ASCII transliteration problems (spanish)

[efejota]

I have been looking around and still cannot fix the issue. I saw there was a ticket (#4782) but I had no help from it. The question is: I have a HTML doc, spanish text, perfectly well seen in Firefox using ISO-8859-1, 8859-3 or 8859-15 or cp1252 (what was supposed to fix the ticket previously mentioned). I set the input in any of those codifications and I cannot get it working properly. When formatting it into another format (btw, I can`t get these docs to generate a proper epub file, conversion always hangs at that point, so I use lrf file, instead), all I get is no "ñ", no "ó", but ns and os and so on. If I force to use iso or cp as mentioned before in the output section, I have strange codes instead. I upload a file just for anybody here to try, if wanted.

I have tried the debug thing and it does not transliterale properly in the input section, so the problem seems to me that it it further up, in the process of converting html to zip.

Any help or hint?

Thanks

[kovidgoyal]

Read point number 2 in http://calibre-ebook.com/user_manual...r-smart-quotes

[efejota]

Thanks for your answer. I have had already done what it is said in that point, out of the embedding characters, as I have had no problem with spanish characters at all before. So, to be clear, I changed the input codification to match the one in which the document showed perfect in Firefox (let's say, ISO 8859-1) and tried with the other codifications that worked in firefox too. The results where always more or less the same. If I left blank the output look and feel in the conversion window, I had no special characters. If I put any other codification there, the resulting file had strange codes, varying depending of codifications. BTW, I have latest version of Calibre for MacOSX. Would it be helpful for anybody if I upload the original file and the resulting converted files?

FJ

[Starson17]

Originally Posted by efejota

Thanks for your answer. I have had already done what it is said in that point,

If you did it correctly, the following would have happened:

• You would have started with an html file.
• You would have specified the correct encoding in Calibre for importing the html file as a zip into Calibre.
• Calibre would have converted the input encoding of the html to utf-8 and packed the file into a zip.

Assuming that you got that far, viewing the zip file in Calibre's viewer would have shown the correct characters. You would not have ever entered anything into the look and feel of the conversion box, as you would not yet have done a conversion (although Calibre would have done one during the importing process - Calibre always stores html/zips as utf-8). It sounds like you tried the above, but never saw correct characters.

To me that sounds like you specified the wrong input encoding or there is a bug. To check, you could try to see what binary code is in your orignal file for a "problem" character (use a hex editor). Then see if that binary code specifies the character you think it should specify in the encoding you have advised Calibre is used for that file (find a table for that encoding). If it matches, there's a bug in Calibre's utf-8 conversion. If not, then you've got the wrong encoding specified and you need to find the right one.