Webware for Python

Ian Bicking wrote:

 > I personally tend towards fairly open security, but I'm not sure what 
other people feel.  We probably should have a backup plan at least for 
the svn repository.

Yep, that was my first thought as well. The Wiki should be also backuped 
regularly. Then we can be a bit more relaxed.

Does Tummy.com provide a backup already? And are safety patches 
regularly installed?

-- Chris

On Tue, 2005-09-27 at 16:23 -0500, Ian Bicking wrote:
> But now I feel I should have checked here first.  I don't know what our
> decision process is on this server.

I don't know either, but I don't have a problem with you having added
Christoph.  I trust the judgement of everyone who currently has sudo
access to useful their best judgement when creating additional accounts.
If in doubt, sending an email beforehand to this list is probably a
good idea.

On a related note, I'm the only one who gets the "root" email (i.e. errors
from cron jobs, etc.).  One a few occasions I've forwarded stuff to Ian
(because it looked like stuff he set up), but sometimes I don't have time
to deal with it.  One thought would be to simply forward this stuff to
webware-devel, but that is probably a bad idea security-wise as these 
emails could potentially contain sensitive information.  Maybe someone has a
better idea.  In any case, I'd invite a few more people to add their
addresses to /etc/aliases, just so I'm not the only one who sees this.

> Maybe to keep things under control we should at least cull things over
> time.  I think everyone currently with sudo access is fine (ianb,
> jdhildeb, echuck, eradman, gtalvola, and now chrisz).  Other user
> accounts on the system: fbar, mophilly, wlarsen.  I'm not sure what each
> is used for... mophilly is Mark Phillips, so he can do the release work.
> fbar is Frank Barknecht, but I'm not sure what the account was for.
> wlarsen is Wayne Larsen, which is not a familiar name to me.

Wayne Larsen was active on webware-discuss about a year and a half ago
(around the time we got the tummy server) and he was interested in
working on a webware-based CMS.  I created the account for him to use
for this purpose.  I don't think he's used the account much though,
and hasn't logged in in over a year.

--
Jason D. Hildebrand
T: 204 775 1212
E: jason@...

On Tue, 2005-09-27 at 16:23 -0500, Ian Bicking wrote:
> I personally tend towards fairly open security, but I'm not sure what
> other people feel.  We probably should have a backup plan at least for
> the svn repository.

The company I work for (PeaceWorks Computer Consulting) is willing to 
provide backup space to the Webware project.

I would implement the backup as a daily cron job which tars up the 
subversion repository and wiki directories and rsyncs them to one of our 
servers, and would also keep a couple of older backups (from previous weeks)
around so that we have more than just the previous day.

Comments?  I will implement this within the next few days unless there are
objections.

peace,
Jason

--
Jason D. Hildebrand
T: 204 775 1212
E: jason@...

On 10/21/05, Jason Hildebrand <jason@...> wrote:
> On Tue, 2005-09-27 at 16:23 -0500, Ian Bicking wrote:
> > I personally tend towards fairly open security, but I'm not sure what
> > other people feel.  We probably should have a backup plan at least for
> > the svn repository.
>
> The company I work for (PeaceWorks Computer Consulting) is willing to
> provide backup space to the Webware project.
>
> I would implement the backup as a daily cron job which tars up the
> subversion repository and wiki directories and rsyncs them to one of our
> servers, and would also keep a couple of older backups (from previous wee=
ks)
> around so that we have more than just the previous day.
>
> Comments?  I will implement this within the next few days unless there ar=
e
> objections.

Sounds great to me. Thanks to you and your company.

-Chuck

I really would like to go for the 0.9 release now.

Mark, if you do not find the time, and nobody else volunteers, I can try 
to do it this weekend. Let me know whether I shall do it and which 
things to heed besides the ReleaseProcedures.html which seem to be outdated.

I'm suggesting to make the current SVN trunk the 0.9 version. I have 
fixed so many problems in the 0.9b3 that I think it would not be a good 
idea to still offer 0.9b3 for download. The current version has improved 
doco, working Admin and Examples context, cleaner HTML output, working 
unit tests, improved launch scripts, an up-to-date KidKit plugin with 
its own Examples context and many more improvements. I have tested it on 
several Unix versions (Linux, BSD, Solaris) and WinXP and with several 
applications in the last week. So I'm pretty confident that it is usable 
as a release - at least better than nothing and better than 0.9b3.

Ideally, if there are people who have time to test it, we could also 
make it a release candidate now, let some more people test it and let 
the actual release follow very quickly (one or two weeks later maximum).

-- Christoph

On 10/21/05, Christoph Zwerschke <cito@...> wrote:
> I really would like to go for the 0.9 release now.
[snip]

I support this and will put in some testing tomorrow (Saturday).

-Chuck

On 10/21/05, Chuck Esterbrook <chuck.esterbrook@...> wrote:
> On 10/21/05, Christoph Zwerschke <cito@...> wrote:
> > I really would like to go for the 0.9 release now.
> [snip]
>
> I support this and will put in some testing tomorrow (Saturday).
>
> -Chuck

I've completed my testing. I fixed a bug in MiddleKit and checked in a
few other tweaks. I then banged on one of my applications and had zero
problems. Looks good to me,

-Chuck

On 10/22/05, Chuck Esterbrook <chuck.esterbrook@...> wrote:
> On 10/21/05, Chuck Esterbrook <chuck.esterbrook@...> wrote:
> > On 10/21/05, Christoph Zwerschke <cito@...> wrote:
> > > I really would like to go for the 0.9 release now.
> > [snip]
> >
> > I support this and will put in some testing tomorrow (Saturday).
> >
> > -Chuck
>
> I've completed my testing. I fixed a bug in MiddleKit and checked in a
> few other tweaks. I then banged on one of my applications and had zero
> problems. Looks good to me,
>
> -Chuck

Woops. I spoke too soon. Under my old Webware workspace, which could
be several weeks old or older, hitting Ctrl-C to stop the app server
worked fine (I do this when developing, not in production):

Listening for Adapter on ('127.0.0.1', 8086)
Listening for HTTP on ('127.0.0.1', 8080)
Ready (0.56 seconds after launch).
^C
Shutdown Called Sat Oct 22 16:48:13 2005
ThreadedAppServer: Shutting Down
Stopping AutoReload Monitor
Autoreload Monitor stopped
Shutting down the AppServer
Application is Shutting Down
Application has been succesfully shutdown.
AppServer has been shutdown
Terminate batch job (Y/N)? y
C:\...>


In the new workspace there are 2 problems. The output isn't nearly as
friendly and my command window is completely frozen. Additional ^C and
Break have no effect and I cannot get the prompt back. I have to close
the window.

^C
App server has been signaled to shutdown.
Shutting down at Sat Oct 22 16:45:57 2005
Traceback (most recent call last):
  File "Launch.py", line 35, in ?
    main(sys.argv)
  File "Launch.py", line 31, in main
    WebKit.Launch.launchWebKit(args[1], appWorkPath, args[2:])
  File "C:\All\echuck\Work\NetEventsStudio\rundir\Launch.py", line
133, in launchWebKit

  File ".\WebKit\ThreadedAppServer.py", line 974, in main
  File ".\WebKit\ThreadedAppServer.py", line 857, in run
  File ".\WebKit\DebugAppServer.py", line 92, in mainloop
  File ".\WebKit\ThreadedAppServer.py", line 219, in mainloop
  File ".\WebKit\ThreadedAppServer.py", line 903, in shutDown
KeyboardInterrupt
*** TOTAL FREEZE AT THIS POINT ***

Isn't that a weird looking traceback too?

This is on WinXP Pro with Python 2.4. I have tried on Linux yet.

I'm out of time for this today, but we should resolve this before
cutting 0.9rc1. Has anybody made any changes in the past few months
that might relate to this?

-Chuck

> Woops. I spoke too soon. Under my old Webware workspace, which could
> be several weeks old or older, hitting Ctrl-C to stop the app server
> worked fine (I do this when developing, not in production):

Yes, I made some changes here, because I got exceptions when hitting 
Ctrl-C with the *old* version (0.9b3) (same platform WinXP+Py2.4).

The 0.9b3 gave errors like this when I hit Ctrl-C:

-------------------------------------------------------
Shutdown Called Sun Oct 23 02:38:12 2005
Traceback (most recent call last):
   File "WebKit\ThreadedAppServer.py", line 821, in run
ThreadedAppServer: Shutting Down
Stopping AutoReload Monitor
Shutting down the AppServer
Application is Shutting Down
     time.sleep(1.0)
IOError: [Errno 4] Interrupted function call

Exiting AppServer due to above exception
Exception in thread Thread-12:
Traceback (most recent call last):
   File "C:\Programme\Python\lib\threading.py", line 442, in __bootstrap
     self.run()
   File "C:\Programme\Python\lib\threading.py", line 422, in run
     self.__target(*self.__args, **self.__kwargs)
   File "WebKit\ThreadedAppServer.py", line 812, in windowsmainloop
     server.mainloop()
   File "WebKit\ThreadedAppServer.py", line 213, in mainloop
     client, addr = sock.accept()
   File "C:\Programme\Python\lib\socket.py", line 169, in accept
     sock, addr = self._sock.accept()
   File "C:\Programme\Python\lib\socket.py", line 144, in _dummy
     raise error(EBADF, 'Bad file descriptor')
error: (9, 'Bad file descriptor')
--------------------------------------------------------

I made changes so that sockets are properly closed etc. and when I hit 
Ctrl-C under Windows, everything closes smoothly in the current version 
like that:

App server has been signaled to shutdown.
Shutting down at Sun Oct 23 02:35:48 2005

Exiting AppServer due to keyboard interrupt.
ThreadedAppServer is shutting down...
Stopping AutoReload Monitor...
AppServer is shutting down...
Application is shutting down...
Application has been succesfully shutdown.
AppServer has been shutdown.

--------------------------------------------------------

How exactly did you start Webware? I opened a DOS box in the WebKit 
directory, typed "appserver" and then hit Ctrl-C.

Anyway, we should check and resolve this issue. But I think this can be 
done very quickly. We should also check this under Linux again. I will 
make some more tests.

-- Christoph

> In the new workspace there are 2 problems. The output isn't nearly as
> friendly and my command window is completely frozen.

I've tried some more times, hitting Ctrl-C multiple times and things 
like that, but it always stopped gracefully. I also tested under Linux.

But I think I found the problem. It seems to happen when you start the 
DebugAppServer, not the ThreadedAppServer. Can you confirm? If that's 
the case I think I can fix it very quickly.

-- Christoph

>> In the new workspace there are 2 problems. The output isn't nearly as
>> friendly and my command window is completely frozen.

The problem seems to be that the DebugAppServer freezes always when an 
exception is raised - in this case, a KeyboardInterrupt exception. In 
the old version, the Ctrl-C was trapped a little bit differently without 
raising a KeyboardInterrupt, so it did not happen there. I have checked 
in a change so that DebugAppServer will not care for KeyboardInterrupts, 
so the Ctrl-C issue should be fixed.

But the problem that DebugAppServer freezes when an exception is raised 
still exists, and it existed already in the 0.9b3. Can you have a look 
at that? I'm unsure what the expected behavior of the DebugAppServer 
should be.

-- Christoph

On 10/22/05, Christoph Zwerschke <cito@...> wrote:
> > In the new workspace there are 2 problems. The output isn't nearly as
> > friendly and my command window is completely frozen.
>
> I've tried some more times, hitting Ctrl-C multiple times and things
> like that, but it always stopped gracefully. I also tested under Linux.
>
> But I think I found the problem. It seems to happen when you start the
> DebugAppServer, not the ThreadedAppServer. Can you confirm? If that's
> the case I think I can fix it very quickly.
>
> -- Christoph

Yep.

-Chuck

On 10/22/05, Christoph Zwerschke <christoph.zwerschke@...> wrote:
> >> In the new workspace there are 2 problems. The output isn't nearly as
> >> friendly and my command window is completely frozen.
>
> The problem seems to be that the DebugAppServer freezes always when an
> exception is raised - in this case, a KeyboardInterrupt exception. In
> the old version, the Ctrl-C was trapped a little bit differently without
> raising a KeyboardInterrupt, so it did not happen there. I have checked
> in a change so that DebugAppServer will not care for KeyboardInterrupts,
> so the Ctrl-C issue should be fixed.
>
> But the problem that DebugAppServer freezes when an exception is raised
> still exists, and it existed already in the 0.9b3. Can you have a look
> at that? I'm unsure what the expected behavior of the DebugAppServer
> should be.
>
> -- Christoph

I'll look into it sometime on Sunday.

-Chuck

The problem with DebugAppServer was that it finished without stopping 
the "close thread" when an exception was raised and not handled. I've 
just checked in a fix for the DebugAppServer so that it shuts down 
properly. Please check whether it is ok that way.

-- Christoph

On 10/23/05, Christoph Zwerschke <cito@...> wrote:
> The problem with DebugAppServer was that it finished without stopping
> the "close thread" when an exception was raised and not handled. I've
> just checked in a fix for the DebugAppServer so that it shuts down
> properly. Please check whether it is ok that way.
>
> -- Christoph

Works great. I can tell the app server to exit gracefully with Ctrl-C.
Setting breakpoints works fine. Use the application works fine. I
tested from both the command prompt (well, not breakpoints) and from
WingIDE.

Thanks for the turn around.

I don't know of any other outstanding issues with the current Webware
repository.

-Chuck

Shall we make a 0.9rc first or directly release it as 0.9 and fix any 
problems in 0.9.1?

Unfortunately, here in Germany the weekend is already over so I will cut 
the release in the course of the next week unless somebody else jumps 
in. Today I'll just try to update the ReleaseProcedures.html. (Chuck, 
Mark: If you have any notes for changes here, let me know. Also, I will 
need admin privileges for the SourceForge project page. I already have 
the necesarry privleges for the w4py.org server.)

-- Christoph

Some things are still unclear concerning the ReleaseProcedures.

* A download mirror at ebad.pair.com is mentioned. I assume it is not 
used any more. We have the download directory at w4py.org and we can 
also use the SourceForge files section.

* The ReleaseHelper script cuts the release directly from the CVS 
repository -- but first, the repository is now SVN, and second, the SVN 
version has version tags in the Properties and html files instead of the 
actual version which must be first replaced by the actual version with 
"setversion.py" before making the tarball.

Is there a newer ReleaseHelper script? How have the 0.9 beta versions 
been cut? (Mark?) I could adapt these things but I don't want to do 
things which have been already done. Some hints would be helpful here.

-- Christoph

> * The ReleaseHelper script cuts the release directly from the CVS
 > repository -- but first, the repository is now SVN, and second,
 > the SVN version has version tags in the Properties and html files
 > instead of the actual version which must be first replaced by the
 > actual version with "setversion.py" before making the tarball.

Just had another look at it and I think I understand: You make a tag, 
replace the version tags, check this in under the tag and then cut the 
release directly from the tag. The trunk keeps the version tags. Right?
I was a bit confused because usually, tags are not changed after they
have been copied. I will document this in the ReleaseProcedures.

Still, is there a newer version of ReleaseHelper that uses SVN instead 
of CVS and that handles the ReleaseNotes-X.Y files properly? For the 
tag, they should be *moved* to ReleaseNotes-0.9, and in the tag if it is 
a final release, they should be *copied* to ReleaseNotes-0.9, and 
ReleaseNotes-X.Y should be emptied (prepared for the next release).

-- Christoph

> and in the tag if it is a final release, they should be *copied*
 > to ReleaseNotes-0.9, and ReleaseNotes-X.Y should be emptied

I meant, in the *trunk*.

Christoph and everyone,

If you have time, please go ahead. As much as I want to help with this, 
I have not been able to make time.

  - Mark

On Oct 21, 2005, at 3:37 PM, Christoph Zwerschke wrote:

> Mark, if you do not find the time, and nobody else volunteers, I can 
> try to do it this weekend. Let me know whether I shall do it and which 
> things to heed besides the ReleaseProcedures.html which seem to be 
> outdated.