dataedit-update-operations

[Ian Eslick]

Am I missing something? dataedit-update-operations is getting called in the rendering function for gridedit which results in a circularity warning during dirty widget rendering. Is gridedit deprecated? I'm surprised someone else hasn't run across this.

Thank you,
Ian

[Ian Eslick]

In fact dataseq appears to be full of circular dirty writes for pagination, etc within render-widget-body.
This is the latest weblocks-dev I'm using.

Thanks,
Ian

> --
>
> You received this message because you are subscribed to the Google Groups "weblocks" group.
> To post to this group, send email to webl...@googlegroups.com.
> To unsubscribe from this group, send email to weblocks+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/weblocks?hl=en.
>
>

[Stephen Compall]

On Dec 22, 2009, at 12:25 PM, Ian Eslick wrote:

> Am I missing something? dataedit-update-operations is getting called in the rendering function for gridedit which results in a circularity warning during dirty widget rendering. Is gridedit deprecated? I'm surprised someone else hasn't run across this.

No, you're seeing that right. It's okay in gridedit's case because the updates it's making are idempotent. That's something we can't detect, and thought it would be better to warn anyway.

--
Stephen Compall

[Ian Eslick]

It's just that my logs are filling up with warnings. :) Is there a reason that update can't happen during the tree-shaking phase? Or can we mark the updated slots as non propagating? I haven't looked at it closely to evaluate it myself yet.

By the way, I just checked in some uri-parameter code and some improvements to the elephant store to support cached slots (a new elephant feature). Let me know if anyone detects problems with these checkins, I had some merge issues which I believed I resolved correctly but it's been awhile since I've done core weblocks development.

Thank you,
Ian

[Leslie P. Polzer]

On Dec 23 2009, 9:05 pm, Ian Eslick <esl...@media.mit.edu> wrote:
> It's just that my logs are filling up with warnings.  :)  Is there a reason that update
> can't happen during the tree-shaking phase?

I can't remember exactly but I think it's not possible because
gridedit makes some
late (i.e. rendering time) decisions. Someone would have to take a
closer look.

In the meantime you could just catch those style warnings and direct
them to a
separate log or ignore them.

> By the way, I just checked in some uri-parameter code and some improvements to the elephant store to support cached slots (a new elephant feature).  Let me know if anyone detects problems with these checkins, I had some merge issues which I believed I resolved correctly but it's been awhile since I've done core weblocks development.

Some comments...

http://bitbucket.org/S11001001/weblocks-dev/changeset/2240c38a10da/#chg-src/server.lisp

should be merged with the existing code in request-handler.lisp
(handle-client-request :around).

http://bitbucket.org/S11001001/weblocks-dev/changeset/1f79ea6672bb/#chg-src/request-handler.lisp

Line 192 should be #-(or sbcl ccl)

Great work with those uri parameters!

Leslie

[Ian Eslick]

Hi Leslie,

Good to hear from you. Would you like me to go ahead and make those changes and submit the patch?

[Leslie P. Polzer]

On Sun, Jan 10, 2010 at 11:19:08AM -0500, Ian Eslick wrote:

> Good to hear from you. Would you like me to go ahead and make those changes and submit the patch?

That would be great!

[Ian Eslick]

Done.