On Jun 18, 2010, at 3:08 PM, Michael Foord wrote: > I'm still baffled as to how a bug in the cgi module (along with the > acknowledged email problems) is such a big deal. Was it reported and then > languished in the bug tracker? That would be bad ion its own but if it was > only recently discovered that indicates that it probably isn't such a big > deal - either way it needs fixing, but using Python for writing cgis hasn't > been a big use case for a long time.
That's one possible explanation. Another possible explanation is the product isn't being heavily exercised for serious work and that it has yet to be shaken-out thoroughly. There has been a disappointing lack of bug reports across the board for 3.x. That doesn't mean that the bugs aren't there and that they won't be reported when adoption is heavier. In the cases of email, mime handling, cgi and whatnot, the important point is not whether a given technology is popular. The important part is that it hints at the kind of bytes/text issues that people are going to face and that we will need to help them address (i.e. such as blobs containing multiple encodings, a need to use byte oriented tools such as md5 in conjunction with text oriented applications, etc.) One other thought: In addition to not getting many 3.x specific bug reports, we don't seem to be getting many 3.x specific help questions (i.e. asking about dictviews or how to make a priority queue in a environment where many callable don't support ordering operations, etc.). > Mark Lutz wrote > What I'm suggesting is that extreme caution be exercised from > this point forward with all things 3.X-related. Whether you > wish to accept this or not, 3.X has a negative image to many. > This suggestion specifically includes not abandoning current > 3.X email package users as a case in point. Ripping the rug > out from new 3.X users after they took the time to port seems > like it may be just enough to tip the scales altogether. A couple other areas that need work (some of them are minor): * BeautifulSoup was left behind when SGML parsing was removed from the standard lib. * Shelves were crippled for Windows users when bsddb was ripped out. * Lists containing None for missing values are no longer sortable. * The basic heapq approach to making a priority queue not longer works well. Simply decorating with (priority_level, callable_or_object) fails with two tasks at the same priority if the callable or other objects aren't orderable. Raymond P.S. I do think it would be great if we could direct some attention to parts of 3.x that are really nice. Am hoping that this conversation doesn't drown in negativity. Instead, it should focus on what improvements are needed to win broader adoption. _______________________________________________ Python-Dev mailing list Python-Dev@python.org http://mail.python.org/mailman/listinfo/python-dev Unsubscribe: http://mail.python.org/mailman/options/python-dev/archive%40mail-archive.com
