On 11/14/2013 01:43 PM, James Pearson wrote:
> I was browsing around Freenet's github profile this morning and ran across
> PyFCP (lib-pyFreenet).  There were a few things that immediately jumped out
> to me as projects I could work on, but given their scope I figured I'd get
> some feedback before investing the effort.
> 
> First, I'd like to convert the documentation from epydoc to sphinx.  Sphinx
> is much prettier by default, has a number of great themes, is not dead
> (Epydoc's last release was almost six years ago!), lends itself to
> user-documentation (not just API docs), and is supported by Read the Docs,
> a fantastic documentation project that has become the standard for Python
> projects.

Sounds good!

> Secondly, the package is in actuality many different things, tied together
> merely by subject (Freenet) and implementation language (Python); a good
> half of the readme is a description of the various projects.  Why should
> someone wanting to work with the Freenet protocol in Python need to install
> an IRC bot, or an XML-RPC server?  A split should, of course, consist of a
> number of forks from the current HEAD, so as to maintain revision history
> for each project.

Wow, that is incredibly crufty! An IRC bot for exchanging noderefs
should no longer even exist - automatically exchanging noderefs is
discouraged as it removes the need to socially engineer people to get
into a darknet. That sounds like a holdover from Freenet 0.5 days if
anything.

> (BTW, the 'requires' argument is misspelled 'requries' in the setup.py; I
> imagine setuptools would complain about this any time you attempt to build
> the package.)
> (BTW #2, what is the point of having separate -official and -staging
> repositories?  Is this not the point of branches?)

My understanding is that there are separate repositories because GitHub
does not support separate permissions for different branches within a
repository. I'm not sure if this is a legitimate reason. There's a fair
amount of momentum behind this setup, at least for fred, (it's in a
number of scripts) but I for one would be open to consolidation.

> My apologies if this is stepping on any toes; I'm not at all familiar with
> the development history of Freenet or PyFCP, so I am relying upon you to
> point out any reasons things should stay as they are.

I am not aware of such toes or reasons. I know ArneBab reviewed the
library a few months ago, but if I recall correctly that may have been
focused on the FCP parts.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: OpenPGP digital signature
URL: 
<https://emu.freenetproject.org/pipermail/tech/attachments/20131116/1120c574/attachment.pgp>

Reply via email to