Icon View Thread

The following is the text of the current message along with any replies.
Messages 1 to 6 of 6 total
Thread =?iso-8859-1?B?7yBhbmQg6Q==?=
Thu, Apr 17 2014 2:59 PMPermanent Link

Ronald

Hi,

I have some problems with Unicode I presume. My EWB app makes a call to my
cgi app on the server that reads an rss feed and converts that to a .ini
file format (for easier reading). If I make this call in my browser I get
the raw .ini content and I can see that the word Oekra´ne (dutch for Ukrain)
is spelled correct. I can see the double dots above the i. Id I read this
content from my EWB app I get rectangles where the ´ should be. Why is that?
I have tried all kind of conversions to utf-8, but I do not get the right
characters in my app. What part of unicode don't I understand?

Thanks,
Ronald
Tue, Apr 22 2014 4:52 AMPermanent Link

Matthew Jones

I've had this sort of issue, and the main cause has been the web server's encoding
string. On the RemObjects SDK server, the Indy code puts out the wrong encoding,
but when I fix it it works properly in data to EWB. Obviously you have to handle it
properly all through the chain, and I still have issues with that in my Delphi code
somewhere, but in general with that fix I can round-trip Unicode via UTF-8 and
DBISAM databases. So checking the encoding is the first thing to do.

/Matthew Jones/
Tue, Apr 22 2014 2:42 PMPermanent Link

Ronald

Thanks, I will have another look. I already tried to explicitely convert it
to UTF8String, but I will try some more. I will report on my progress.


"Matthew Jones"  schreef in bericht
news:memo.20140422095252.7924H@nothanks.nothanks.co.uk...

I've had this sort of issue, and the main cause has been the web server's
encoding
string. On the RemObjects SDK server, the Indy code puts out the wrong
encoding,
but when I fix it it works properly in data to EWB. Obviously you have to
handle it
properly all through the chain, and I still have issues with that in my
Delphi code
somewhere, but in general with that fix I can round-trip Unicode via UTF-8
and
DBISAM databases. So checking the encoding is the first thing to do.

/Matthew Jones/
Wed, Apr 23 2014 4:59 AMPermanent Link

Matthew Jones

> I already tried to explicitely convert it
> to UTF8String

And you probably did, but if the web browser and server don't tell each other it is
UTF-8, they will convert it according to what is actually set. Chrome's F12 debug
system is good for this as you can see the headers being used.

/Matthew Jones/
Thu, Apr 24 2014 4:51 PMPermanent Link

Ronald

If I set the charset to UTF-8 then I get the "diamaonds" with the
quetionmark.
If I set it to iso-8859-1 then it shows the diacritic chars correct. I
thought that utf-8 was a sort of  "general purpose" set that could be used
to show all most-used characters.

"Matthew Jones"  schreef in bericht
news:memo.20140423095911.7924I@nothanks.nothanks.co.uk...

> I already tried to explicitely convert it
> to UTF8String

And you probably did, but if the web browser and server don't tell each
other it is
UTF-8, they will convert it according to what is actually set. Chrome's F12
debug
system is good for this as you can see the headers being used.

/Matthew Jones/
Fri, Apr 25 2014 4:19 AMPermanent Link

Matthew Jones

> If I set the charset to UTF-8

Where? It is important for all aspects of the transmission to send with the right
character set information. http uses a content-type or similar. I found I had to be
careful on my server code for some things too. It is worth tracing all the way
through.

/Matthew Jones/
Image