hi there, folks:
I'd really like to release 0.7.0 but I would like it to be at least a
little bit tested before I do so. Could those of you with CVS trees check
everything out and see if it performs as advertised? Deeper bugs than
that will have to wait for the next release, but I'd at least like to know
if it works for someone other than me.
______ __ __ _____ _ _
| ____ | \_/ |_____] |_____|
|_____| |_____ | | | |
@ t w i s t e d m a t r i x . c o m
I have started working with Twisted and found it to be a very powerful
package. That said, some modules are not documented very thoroughly (big
surprise in the world of software, right?) and I also found some
questionable code, too.
If I were to contribute fixes/improvements, what is the submission
I just glanced at https://github.com/twisted/twisted/pull/1417 <https://github.com/twisted/twisted/pull/1417> after it was merged and noticed a few problems:
- It exposes a new top-level name rather than setting the protocol version as a parameter
- The NEWS files are malformed which is going to lead to some confusing duplication in the changelog
- AMPv2 itself does not appear to be "standardized" - the page at https://amp-protocol.net/conversations_v2.html <https://amp-protocol.net/conversations_v2.html> appears to be a preliminary suggestion for how long values might be handled rather than something broadly implemented; for one thing, I'd never heard of it before and for another the site itself doesn't link the document. This would be fine as a first draft, but I think it needs to be given some more revisions, given that (for example) there are a number of backwards-compatible ways to do this.
- The layering is wrong because it puts the protocol-parsing into a leaf class in the hierarchy, when the parsing logic was deliberately isolated to a lower level to facilitate different framing mechanisms.
- As JP pointed out, the tests have a potential bug where they can leak errors between cases.
I don't normally like to revert folks' work once it's been reviewed and accepted, particularly when there's no process violation (broken CI, lack of code coverage etc), but I'm particularly concerned about lending Twisted's imprimatur to this protocol extension as a whole new version without much more context on who is implementing it and what other options were considered, particularly when I personally (nominally the inventor of AMP) don't like the direction it has taken :).
What do other folks think? Anyone else have more of a finger on the pulse of where the "v2" conversation has been happening?
I have continued to work slowly on this project. Named pipes are
supported, the samba client can connect and list available shares.
i am currently working on a vfs layer using deferToThread to provide
asynchronous file access
I have tried to keep the code divided into logically separate chunks .
my first chunk received comments from glyph 2 months ago but has
remained in the review queue since.
is there anyone available to review this code to progress it ?
There was a discussion over IRC about re-running failed Azure tests.
I was confused, as if I go to the Azure build page, I don't see any re-run
Maarten told me that the re-run link is on the GitHub page, and not on
See the screenshot below.
I hope this helps :)
Where does towncrier stand on review policies and etiquette? I
generally don't like to just jump into new projects and start reviewing
and merging but I don't see other activity on that front nor do I see
any guidance on how it should be done. If I know how to proceed
properly then I might dedicate some review time to it.
Can we have a new Twisted release within the next two months, say in
Nov./Dec. timeframe, or
sooner if you'd like?
In Twisted trunk, there are a lot of things that would be good to have in a
Here are some of the high order items:
*Python 3 Fixes*
- twist dns --pyzone now works on Python 3 (I had a project 6 months
ago where I would have liked to have this in a release)
- twisted.web.twcgi now works on Python 3
*Python 3.8 Fixes*
- all tests pass on all platforms on Python 3.8, including asyncio tests
*Python 3.9 Fixes*
- all tests pass on Python 3.9.0rc2 (at least on Mac and Linux), (thanks
to Michał Górny)
- Python 2.7 support has been dropped in trunk
- black code formatter has been integrated into the CI for Twisted, and
all Twisted code is now formatted with Black (thanks to Tom Most)
- mypy is now run as part of Twisted CI and checks the code for every
PR. Type annotations compatible with Python 3.5+ are slowly being added to
There are a lot of other small fixes and enhancements that people have
contributed since the last Twisted release.
I would like this new release to be the last release supporting Python 3.5,
so that Python 3.5 support can be removed in trunk, and PEP-526 style
variable annotations can be added in trunk.
Thanks for your consideration.