[Twisted-Python] Weekly Bug Summary

Bug summary ______________________ Summary for 2006-04-09 through 2006-04-16 Bugs opened: 18 Bugs closed: 9 Total open bugs: 589 (+9) |== Type Changes |== Priority Changes |== Component Changes |Defect: +5 |Highest: -1 |Conch: +1 |Enhancement: +5 |Normal: +10 |Core: +1 |Task: -1 |Low: +0 |Mail: +1 |Pydoctor: +6 |Release Management: +1 |Trial: +2 |Web2: -1 |Website: -1 |Words: -1 New / Reopened Bugs ______________________ ===== Highest ===== [#1628] TwistedEmacs should be in emacs/ in Twisted (opened by jml) (CLOSED, fixed) enhancement core http://twistedmatrix.com/trac/ticket/1628 [#1634] Add ralphm to the matrix (opened by ralphm) defect release management http://twistedmatrix.com/trac/ticket/1634 [#1637] [PATCH] twisted.mail examples use deprecated imports (opened by jerub) defect mail http://twistedmatrix.com/trac/ticket/1637 ===== Normal ===== [#1629] Unjellyable.unjellyFor does not return self (opened by ml.flex) defect core http://twistedmatrix.com/trac/ticket/1629 [#1630] trial test hangs if using the twisted enterprise connection pool (opened by harshaw) defect trial http://twistedmatrix.com/trac/ticket/1630 [#1631] Make tables showing listings of members of packages, classes, etc, sortable (opened by moof) enhancement pydoctor http://twistedmatrix.com/trac/ticket/1631 [#1632] "browse Twisted conch tickets" link wrong (opened by slamb) (CLOSED, fixed) defect website http://twistedmatrix.com/trac/ticket/1632 [#1633] Undocumented dependency on "Crypto" (opened by slamb) defect conch http://twistedmatrix.com/trac/ticket/1633 [#1635] Completely replace t.p.c.Interface with z.i.Interface (opened by ralphm) (CLOSED, duplicate) enhancement core http://twistedmatrix.com/trac/ticket/1635 [#1636] Completely replace t.p.c.Interface with z.i.Interface (opened by ralphm) enhancement core http://twistedmatrix.com/trac/ticket/1636 [#1638] provide a way for test collector to find custom test suites (opened by acapnotic) enhancement trial http://twistedmatrix.com/trac/ticket/1638 [#1639] Full package and class hierarchy pages. (opened by jknight) enhancement pydoctor http://twistedmatrix.com/trac/ticket/1639 [#1640] Child process environment cannot be fully overridden on Windows (opened by bwh) defect core http://twistedmatrix.com/trac/ticket/1640 [#1641] sort subclasses in pydoctor's class hierarchy pages (opened by mwh) enhancement pydoctor http://twistedmatrix.com/trac/ticket/1641 [#1642] pydoctor's nevowhtml needs tests (opened by mwh) enhancement pydoctor http://twistedmatrix.com/trac/ticket/1642 [#1643] package listings in pydoctor should be alpha-sorted (opened by mwh) enhancement pydoctor http://twistedmatrix.com/trac/ticket/1643 [#1645] force the rest of win32-select tests to pass (opened by glyph) defect core http://twistedmatrix.com/trac/ticket/1645 ===== Low ===== [#1644] pydoctor's nevowhtml is probably easy to speed up (opened by mwh) enhancement pydoctor http://twistedmatrix.com/trac/ticket/1644 Closed Bugs ______________________ ===== Highest ===== [#1560] Death to util.wait() (opened by jml, closed by jml, fixed) task core http://twistedmatrix.com/trac/ticket/1560 [#1628] TwistedEmacs should be in emacs/ in Twisted (opened by jml, closed by jml, fixed) enhancement core http://twistedmatrix.com/trac/ticket/1628 [#1612] Words (jabber code) still uses deprecated components API (opened by itamarst, closed by ralphm, fixed) defect words http://twistedmatrix.com/trac/ticket/1612 [#1615] stdio tests should be run for all reactors (opened by exarkun, closed by exarkun, fixed) enhancement core http://twistedmatrix.com/trac/ticket/1615 ===== Normal ===== [#1632] "browse Twisted conch tickets" link wrong (opened by slamb, closed by exarkun, fixed) defect website http://twistedmatrix.com/trac/ticket/1632 [#1635] Completely replace t.p.c.Interface with z.i.Interface (opened by ralphm, closed by ralphm, duplicate) enhancement core http://twistedmatrix.com/trac/ticket/1635 [#1596] Leftover use of childNotFound() in t.w2.twcgi.CGIDirectory (opened by grimmtooth, closed by dreid, fixed) defect web2 http://twistedmatrix.com/trac/ticket/1596 [#1320] Show current version of projects in sidebar, projects page (opened by itamarst, closed by exarkun, wontfix) enhancement website http://twistedmatrix.com/trac/ticket/1320 ===== Low ===== [#1487] Make LoopingCall deal with deferred returning callables (opened by ralphm, closed by ralphm, fixed) enhancement core http://twistedmatrix.com/trac/ticket/1487

glyph@divmod.com writes:
No. As its only current use-case is documenting twisted, I've been asking people to file tickets in tm.com's trac and after a few piled up in the website component, exarkun added a pydoctor component. Does this bother you somehow? Cheers, mwh -- Premature optimization is the root of all evil. -- Donald E. Knuth, Structured Programming with goto Statements

On Mon, 17 Apr 2006 09:20:58 +0100, Michael Hudson <mwh@python.net> wrote:
That makes sense.
Does this bother you somehow?
Nope. Just seems like yet another indication that it should really be lore --api ;-) Or rather, lore, like trial, should be able to take a file (HTML or Python) or module name.

glyph@divmod.com writes:
Good.
It doesn't and can't really work like that though, pydoctor really needs to work on a whole system/package at a time. Cheers, mwh -- I think my standards have lowered enough that now I think ``good design'' is when the page doesn't irritate the living fuck out of me. -- http://www.jwz.org/gruntle/design.html

On Mon, 17 Apr 2006 15:30:22 +0100, Michael Hudson <mwh@python.net> wrote:
Absolutely - you are suggesting that the parity is even greater than I am. 'trial twisted' ~= 'lore twisted'. Or perhaps 'lore zope.interface twisted <something that means the stdlib too>' One really nice thing that lore could do is look for API references in "howto" documentation (man, that stuff needs a better name) and generate the API documentation concurrently so that it can print warnings if such APIs don't exist. Of course, by no means impossible if they were separate projects. Pydoctor generates actually parseable output, as opposed to the crap soup of HTML that comes out of epydoc. Still, it would be super nice if there could be a single, integrated step for building the whole book in HTML and PDF form, then packaging the API docs into an indexed, annotated appendix, included within the LaTeX document, for example. That sort of thing seems like it'd be hard to pull off without the code being closely related, however, I guess lore could always just import pydoctor (assuming no pydoctor=>twisted dependency).

On Mon, 17 Apr 2006 22:32:42 -0400, James Y Knight <foom@fuhm.net> wrote:
It seems to me that pydoctor should remain separate -- I suspect it would be more useful for other projects if it was not integrated with lore.
Any reason why? Or just a general feeling of suspicion?

I concur. I'm currently using pydoctor (thanks Michael!), and have no need for lore. -- Nicola Larosa - http://www.tekNico.net/ Hateful, blasphemous, prejudiced, vulgar, rude, or ignorant remarks are the music of a free society, and the relentless patter of idiots is how we know we're in one. When all the words in our public conversation are fair, good, and true, it's time to make a run for the fence. -- Daniel Gilbert, December 2005

glyph@divmod.com writes:
Ah, right.
pydoctor also optionally generates a system pickle with all the information in, so you shouldn't really even need to parse anything...
I would also like a pony :)
pydoctor depends on nevow (practically speaking, anyway; there's a writer that just builds up HTML by string interpolation but all the funky new stuff is just in the nevow writer). It is also vaguely designed to be used as a library, so I think lore could import it. Cheers, mwh -- Also, remember to put the galaxy back when you've finished, or an angry mob of astronomers will come round and kneecap you with a small telescope for littering. -- Simon Tatham, ucam.chat, from Owen Dunn's review of the year

glyph@divmod.com writes:
No. As its only current use-case is documenting twisted, I've been asking people to file tickets in tm.com's trac and after a few piled up in the website component, exarkun added a pydoctor component. Does this bother you somehow? Cheers, mwh -- Premature optimization is the root of all evil. -- Donald E. Knuth, Structured Programming with goto Statements

On Mon, 17 Apr 2006 09:20:58 +0100, Michael Hudson <mwh@python.net> wrote:
That makes sense.
Does this bother you somehow?
Nope. Just seems like yet another indication that it should really be lore --api ;-) Or rather, lore, like trial, should be able to take a file (HTML or Python) or module name.

glyph@divmod.com writes:
Good.
It doesn't and can't really work like that though, pydoctor really needs to work on a whole system/package at a time. Cheers, mwh -- I think my standards have lowered enough that now I think ``good design'' is when the page doesn't irritate the living fuck out of me. -- http://www.jwz.org/gruntle/design.html

On Mon, 17 Apr 2006 15:30:22 +0100, Michael Hudson <mwh@python.net> wrote:
Absolutely - you are suggesting that the parity is even greater than I am. 'trial twisted' ~= 'lore twisted'. Or perhaps 'lore zope.interface twisted <something that means the stdlib too>' One really nice thing that lore could do is look for API references in "howto" documentation (man, that stuff needs a better name) and generate the API documentation concurrently so that it can print warnings if such APIs don't exist. Of course, by no means impossible if they were separate projects. Pydoctor generates actually parseable output, as opposed to the crap soup of HTML that comes out of epydoc. Still, it would be super nice if there could be a single, integrated step for building the whole book in HTML and PDF form, then packaging the API docs into an indexed, annotated appendix, included within the LaTeX document, for example. That sort of thing seems like it'd be hard to pull off without the code being closely related, however, I guess lore could always just import pydoctor (assuming no pydoctor=>twisted dependency).

On Mon, 17 Apr 2006 22:32:42 -0400, James Y Knight <foom@fuhm.net> wrote:
It seems to me that pydoctor should remain separate -- I suspect it would be more useful for other projects if it was not integrated with lore.
Any reason why? Or just a general feeling of suspicion?

I concur. I'm currently using pydoctor (thanks Michael!), and have no need for lore. -- Nicola Larosa - http://www.tekNico.net/ Hateful, blasphemous, prejudiced, vulgar, rude, or ignorant remarks are the music of a free society, and the relentless patter of idiots is how we know we're in one. When all the words in our public conversation are fair, good, and true, it's time to make a run for the fence. -- Daniel Gilbert, December 2005

glyph@divmod.com writes:
Ah, right.
pydoctor also optionally generates a system pickle with all the information in, so you shouldn't really even need to parse anything...
I would also like a pony :)
pydoctor depends on nevow (practically speaking, anyway; there's a writer that just builds up HTML by string interpolation but all the funky new stuff is just in the nevow writer). It is also vaguely designed to be used as a library, so I think lore could import it. Cheers, mwh -- Also, remember to put the galaxy back when you've finished, or an angry mob of astronomers will come round and kneecap you with a small telescope for littering. -- Simon Tatham, ucam.chat, from Owen Dunn's review of the year
participants (5)
-
exarkun@twistedmatrix.com
-
glyph@divmod.com
-
James Y Knight
-
Michael Hudson
-
Nicola Larosa