Naomi et al, many thanks for your helpful replies. It seems we were tripped up the recent VFP (7/8?) change to the way code pages are processed. By changing the PCs Windows regional language setting from Chinese to English, things now work. However, we sacrifice the Chinese characters stored in some fields. Mostly it is in English but some fields (in this database, common names for species and place names) are also stored in Chinese.

So - to get things running on all 4 cylinders with Chinese and English characters, am I right to say we need to a) change the Windows setting back to Chinese b) change the code pages for all the DBF files in this project to the Chinese one and c) add a CODEPAGE= statement to config.fpw.... and a 'SET COLLATE TO ' if we want indexing correct too ... ? Many thanks, Denis




Put MASHINE instead of AUTO. Also if you checked the thread I pointed you to you may conclude, that you need to make sure that forms use the same SET COLLATE TO MACHINE in their DE. Finally, you may want to recreate all your existing indexes making sure collate is set to machine.

>Dragan, thanks for reply. I've tried most things but we still get this message 'Collating sequence PINYIN not found' when reindexing a particular file. That file as distributed from here has a code page value 1252 as determined using the cpdbf() function. Their config.fpw includes the line CODEPAGE=AUTO.
>
>The enduser in China has Windows XP (Service Pack 2). I don't know much about the .nls you mention files. Is there an easy way to check that? Im not sure what to next check. We deleted all CDX files too.
>
>Seems to me that documentation of this whole area could be unified bring together under the 'internation applications' section a clearer list of things to check config.fpw, set collate, codepage, .nls as you mention and so on. Many thanks again, Denis
>
>
>
>>>Borislav
>>>
>>>>What is the very first error your users see?
>>>
>>>Thanks for your help here. it seems the first error message seen is the
>>>'Collating sequence PINYIN not found' one. I am trying to trace what might have changed since the last revision of the same software - as this did work! But so far, nothing obvious, Denis
>>
>>Changed Config.fpw (CODEPAGE=?) ?
>>If the current codepage not support the desired collating this error happens.
>>New indexes in Tables?
>>New tables?
>>Changed SET COLLATE TO?
>
>Or if there's an existing index in that collating sequence, but the table's codepage is not proper for that sequence. You can't index a table which has codepage=1251 in a Swedish sequence - but you once could, up to VFP6 or 7.
>
>Now if the tables were rebuilt from GenDbc generated code, and the current codepage at that moment wasn't proper for Pinyin script... ahem, how did it index the tables then? Or maybe there's a separate indexing utility written in an older version of Fox?
>
>Also, the installed codepage on the given machine needs to be present (the .nls files, I think). Up to v7 Fox was using its own foxpro.int, but since (I think) 8 it uses system level nls files, and if the proper nls file is missing, that could also explain this.