[reportlab-users] Python 3.x progress

Michael Hipp michael at redmule.com
Tue Dec 3 21:19:28 EST 2013


On 12/3/2013 10:34 AM, Andy Robinson wrote:

> Hi everyone,

>

> More progress towards Python 3.3 compatibility, and some tentative

> decisions. Shout now if any of these worry you.

>

> Many of the failing tests now relate to old demos or documentation

> tools I would prefer to simply get rid of. After some debate

> internally about "using docutils" versus "eating our own dogfood", we

> want to keep the User Guide, which makes itself in PDF with no

> external dependencies, as it is our best test and a compelling demo.

>

> I am suggesting that we leave some things out of the next version...

>

> 1. Drop the auto-generated library reference (which walks through

> docstrings in a very weird manner). Replace it with Sphinx

> documentation.

> 2. Drop the auto-generated graphics reference and replace with some

> Sphinx extension to the above. It's just easier to browse this stuff

> on the web

> 3. Drop Pythonpoint, then possibly rebuild it later using new parser technology.

>

> We are hoping we might make an Ubuntu cut-off in February to get into 14.04.

>

> Other suggestions are welcome for "dead code" we could throw away.

>

> There is a separate issue on parsers used for paragraph content. This

> will involve a choice about backwards compatibility. The ideal

> engineering solution would be to demand strictly well-formed XML

> fragments when you construct a paragraph, and might punish anyone

> using ReportLab to print unquoted data. But that will be another post

> in a few days.


My (mostly unhelpful) suggestion is to do whatever enables you to get it
done the quickest.

I don't say that to be trite, but to recognize that it appears Python 3
is nearing some critical mass of widespread usage where it will look bad
on ReportLab if it is not ready.

I like the pdf User Guide, it is an excellent demonstration of its own
ability.

Thanks for your hard work,
Michael Hipp
Heber Springs, Arkansas, USA


More information about the reportlab-users mailing list