.

Item 10. On Windows WordPad defaults to Arial 10pt in a Western locale (e.g. 
Control Panel|Regional and Language Options|Regional Options set to English 
US), but Cordia 14pt in a Thai locale. In a Western locale, if I change the 
keyboard layout to Thai, the font changes to Cordia 14pt.

Item 32. With Notepad on Windows XP, the cursor shape changes from a vertical 
bar to a sort of very narrow backwards L shape whenever the keyboard layout 
changes from English to Thai.

Item 33. In Windows, the Control Panel under Supplemental Language Support 
uses the terms "East Asian languages" and "complex script and right-to-left 
languages (including Thai)", which is a lot better than OOo's "Asian" 
and "CTL".

Item 34. On Windows XP, WordPad has a drop down box for the current script, 
next to the font name and font size drop down boxes; when the keyboard layout 
is changed, the script in the drop down box changes automatically.  The list 
of scripts show in the box corresponds to the scripts supported by the font 
shown in the drop down box.  If the current keyboard layout is English and 
current font is Arial (or any other font that doesn't support Thai) and you 
then change the keyboard layout to Thai, the font shown in the drop-down box 
will switch to Cordia.

Item 41. The OpenOffice TLE 1.1.2 extras (25M) are at:

http://www.softwarebank.org/download.php/49/OOTLE_1.1.2_extras.tar.gz

Does this mean we are going to look at each item in the list;
see if it's valid or not, for the current OOo 1.9.x build; and if it is, create
new issue in the IssueZilla (to bring it to attention of wider OOo developers)?

James->Art:  Yes, that's what we're in the process of doing.  SIPA has
contracted Samphan to do the work.  He's done quite a few already as you can see
from the comments and Issue links.  The descriptions are intended to be just
enough to remind those of us who have been discussing the problems of the
various issues we've talked about.  In many cases, I fear they may be too short
to make sense to those who haven't bee involved in the discussions.

But if any of the descriptions here make sense and you can help turn them into
proper issues, that would be great.  Falko (ft) is helping us coordinate with
the OpenOffice team, so we're assigning new bugs to him.  Also we have a
tracking bug issue #41707 to which we're adding all our issues as we create them.

The plan is for SIPA to provide some significant funding to Sun in Hamburg
(and/or anybody else who can help get the job done) to get at least some of
these things fixed in OOo 2.x, so that eventually there will be no need to
maintain two separate Thai versions (OfficeTLE and Pladao).  I am sure you agree
that the best long-term solution is for the official version of OOo to have
first-rate Thai support.

Many of the issues on the list come from the OfficeTLE team at NECTEC. I would
like to ensure that we also include any issues that the Pladao team has
discovered; I have tried to make contact, but so far without success.  I haven't
even been able to find the source code for latest version of Pladao: everything
in their CVS Web seems to be about two years old.  I think SIPA would be happy
to provide some funding to the Pladao team to get their changes into a suitable
form for merging into the OOo 2.x.  Perhaps you could help liaise.

Pladao Office 3.1 source code

http://pladao.org/modules.php?name=Downloads&d_op=getit&lid=93

an actual url for the Pladao Office 3.1 source code

http://www.pladao.org/files/3.1/PladaoOffice3.1-src.tar.gz

Pladao Office 3.1 readme -- for known issues with Pladao Office 3.1
http://www.pladao.org/files/3.1/readme.html

(in Thai language)

----

FYI, Pladao Office 3.1 is OO.o 1.x-based.

Revise the list :-

0. Tracking bug for Thai-related issue Issue #41707
1. Numbers inappropriately formatted using Thai digits Issue #41672
2. Thai line-breaking non-functional Issue #41671
3. Display of invalid Thai combining character sequences broken on Windows Issue
#42171
4. Sequence input checking not working Issue #42469
5. Thai justification inaccuracy Issue #41860
6. Need input sequence correction feature Issue #42661
7. Interop of Thai text with .doc broken for Word >= 2000; maybe Issue #23784
8. Need number spellout feature Issue #3702, Issue #17554
9. Interop with Thai MS Excel 97: BAHTTEXT(), "t" format prefix Issue #42727
10. Font size for default fonts needs to be configurable for each font Issue #42725
11. Display of invalid combining sequences broken with some Unicode fonts (e.g
Tahoma) Issue #42663
12. Missing dotted circle breaks display of invalid combining sequences Issue #42662
13. Thai locale data incomplete Issue #42723
14. Windows search cannot search inside .sxw (fixed in 2.0) Issue #21108
15. Optional bell/flash when invalid input sequence entered Issue #42728
16. psprint cache makes printing become gradually slower and slower Issue #20287?
17. On OOo Web site, link to Thai OpenOffice goes to Pladao
18. Mail merge should be able to generate one file containing all recipients
instead of one for each (maybe fixed in OOo 2.0) Issue #20057
19. Inaccurate cursor positioning for Thai (fixed in OOo 2)
20. Make Pladao and OfficeTLE fonts conveniently available for OOo Issue 42738
21. In non-English locale, numbering menu has bogus entries (get Sila to explain
how to reproduce)
22. ICU Thai word-breaking does not work well in real-life
23. Integrate Thai spell-checking Issue #31230
24. Problem with mixed-language spell-checking that shows up when Thai
spell-checking integrated (see patch in Issue #31230 from hin)
25. Thai thesaurus
26. Sequence checking at beginning of paragraph or in between English and Thai
(need 4 fixed first)
27. On Windows machine with Thai locale, installation needs to be set up
appropriately for Thai (e.g. with CTL enabled with default language of Thai)
(fixed already in OOo 2?) Issue #42730
28. Need feature for easy manual override of incorrect word breaking Issue #42660
29. Thai in title bar and filename (fixed in OOo 2)
30. Showing non-printing characters with font with no middle dot on Windows,
characters following middle dot get font that contains middle dot
31. Should have separate font for each script as in Mozilla, instead of
Western/Asian/CTL Issue #42735
32. Should have visual feedback for current keyboard layout (e.g. change in
cursor shape) Issue #42731
33. UI should not use term "CTL" Issue #42733
34. Font dropdown should be aware of current keyboard layout Issue #42732
35. Thai language pack (get NECTEC to do it?)
36. Need to integrate word-breaking and spell-checking
37. Inconsistent treatment of description property between Word and OOo creates
interop problem (from NECTEC)
38. Thai forward word broken (ICU bug; fixed in OOo 2)
39. Need to be able to switch UI language at runtime (fixed in OOo 2)
40. Installation of OOo with Thai UI should install English UI as well (fixed in
OOo 2?)
41. Make OfficeTLE non-code extras (e.g. clip art?) available for OOo
42. Problem fixed by ootle112_mapchar_en_th.patch
43. Ignore obsolete chars when numbering using Thai letters
44. Extend number of outline numbering levels from 8 to 16
45. Problem fixed by ootle112_adjust_matric_page_margin.patch

Another one we discussed that I forgot to add to the list:

46. Don't understand semantics of check-box to control whether Sequence Input
Checking is or is not Restricted

I think by far the most important item that we don't yet have a good issue for
is 7. Next is 42 (this looks like it is very important as soon as you have a
Thai UI).

Current state

Issued:-
0. Tracking bug for Thai-related issue Issue #41707
1. Numbers inappropriately formatted using Thai digits Issue #41672
2. Thai line-breaking non-functional Issue #41671
3. Display of invalid Thai combining character sequences broken on Windows Issue
#42171
4. Sequence input checking not working Issue #42469
5. Thai justification inaccuracy Issue #41860
6. Need input sequence correction feature Issue #42661
7. Interop of Thai text with .doc broken for Word >= 2000; maybe Issue #23784
8. Need number spellout feature (old Issue #3702, Issue #17554) Issue 43043 
9. Interop with Thai MS Excel 97: BAHTTEXT(), "t" format prefix Issue #42727
10. Font size for default fonts needs to be configurable for each font Issue #42725
11. Display of invalid combining sequences broken with some Unicode fonts (e.g
Tahoma) Issue #42663
12. Missing dotted circle breaks display of invalid combining sequences Issue #42662
13. Thai locale data incomplete Issue #42723
15. Optional bell/flash when invalid input sequence entered Issue #42728
20. Make Pladao and OfficeTLE fonts conveniently available for OOo Issue 42738
27. On Windows machine with Thai locale, installation needs to be set up
appropriately for Thai (e.g. with CTL enabled with default language of Thai)
Issue #42730
28. Need feature for easy manual override of incorrect word breaking Issue #42660
31. Should have separate font for each script as in Mozilla, instead of
Western/Asian/CTL Issue #42735
32. Should have visual feedback for current keyboard layout (e.g. change in
cursor shape) Issue #42731
33. UI should not use term "CTL" Issue #42733
34. Font dropdown should be aware of current keyboard layout Issue #42732
41. Make OfficeTLE non-code extras (e.g. clip art?) available for OOo Issue
42971, Issue 42738
42. Problem fixed by ootle112_mapchar_en_th.patch Issue 42964
43. Ignore obsolete chars when numbering using Thai letters Issue 43045
44. Extend number of outline numbering levels from 8 to 16 Issue 42965
45 Don't understand semantics of check-box to control whether Sequence Input
Checking is or is not Restricted Issue 42967

Fixed:-
14. Windows search cannot search inside .sxw (fixed in 2.0) Issue #21108
16. psprint cache makes printing become gradually slower and slower (fixed in
OOo 1.1.1?) Issue #20287 
18. Mail merge should be able to generate one file containing all recipients
instead of one for each (maybe fixed in OOo 2.0) Issue #20057
19. Inaccurate cursor positioning for Thai (fixed in OOo 2)
29. Thai in title bar and filename (fixed in OOo 2)
38. Thai forward word broken (ICU bug; fixed in OOo 2)
39. Need to be able to switch UI language at runtime (fixed in OOo 2)
40. Installation of OOo with Thai UI should install English UI as well (fixed in
OOo 2)
	Users should install English OOo then Thai language pack instead?

New feature:-
23. Integrate Thai spell-checking Issue #31230
25. Thai thesaurus
35. Thai language pack (NECTEC will do it)

No idea:
17. On OOo Web site, link to Thai OpenOffice goes to Pladao

Need other fix
26. Sequence checking at beginning of paragraph or in between English and Thai
(need 4 fixed first)

Can't reproduce
30. Showing non-printing characters with font with no middle dot on Windows,
characters following middle dot get font that contains middle dot

To be issued :-

General, non-thai specific
37. Inconsistent treatment of description property between Word and OOo creates
interop problem (from NECTEC)

Need to build to understand:
21. In non-English locale, numbering menu has bogus entries
24. Problem with mixed-language spell-checking that shows up when Thai
spell-checking integrated (see patch in Issue #31230 from hin)

Need investigation:
45. Problem fixed by ootle112_adjust_matric_page_margin.patch
36. Need to integrate word-breaking and spell-checking

Issue at ICU
22. ICU Thai word-breaking does not work well in real-life

> 17. On OOo Web site, link to Thai OpenOffice goes to Pladao

On which page please? If you can give me a url, I may ask people who take care
the website to fix it. .. Replace it with http://th.openoffice.org/ ?

Item 17.

If you go to:

http://download.openoffice.org/1.1.4/index.html

Then select Thai.  On the page this takes you to, select "Continue to download".
 Then it takes you to a page that says:

คุณสามารถดาวน์โหลด OpenOffice.org รุ่นภาษาไทย ได้จากไซต์เหล่านี้:

    * http://pladao.org (ประเทศไทย)

We need to make sure this it doesn't say this for the 2.x release.

I think it would be helpful to provide some guidance on the relative priority of
issues. Fixes for issue 41671 and issue 41672 are in the pipeline.  Apart from
those, my top issues (from highest to lowest priority) would, I think, be:

issue 41860
issue 42909
issue 42469
issue 42725

Anybody else have thoughts on this?

Item 17. URL is http://th.openoffice.org/about-downloads.html

Item 17, fixed. 
Now points to both Thai-hacked versions (Pladao Office, OfficeTLE) and also
standard version (OpenOffice.org).

http://th.openoffice.org/about-downloads.html

45. Problem fixed by ootle112_adjust_matric_page_margin.patch
The patch came from the fact that Thai use Metric mesurement system as specified
in th_TH.xml but usually (always?) use Inch in measuring paper document (like
1.25" margin, 0.5" tab stop).

The problem is that the codes that the patch fixed have hard-coded some numbers
depend on the locale's mesurement system like 2 cm vs 1" (e.g. margin), 1.25 cm
vs 1/2" (tab stop).
The solutions :-
1) Specify in th_TH.xml that Thai use US (not metric) measurement system because
Thai use inch for papers. No change in code.
2) Have another entry in the locale data to specify 'measurement system for
document' that will be reflected in default values for various things and the
unit of the rulers.
3) Not hard-code anything. Read every defaults from locale data.

45.  I like your solution 2.

45. Just discussed with people on TLWG newsgroup and talked with Sila, concluded
that there's no standard for measure system for Thai document. Some people say
CM (and left/right margin of 2cm). Some say inch (and left/right margin of 1".
One say that it came from the use of MS Word in the past.). Sila also suggest
not to submit this patch. So I think we should left it as CM to conform with the
national measurement system.

17. On OOo Web site, link to Thai OpenOffice goes to Pladao - Fixed
37. Inconsistent treatment of description property between Word and OOo creates
interop problem (from NECTEC) Issue #43588
45. Problem fixed by ootle112_adjust_matric_page_margin.patch - Non-issue
36. Need to integrate word-breaking and spell-checking - Issue #43583

21. In non-English UI, bullet & numbering menu has bogus entries - FIXED
   Tested by building OOo 680_m79 with Thai UI, then look at
   Format->Bullet & Numbering->Options->Numbering (detail from Sila), 
   no strange or blank entry. However, there're some errors in translation
   which could be fixed by NECTEC in the new Thai language pack.

K Nusorn pointed me to this

http://marketing.openoffice.org/conference/presentations-pdf/fri1330/PladaoOffice2003.pdf

This is a nice discussion of everything Pladao added.

I noticed references to a couple of things that we haven't discussed yet:

- drop caps

- underlining

Created attachment 23108 [details]
An example of Thai drop cap text

We can specify the number of characters to drop in the drop-cap feature. If we
specified only part of the first cluster (e.g. 1 char when the cluster contain 3
chars), there'll be a bug (repetition of the first cluster, as shown in the
first paragraph in my attached example). But if we specified the number of
characters of the whole first cluster, the feature work (2nd paragraph).
Drop-cap of the first word also works (3rd paragraph).

I think I should file a bug on this (repetition of the first cluster).
Do you think it should detect the cluster boundary itself? (specify 1-3
characters should work in my example).

I think it should count in clusters.  "1" "character" should mean 1 cluster. 
It's no good to require the user to specify 3 characters for a cluster with
consonant+vowel+tone-mark, because you want to be able to have a paragraph style
which always drops exactly the first cluster, regardless of whether the cluster
has 1, 2 or 3 characters. In any case, it makes no sense to make a base
character dropped without also making its following combining characters dropped.

for unlining please see more info at Mozilla's Bugzilla:

Bug 156881 : Underline should skip character/part-of-character that is below the
base line (text-decoration-mode)
https://bugzilla.mozilla.org/show_bug.cgi?id=156881

Underlining, also this one,

issue 10708 : RFE: implement CSS3 text mode - text decoration
http://www.openoffice.org/issues/show_bug.cgi?id=10708

All valid items have their own issue. See issue 41707 - Tracking bug for
Thai-related issue. Discussion will be there.
22 - work-breaking will be issued at ICU.
Closed.

The Issue you raised has been marked as 'Resolved' and not updated within the
last 1 year+. I am therefore setting this issue to 'Verified' as the first step
towards Closing it. If you feel this is incorrect, please re-open the issue and
add any comments.

Many thanks,
Andrew
 
Cleaning-up and Closing old Issues
~ The Grand Bug Squash, pre v3 ~
http://marketing.openoffice.org/3.0/announcementbeta.html

As per previous posting: Verified -> Closed.
A Closed Issue is a Happy Issue (TM).

Regards,
Andrew