Hi,
On Thu, Jul 23, 2015 at 1:30 PM, Andreas Tille <
andreas@an3as.eu> wrote:
>
> Hi Akshita,
>
> On Wed, Jul 22, 2015 at 07:55:24PM +0200, Andreas Tille wrote:
> > > The point is I do not have this UDD dump and hence I am unable to pin point
> > > where the encoding error occurs. Is there a way by which I can get this
> > > older UDD dump ?
> >
> > I'll get a fresh dump and try again.
>
> Just for the sake of completeness to answer your question: The problem
> remains with dump from 2015-07-22 (as somehow suspected).
>
I had also downloaded the UDD dump yesterday. But, I am unable to reproduce this error. You had mentioned in one of your previous mails that index.de.html (german) is of size 0. But I do not find any index files of size 0 when I run ./tasks_udd debian-med nor are there any other files created which are of size 0.
PFA the attached output of "$ ls -l index.* " in www/med/tasks_udd/ folder.
> I'd happily test the Python 3 version. :-)
>
I have almost completed porting tasks_udd.py and blendstasktools_udd.py to Python 3 using 2to3 program. Apart from the obvious changes required in the scripts (which were done by the 2to3 tool mostly):
1) I had to make some changes to the .xhtml files present in template folder.
2) Also, genshi seems to be broken for Python 3.4. The error I was getting :
Exception: Unhandled node type <class '_ast.NameConstant'>
was known [0] and I fixed it by applying all the patches for this specific error.
3) In addition to this, I needed to use set_client_encoding('utf-8') for psycopg2 as there was Encoding error in curs.fetchall(). This brings to my mind the question that while creating the database UDD, why can't we directly use encoding UTF-8 ? The reason seems to be that the encoding is SQL_ASCII since that is the format of export [1]. Won't it be beneficial if we encode using UTF-8 ?