[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#724896: RFS: wiktionarytodict/20130929 [ITP] -- dictionaries for dictd




On 06/10/13 13:29, Tim Edwards wrote:

On 05/10/13 03:53, Paul Wise wrote:
On Fri, Oct 4, 2013 at 9:26 PM, Tim Edwards wrote:

No, not yet. It should be a pretty straightforward set of packages to
include as it's really just a packaging of dictionaries generated by my
scripts available at https://github.com/tkedwards/wiktionarytodict
How are the dictionaries modified? What are the dictionaries generated
from? Please note (DFSG item 2) that you have to include what they are
generated from (the preferred form for modification) in the source
package and generate the dictionaries during the package build
process. The package you have provided is not doing that right now so
it cannot be uploaded to Debian yet.

I can include the generation script (it's GPLv2 and available from https://github.com/tkedwards/wiktionarytodict) but I don't think that's very practical. The script regenerate_all.sh will download a 1.5GB file from Wiktionary.org and extract it, creating a nearly 4GB file. It then takes 2 hours on a quad-core desktop machine with all 4 cores at 100% CPU usage to generate the dictionaries.

The upstream project, http://tkedwards.github.io/wiktionarytodict/ (which is also run by me) produces the dictd format dictionaries as its output, which I'm then packaging up for Debian with no modifications - none are needed. Similar to how the tzdata package takes its data from upstream and just packages it up. I don't understand how this is against the DFSG.

Tim
If I look in the freedict dictionary packages (e.g. dict-freedict-eng-spa) which are already in Debian, so must conform to the DFSG, I don't see how the dictionaries were generated. The packaging only includes small modifications to the dictd files released by the upstream freedict project. At this stage my packages don't need to modify anything from the dictd files released by the upstream project http://tkedwards.github.io/wiktionarytodict/

Tim


Reply to: