Multiple keystroke foreign language characters not correctly selectable

[MeiYu]

I am trying to convert Indian language input into Mandarin and for some reason, the individual characters are selected incorrectly.

For instance, நன்றி should be 3 characters (the obvious separation is correct), but in FileMaker the 1st character is correct and the 2nd character is all except the last vertical line. The last vertical line, according to FileMaker is the 3rd character.

Looking at the Code() for each character is 2984, 2.993e+13 and 3007. While the 1st and 3rd are within normal ranges (still not right for the 3rd, but within normal ranges), the 2nd is insane.

I have tried to export the field, save it in .xls format and reimport it without any effect. Each of the characters is correct in Excel and Notepad.

Thanks for any enlightenment!

[comment]

For instance, நன்றி should be 3 characters (the obvious separation is correct)

I copied that and pasted it into a text editor, and it came out as 5 characters. Same thing in Filemaker. The characters are (from left to right):

2984 TAMIL LETTER NA

2985 TAMIL LETTER NNNA

3021 TAMIL SIGN VIRAMA

2993 TAMIL LETTER RRA

3007 TAMIL VOWEL SIGN I

I am not sure what "the obvious separation" is.

[MeiYu]

And my friends say I'M too specific!

The obvious refers to characters that aren't physically connected, or SEPARATED!

In Google Translate, Notepad or Excel, there are three physically selectable characters. Of everything I've tested this in, FileMaker is the only program that doesn't treat this as 3 characters (which would be perfect for what I'm trying to do).

[RodSierra]

This may be related to Filemakers incompatibility with other languages, such as Thai. In the case of Thai, Filemaker refuses to accept entry of certain common characters, only beeps a warning, however if the characters are pasted into the field they appear properly. This has been around since FM7 when unicode was introduced, and although I among others have requested corrective actions, this issue still persists today. I have come to the conclusion that Filemaker is not fully unicode compatible, and have pretty much given up on any hope of a fix. In cases where we must use Filemaker for text entry in these languages we have come up with a cluge where the text is entered into a web viewer popup and then pasted into the field.

[comment]

I don't speak or write Tamil. If I understand correctly, you are complaining about cursor behavior - not about any calculation results? If so, make sure the field is set to use Tamil as the default language (under Field Options > Storage). Selecting a dedicated Tamil font may also improve things. If all else fails, try these guys:

http://www.winsoft-international.com/en/products/database.html

[MeiYu]

The text is already input into FileMaker. Copying and pasting it into something else only returns the same result when pasted back into FileMaker.

I did notice that some of the Tamil characters are converted to a unicode string (2984, 29930302102985 and 3007); the middle sequence is three unicode characters which could be converted to their equivalents. The only one that is originally what would work for my current design is the first, but if I process the unicode character sequence in quintets, it could possibly work. A lot more work, but anything easy has already been done!

I had tried Tamil specific fonts already and Tamil is not listed under the language choices. The database and FileMaker app is already very well developed. Tamil is simply an addition that raised questions. No need to throw out the near-adult with the bath water.

BTW, I didn't have the same problem with Thai which was the basis for this effort. In over 28,000 records with the longest sequence in any single record being 76 Thai characters, the same problem never occurred.

I was just hoping that I had done something wrong. It seems the only time I'm wrong is when I think I am! :D

[comment]

As I said, I am not familiar with Tamil, so some of your points are cryptic to me.

I did notice that some of the Tamil characters are converted to a unicode string (2984, 29930302102985 and 3007);

Converted from what? The string in your original post is a Unicode string of 5 (five) characters. There is no such thing as character 29930302102985; these are three characters 2993, 3021, and 2985 (from right to left).

If it's possible to express the string 2993, 3021, 2985 as a single precomposed character, then you must either use Substitute() to convert the entry to that character, or enter that character to begin with - see also:

http://fmforums.com/forum/topic/64240-limitations-to-code-and-char/page__view__findpost__p__304655

However, I doubt this will resolve the issue of cursor behavior - if indeed that is the issue here. Probably not, since you say:

I process the unicode character sequence in quintets

Tamil is not listed under the language choices

Sorry about that; I inadvertently tested this in one of WinSoft products. Apparently Filemaker itself does not support Tamil.