Quote:
Originally Posted by Isomorpheus
This doesn't seem to work with Koreader. Testing with sdcv it's also broken. Just a heads up
|
Yeah, it's a bug with sdcv. For whatever reason, sdcv is REALLY slow when looking up words with synonyms (which you can work around with in
penelope by reprocessing the stardict dictionary using the --flatten-synonyms option which creates separate entries for each synonym. However, sdcv also has another bug where if a word has multiple definitions, it'll only return one result and ignore the rest, which you can then work around using the --merge-definitions option to create one giant entry instead). I originally used
pyglossary to perform the conversions, but it had limited options and the stardict output resulted in the above. The Kobo version was fine.
That said, I have since managed to track down a Windows version of the old
xdxf makedict utility, and it creates none of those problems so I used that to reconvert everything and things work rather well now. So go grab an updated copy.
I also added some <pre> tags to the Kobo version to try and maintain the original formatting of the dictd version so that things don't look like one giant blob of text. I don't have the time to style things myself beyond that simple thing but other people are welcome to try. If people don't like the resulting look because of wrapping or line breaks, an easy thing to try is shrinking the font size of the dictionary (in koreader through Advance Settings (DDICT_FONT_SIZE) or in Kobo Nickel via the "Dictionary text font-family/font-size/line-height" patch) so that the entire text block fits properly inside the dictionary window. Else, you'll need to figure out how to add your own HTML markup to the definitions to make it look the way you want (use the xdxf version as a basis to make your edits (or make your own using the dictd version as the source and using makedict to convert it to xdxf format) and when done, use makedict to convert it to stardict format (and then Penelope to convert to Kobo format if you want) and that way you won't need to worry about regenerating indexes).
Anyway, I've updated both my
WordNet 3.1 and
GCIDE 0.52 posts with all the source files in case people want to play around.