ebook-convert vs windows version

[zhoujianfu]

Hi!

I have 0.6.51 installed on windows 7 but also on an ubuntu server, and when I try to use ebook-convert in the unix command line, it doesn't have as much success as opening the books in windows and running convert from the gui.

Are these implementations somehow different? I have a lot of files (which I can attach) that convert okay via the windows 0.6.51 but give various errors when running ebook-convert on them (to the epub format). Of course, plenty of files work fine in both too!

Is this a known issue? Should I create a bug? I couldn't find anything about it in the bug tracker. Should I upload the actual ebooks?

Thanks Kovid, or anybody!

josh!

P.S. here's one example .fb2 file that converted okay in the gui but got this error using ebook-convert (with no options, to epub):

Traceback (most recent call last):
File "/usr/bin/ebook-convert", line 19, in <module>
sys.exit(main())
File "/usr/lib/calibre/calibre/ebooks/conversion/cli.py", line 254, in main
plumber.run()
File "/usr/lib/calibre/calibre/ebooks/conversion/plumber.py", line 905, in run
self.opts, self.log)
File "/usr/lib/calibre/calibre/ebooks/epub/output.py", line 159, in convert
self.workaround_ade_quirks()"
File "/usr/lib/calibre/calibre/ebooks/epub/output.py", line 474, in workaround_ade_quirks"
for rule in stylesheet.data.cssRules.rulesOfType(CSSRule.STYLE _RULE):"
AttributeError: 'CSSRuleList' object has no attribute 'rulesOfType'

I'll attach it renamed to .txt because .fb2 seems to be not an allowed attachment type!

[kovidgoyal]

you need to upgrade

cssutils to atleast 0.9.7a3

[zhoujianfu]

Awesome, that worked, thanks a bunch!

Heh.. now I still have just one other error on a few files, something to do with character sets..

UnicodeEncodeError: 'ascii' codec can't encode character u'\xf3' in position 26: ordinal not in range(128)

I'm just running "ebook-convert david.lit david.epub"

one more time help please?

thanks a ton!

[kovidgoyal]

post the full error message

[Turt99]

hrm, I was having similar issues with book converting for me in windows but not in linux, I upgraded my Calibre in linux and that helped a little but I will have to look into this more maybe I need to update something else too.

[zhoujianfu]

Sure, sorry about that!

Traceback (most recent call last):
File "/usr/bin/ebook-convert", line 19, in <module>
sys.exit(main())
File "/usr/lib/calibre/calibre/ebooks/conversion/cli.py", line 254, in main
plumber.run()
File "/usr/lib/calibre/calibre/ebooks/conversion/plumber.py", line 905, in run
self.opts, self.log)
File "/usr/lib/calibre/calibre/ebooks/epub/output.py", line 190, in convert
oeb_output.convert(oeb, tdir, input_plugin, opts, log)
File "/usr/lib/calibre/calibre/ebooks/oeb/output.py", line 45, in convert
path = os.path.abspath(unquote(item.href))
File "/usr/lib/calibre/calibre/startup.py", line 80, in my_abspath
path = path.encode(encoding)
UnicodeEncodeError: 'ascii' codec can't encode character u'\xf3' in position 26: ordinal not in range(128)"

[kovidgoyal]

your linux system proabbly isn't set up correctly with a utf-8 locale

what does

locale

return

[zhoujianfu]

Hmm!

# locale
LANG=en_US.UTF-8
LC_CTYPE="en_US.UTF-8"
LC_NUMERIC="en_US.UTF-8"
LC_TIME="en_US.UTF-8"
LC_COLLATE="en_US.UTF-8"
LC_MONETARY="en_US.UTF-8"
LC_MESSAGES="en_US.UTF-8"
LC_PAPER="en_US.UTF-8"
LC_NAME="en_US.UTF-8"
LC_ADDRESS="en_US.UTF-8"
LC_TELEPHONE="en_US.UTF-8"
LC_MEASUREMENT="en_US.UTF-8"
LC_IDENTIFICATION="en_US.UTF-8"
LC_ALL=

Does that seem okay?

Thanks again!

[kovidgoyal]

calibre-debug -c "from calibre.constants import preferred_encoding, filesystem_encoding; print preferred_encoding, filesystem_encoding"

[zhoujianfu]

I ran that and got:

UTF-8 UTF-8

[kovidgoyal]

that doesn't ake any sense, the error is complaining about using the ascii encoding but you're saying the encoding is utf-8

are you launching ebook-convert from some sort of special environment?

[chaley]

Calibre is clearly trying to convert str() to unicode, in this case (assuming latin-1) the character ó. This could happen if command line arguments are bytestrings instead of unicode (and I think they are). It wouldn't happen in the GUI, because it represents and processes paths in UTF-8.

If I am right, things should work if there are no non-ASCII characters anywhere in the full pathname of the book.

[zhoujianfu]

Ah, dohhh!

My apologies, I ran locale and that test command as ROOT and then I've been running the ebook-convert as a user. Running locale as the user gives "POSIX"! I've switched the user to utf-8 and it worked!

(chaley, it looks like the filename part didn't matter.. when I renamed it to just ascii "duh.lit" it still wouldn't convert when I had the wrong locale, and when I keep it with a funny character in the filename it still works as long as I have the right locale!)

Thanks again!!