Re: New upstream data in blends query
- To: firstname.lastname@example.org
- Subject: Re: New upstream data in blends query
- From: Ole Streicher <email@example.com>
- Date: Fri, 04 Mar 2016 13:46:46 +0100
- Message-id: <[🔎] firstname.lastname@example.org>
- References: <20160216140943.GE9931@an3as.eu> <email@example.com> <20160219120451.GA11006@an3as.eu> <firstname.lastname@example.org> <20160219144044.GB16892@an3as.eu> <email@example.com> <20160221081005.GH25698@an3as.eu> <firstname.lastname@example.org> <20160221113755.GI25698@an3as.eu> <email@example.com> <20160221132125.GJ25698@an3as.eu>
while you are discovering remote areas on Earth...
Andreas Tille <firstname.lastname@example.org> writes:
> On Sun, Feb 21, 2016 at 01:44:00PM +0100, Ole Streicher wrote:
>> >> There are no changes in blends/website.git yet?
>> > Pushed now, thanks for reminding me.
>> Thank you very much for this.
> You are welcome. It is just motivating if somebody joins the effort to
> work on the tasks pages. :-)
Something is not completely working here: I get no upstream_repository
entry, and also only a few upstream_bugs entries, although I have quite
many packages with d/u/metadata conteining them. Somehow they don't pass
> BTW, there is an encoding error if there are special HTML characters in the
> references. For example see:
> Please cite: Frédéric Mahé, Torbjørn Rognes, ...
> It would be really cool if you could hunt this down meanwhile. :-)
I had a look into this. The main problem is that the unicode stuff is
somehow "pragmatically" distributed over the files:
* Many strings (including the questioned one) are handled by
to_unicode() from blendsunicode.py, which explicitely encodes them as
XML entities. I now removed the encoding there; however we should
generally get rid of this function.
* The english long description, the publication title and authors are
also additionally explcicitely encoded with XML entities.
IMO, we should use "real" unicode strings (without any escapes)
everywhere in the code, and do the encoding only when we write to
files. Writing to files with a specific encoding can be simply done by
using "codecs.open()" instead of the open() builtin.
I changed this for the old webtools.
> Thanks. We are hoping for polar lights. ;-)
Did you see any?