Dealing with multiple encodings in the same app

[Isak]

Our (rails) app is mainly utf-8, so we set kcode to u.

However, we need to generate pdfs using the cp1252 encoding, as the
FPDF library doesn't have unicode support.
Translating strings work fine using iconv, but the library is using
sprintf internally, which blows up on non-utf8 characters.

E.g.:

s = ("" << 73 << 241 << 116 << 234 << 114 << 110 << 228 << 116 << 105
<< 111 << 110 << 97 << 108 << 105 << 122 << 230 << 116 << 105 << 248
<< 110) # valid cp1252-string
puts sprintf("aa %s bb", s)

...which discards the last two characters of s.


Suggestions appreciated,
Isak

 

[Manfred Stienstra]

Our (rails) app is mainly utf-8, so we set kcode to u.

From Rails 1.2 on KCODE is always set the 'u', so you might not have

to set it anymore.

However, we need to generate pdfs using the cp1252 encoding, as the
FPDF library doesn't have unicode support.

Ok, so FPDF doesn't support Unicode.

Translating strings work fine using iconv, but the library is using
sprintf internally, which blows up on non-utf8 characters.

But you can only feed it utf-8 characters?

 

[Isak]

Ok, so FPDF doesn't support Unicode.
But you can only feed it utf-8 characters?

That's the gist of it, at least when running ruby with kcode = 'u'.

I replaced the 4 occurances of sprintf("%X %s bbb", X, txt) with
sprintf("%X ", X) + txt + " bbb", which fixed our issues.
Reasonable approach, or is there a better way?


Isak