[Springframework-developer] JavaDoc and possible issues in formcontroller

[Alef A. \(JTeam\) <al...@jt...>]

Juergen, Rod (and mayb other people),

While JavaDoc'ing parts of the mvc package (I've extensive documented =
about 5 or 6 of the controllers now), I came across three things in the =
AbstractFormController that sounded a little bit weird:

I can't find the place where the command object is bound in the session =
in case sessionForm is set to true. I've seen this work with the =
SimpleFormController, but never used the AbstractFormController and =
since the whole RequestDataBinder stuff is kind of new to me, you might =
wanna have a look (line 266 or something)

Also, it seems like the ServletRequestDataBinder gets created, even if =
bindOnNewForm is false (but that's probably got to do with the previous =
remark) (line 272)

Then - something slightly different - it seems like referenceData method =
does not get called on showing a new form, but only for the =
showForm()-methods that show a incorrectl validated command object (also =
in the same method I would have expected)

Maybe you can also have a look at the JavaDoc of Controller, =
AbstractController and for instance BaseCommandController and see if =
it's useful at all (otherwise of course I won't spend any time on it)...

Cheers,

Alef Arendsen

-----Oorspronkelijk bericht-----
Van: spr...@li... =
[mailto:spr...@li...] Namens =
j=C3=BCrgen h=C3=B6ller [werk3AT]
Verzonden: Wednesday, July 30, 2003 11:34 PM
Aan: pr...@se...; Spring Developers
Onderwerp: Re: [Springframework-developer] Current release plan


Hi Trevor,
=20
First of all, good questions resp. important ones indeed!
=20
<quote>
1. What still needs to be done for 0.9.1.  From that list, what is =
currently "assigned" to someone and what still needs help? </quote>
=20
Not a lot, actually. It's mainly about polishing the Petclinic and =
Countries sample apps, and testing the current codebase in everyone's =
own projects. Everything should be assigned so far, except the latter of =
course :-)
=20
I'm currently polishing a lot of the source code. I've also added 2 new =
features that aren't committed yet but will be by the end of this week, =
namely Hibernate flush mode support on HibernateTemplate and =
HibernateInterceptor, and read-only transactions (suppressing Hibernate =
flush on the transaction level). Furthermore, I've refined =
TransactionInterceptor and MapTransactionAttributeSource a bit, =
supporting more configuration options then before.
=20
<quote>
2. Same as #1 but for the 1.0 release.
</quote>

The first 1.0 RC will involve a package name change from com.interface21 =
to org.springframework. Besides that, there won't be a lot in terms of =
new features -maybe even less than from 0.9 to 0.9.1. Rod might already =
introduce some source-level attribute stuff to the AOP framework, but =
that isn't a requirement at all. Proper JMS and Web Service support are =
candidates too but can easily wait until 1.1.
=20
<quote>
3. What is the target time-frame for the 0.9.1 and 1.0 releases =
(obviously this is sketchy since it's "volunteer" work, but "gut =
guesses" - probably from Juergen :) - based on work left is what I'm =
hoping for.  The previous estimate was the 1.0RC1 by now. </quote>
=20
0.9.1 should be out any day now. I encourage everyone to try it the =
current CVS version before the weekend. A release some time next week =
should really be achievable. We *need* to get out a follow-up release =
quickly, if just because of all the enhancements to the Hibernate =
support that have a good chance of getting adopted promptly. My =
Hibernate article in the community area and my postings to the Hibernate =
forum have already been discussing some of them for a while.

<quote>
4. From Juergen's email which I referenced, it states "As far as I see, =
we don't need additional functionality for 1.0".  The main thing holding =
me back from adopting the current code base (I'm currently using a =
personally-modified version of 0.8) is the amount and frequency of =
changes to core functionality and the public api.  Some new features or =
minor bug fixes are ok/normal, but these sweeping changes are =
destructive to production code.  How close are we to finalizing the =
public api and the core code?  Is it likely that these major changes =
should be done for 0.9.1? </quote>
=20
Maybe I'm promising too much, but I don't see any major changes on the =
horizon. The introduction of the new DTD was probably the biggest change =
in the last few months, everything else was just about slight changes to =
the public API. I agree that we need stable APIs though to allow =
production apps to rely on them without *any* hassle. 0.9.1 should be as =
stable as can be in that respect, besides the package name change for =
1.0 RC which should just involve a search-and-replace.

<quote>
To expand on my fourth question with a recommendation (hopefully it =
makes sense to everyone).  If possible we should finalize the public api =
for version 0.9.1 rather than at the 1.0RC1.  This will allow usage in =
new development without fear of incompatibilities in the next couple =
months. (...) </quote>
=20
As I indicated above, I fully agree. We've been changing our apps at =
werk3AT quite often too, it would be fine it that wouldn't be necessary =
anymore as soon as possible. And I'd really like to get you on board =
again, in terms of working with a current Spring version :-)
=20
Regards,
Juergen
N=18HS^=E9=9A=8A[){([ Zz=DE=9An =044D w%=D8=A76i=17l=11&=CA=99 =
x+ljwE=E9=A2=BBj zz0=0E'=E5=8C=96Z=C9=A9z{^=DD=AE0=DA=8Av\=13bJ DN=1Bm =
=DE=B5brK=C9=AB&=20
?=C6=B4]4 M=DA=BD Z=DE=B7N M 5J  =
jg=ED=B0=AB=ED=B0=A2=1Dz=ED=BD=96=ED=B2=97x%R=CB=A6 (G^=EC=AE=BDh  lq  =
zm=D8=B6?X (=1E~zw X b=CB=9D?  jg=1Dz=ED=BD=96=ED=B2=97