Page 1 of 1
Writing UTF/8 from a field?
Posted: Wed May 23, 2018 12:10 pm
by thatkeith
I have a quick-n-dirty tool for generating HTML and posting it, so I can keep a schedule in a web page up to date without mucking about writing HTML directly. In LC I use tabbed text in a field, and I parse/convert this content to a table structure (wrapped in the rest of a basic HTML page), and upload that by FTP. This file is shown in an iFrame in a WordPress page.
This all works absolutely fine – but I can't get my head around how to support characters such as ö and Japanese characters. If I generate the page manually as UTF/8 from BBEdit it works fine, but when I update the remote page using my LC stack like so:
Code: Select all
put field "code" into URL "ftp://NAME:PASSWORD@FULL/FTP/PATH/TO/PAGE.HTML"
the upload works fine, as does the iFrame, but the encoding fails; in a browser I see diamond blocks for the 'special' Western chars and strings of question marks for the Japanese content.
The HTML page code I'm writing starts with this:
Code: Select all
<html>
<head>
<meta charset="utf-8"/>
but this doesn't help with the LC-generated output. Any tips on how should I handle writing out text in a form that doesn't trip up over this kind of content?
k
Re: Writing UTF/8 from a field?
Posted: Wed May 23, 2018 1:03 pm
by MaxV
Just use the
htmlText property:
Code: Select all
put the htmlText of field "code" into URL "ftp://NAME:PASSWORD@FULL/FTP/PATH/TO/PAGE.HTML"
Re: Writing UTF/8 from a field?
Posted: Wed May 23, 2018 4:34 pm
by FourthWorld
HtmlText will escape the angle brackets, so while it's sometimes useful for converting non-HTML to a form of HTML (which usually needs further massaging to turn into web-ready HTML), if you're starting with HTML it'll just produce a form that browsers will render with all the tags showing, rather than parsing the tags.
To convert text to an encoding other than LiveCode's internal format, use the textEncode function.
Also, FTP sends passwords in plain text, considered too dangerous to use outside one's internal network, given the hostile nature of the open Internet these days.
With all LiveCode editions you could use a CGI script on the server to receive data posted to it via HTTPS.
With the Business and Indy editions you can also use the tsNet external to send via ftps or sftp.
Re: Writing UTF/8 from a field?
Posted: Wed May 23, 2018 4:49 pm
by jacque
Just to add : LC uses UTF 16, and without textEncode, that's what the browser will get. The tipoff is all the missing characters and the weird symbols. LC recommends using textEncode any time you move data in or out of LC, unless you are sure the external app works with UTF 16.
Re: Writing UTF/8 from a field?
Posted: Thu May 24, 2018 3:10 pm
by MaxV
FourthWorld wrote: ↑Wed May 23, 2018 4:34 pm
HtmlText will escape the angle brackets, so while it's sometimes useful for converting non-HTML to a form of HTML (which usually needs further massaging to turn into web-ready HTML), if you're starting with HTML it'll just produce a form that browsers will render with all the tags showing, rather than parsing the tags.
...
I don't think so, the
htmlText is what he needs to avoid chars problems.
For example "å" is HTML correct just with "å" other solutions are not really reliable. And "å" is the output of htmlText.
Regarding FTP, it depends on the service provider. Here in Italy all service providers use just FTP, and nobody still has problems with it.
Re: Writing UTF/8 from a field?
Posted: Thu May 24, 2018 3:36 pm
by FourthWorld
MaxV wrote: ↑Thu May 24, 2018 3:10 pm
FourthWorld wrote: ↑Wed May 23, 2018 4:34 pm
HtmlText will escape the angle brackets, so while it's sometimes useful for converting non-HTML to a form of HTML (which usually needs further massaging to turn into web-ready HTML), if you're starting with HTML it'll just produce a form that browsers will render with all the tags showing, rather than parsing the tags.
...
I don't think so, the
htmlText is what he needs to avoid chars problems.
For example "å" is HTML correct just with "å" other solutions are not really reliable. And "å" is the output of htmlText.
Go ahead and try using it to produce HTML when the field's contents are already HTML and let us know how that works out.
Regarding FTP, it depends on the service provider. Here in Italy all service providers use just FTP, and nobody still has problems with it.
The security risk with FTP is inherent in the protocol, independent of any ISP using it.
If you leave your house unlocked and have never had a burglary, that's not because leaving your door unlocked is a good security practice but that you've merely been lucky.
Re: Writing UTF/8 from a field?
Posted: Wed May 30, 2018 11:03 am
by thatkeith
FourthWorld wrote: ↑Wed May 23, 2018 4:34 pm
FTP sends passwords in plain text, considered too dangerous to use outside one's internal network, given the hostile nature of the open Internet these days.
Good advice, exactly what should always be said when old-school FTP is invoked in LC. Fortunately, for this specific case it's a tool made just for a colleague and me, and it'll be redundant (and the password changed just in case) after the weekend.
(Which I realise doesn't address the issue of plain text directly, but it's at least not something I'm proposing to release or even use more than half a dozen times.)
