Increase search and replace?

[millhaus]

thanks for such a fast reply. Use the transliterate unicode characters option doesn't help. the reader uses different code page. we usualy use cp 1252 and even if i set it as the source encoding it doesnt help. the best result i can get is to use lrt file with embed font but when using transliberation, it gets rid of all the interpunction š -> s etc. even the one which is correct. and read books without it is not pleasant
there are ways how to fix it by using other programs or by flashing the firmware.
the S & R feature works (with LRT files)! but i would need like 7 fields.
wouldnt be nice to have an all in one solution?

[kovidgoyal]

Umm transliterate unicode characters converts everything to english cahracters which all ebook readers support.

[user_none]

@Kovid, millhaus only wants to change the characters that don't display properly to ASCII not all unicode characters.

[kovidgoyal]

Ah. Now I get it.

[ldolse]

There's the configuration variable in the output profile to add 'just' un-supported letters there - I don't think it would be a great idea to add letters to the defaults because users taking the time to add fonts to their readers would complain, but maybe a 'tweak' could added so users could append their own list of characters.

@millhaus, I don't know if you've checked on the Sony sub-forums, but I thought that you could just stick a font file onto the Sony's filesystem directly and use Calibre's extra_css to have every ebook use it once converted. I might be mis-remembering though and it's possible that it only works with the latest generation devices.

Couldn't find it on Mobilread, though I'm sure the info is here somewhere, anyway here's a link to the regular Sony forum:
http://www.sonyreaderboards.com/foru...le=6&f=2&t=190

[millhaus]

thank you guys so much. i folowed the method from the sony forum and it works great for books.
but not for news. i was even trying to modify the recipe by adding this

 extra_css             = """
                                @font-face { 
  font-family: "arial"; 
  font-weight: normal; 
  font-style: normal; 
  src: url(res:///Data/fonts/arial.ttf); 
} 
body { 
  font-family: "arial", serif; 
  widows: 0; 
  orphans: 0 
}
                            """

(the extra css that works for the books) but without any success.

[mamon77]

If at all possible, I also vote for more S&R fields.

I have a similar problem as Millhaus except with Polish characters.
I added fonts to my 650, and sometimes I embed the fonts as well and generally it works well.

However, every now and then I'll run into a strange pdf which displays Polish characters fine in pdf but after pdf is processed they get converted incorrectly to something like a' instead of ą, s' instead of ś, etc. Flashing firmware, adding fonts, etc won't help because ePub gets written incorrectly. Transliterating to ASCII is not palatable, because I like my Polish text to have Polish letters where they should be.
I have verified that I can deal with that via S&R, it's just that now I have to do multiple passes.

Fom the UI point of view I think it could be dealt with by having a list of S&R expressions and an "add S&R" button which would add to the list. To make people's life easier for the common S&R tasks like removing -<pgnum>- we could even allow to name and save an S&R for later reuse...

[Elric]

Originally Posted by kovidgoyal

Given that you can chain regexes with | I don't really see the need for more.

Please explain chaining regular expressions. Google is coming up with nothing useful and I'm desperate for more S&R fields.

[Manichean]

Originally Posted by Elric

Please explain chaining regular expressions. Google is coming up with nothing useful and I'm desperate for more S&R fields.

See the section titled "But if I had a few varying strings I wanted to match, things get complicated?" in the tutorial.

[Elric]

Originally Posted by Manichean

See the section titled "But if I had a few varying strings I wanted to match, things get complicated?" in the tutorial.

Thanks a bunch for the pointer. I wasn't aware there was a tutorial here. I had been using some info from the Calibre site to get started w/ S&R and since I am already pretty familiar with regular expressions hadn't seen the S&R tutorial on this forum.