Bug#1088869: ITP: multipart -- Python multipart/form-data parser

[Colin Watson <cjwatson@debian.org>]

Package: wnpp
Severity: wishlist
Owner: Colin Watson <cjwatson@debian.org>
X-Debbugs-Cc: debian-devel@lists.debian.org

* Package name    : multipart
  Version         : 1.2.1
  Upstream Contact: Marcel Hellkamp <marc@gsites.de>
* URL             : https://github.com/defnull/multipart
* License         : Expat
  Programming Lang: Python
  Description     : Python multipart/form-data parser

This module provides a fast incremental non-blocking parser for
multipart/form-data (HTML5, RFC7578), as well as blocking alternatives
for easier use in WSGI or CGI applications:

 * PushMultipartParser: Fast SansIO (incremental, non-blocking) parser
   suitable for ASGI, asyncio and other IO, time or memory constrained
   environments.
 * MultipartParser: Streaming parser that reads from a byte stream and
   yields memory- or disk-buffered MultipartPart instances.
 * WSGI Helper: High-level functions and containers for WSGI or CGI
   applications with support for both multipart and urlencoded form
   submissions.

For background on this, please see https://bugs.debian.org/1085728.  In
short, there has historically been an unfortunate namespace collision
between the multipart and python-multipart packages on PyPI which makes
it difficult for both to exist in the same environment, and now that
multipart is recommended as a replacement for parts of the old cgi
module that was removed from the standard library in Python 3.13, this
is a problem for some outlying parts of the addition of Python 3.13 to
Debian.

I'm aware that it's normally better to prefix "python-" to Python module
source package names to avoid name conflicts with other packaging
ecosystems.  However, in https://bugs.debian.org/1085728 Sandro made it
very clear that he didn't want to rename the existing python-multipart
source package to allow for this, and since that bug report already got
a bit more contentious than I'd have liked, I really don't want to turn
this into a bigger argument.  In the circumstances, just "multipart"
seems tolerable and the least bad option.

I plan to maintain this within the Debian Python Team.  It can
immediately be used to fix RC bugs in trac-customfieldadmin and
trac-wysiwig, and to clean up some vendoring in wadllib; there may be
others I haven't spotted yet.

-- 
Colin Watson (he/him)                              [cjwatson@debian.org]