Percent encoding for web searches

[jennie]

This came up at the Greek Ubuntu forum, with a user trying to add this website search

Code: Select all

http://www.in.gr/dictionary/lookup.asp?Word=%GDWORD%

. The site uses ISO-8859-7 encoding, so it refused to understand the search. I tried using

Code: Select all

http://www.in.gr/dictionary/lookup.asp?Word=%GDISO7%

but this didn't work either. Is there another workaround for this? I was wondering if it would be doable to have the application convert keywords to percent-encoding (url encoding) before looking them up on websites.

[ikm]

Yes, only ISO1 was implemented. Now I've added support for all ISO encodings via %ISO1%, %ISO2% etc. Try the latest SVN version -- your site now works fine with ISO7.

p.s. Regarding percent encoding -- it should be done already, transparently. But percent encoding does not solve the initial encoding problem -- it is merely a way to encode 8-bit chars to their hex representation, and as such, it encodes strings which must already be in some 8-bit encoding. So you always have to choose the right text encoding anyway.

[jennie]

Thanks for the reply. I will try the new version.
I 'll also add a few really minor things here:
Although I have Ancient Greek fonts installed, ancient Greek accented characters are rendered as blank squares (e.g. ἁ, ῖ). Modern Greek displays correctly.

I have a stardict dictionary that I compiled myself, with cross-references to keywords of the type "Ρ2.1". Clicking on the link will not work. If I choose "Lookup" however, I am shown the Ρ2.1 entry. Links to "Ρ2" work as expected, so the problem seems to be the dot (.). I have already replaced the faulty links with dotless ones, but I thought I should mention this.

[ikm]

Please upload the dictionary somewhere (if you don't mind of course), I'll see and fix those problems with dot.

[jennie]

Hi! Did you get the email I sent you (to the address found in the Contact page)?

[ikm]

Yes. I'll get to it when I have time.

[ikm]

See if the issue was fixed in SVN r369.

Note that I have only found one dotted link in your dictionaries, namely "Ρ2.1α", and it still doesn't work since there's no such article in neither of the two dictionaries. But now it is decoded correctly.

[jennie]

Sorry for the late reply. I can confirm that this issue has been fixed. Thanx a lot!!