The default value of :affects-dirty-status-p

[Daniel Borchmann]

Hi,

In df17f29, the default value for :affects-dirty-status-p was changed
From t to nil. I find this change quite severe, as it breaks backward
compatibilty and a lot of existing weblocks code (including mine). If
we would follow semantic versioning, this would even require a new major
release number. My wish would be to revert this change.

The corresponding commit refers to a “repeated rendering bug.” Can
somebody please elaborate what this is exactly? And, if it can be fixed
by changing the default value of :affects-dirty-status-p, couldn't we
just as well add some explicit statments like `:affects-dirty-status-p
nil` for the widgets involved in the particular instance of this bug?

Thanks for maintaining weblocks!

Daniel

--
Daniel Borchmann http://daniel.kxpq.de
GPG (Mail) 0849 473E 6BF0 B504 DF0B D640 455E 3610 01FF 778F

[o_z]

Hi, I've met situation where too many widgets were rendered only because their slots values were changed.

For any widget there are only few slots on which widget state depends. So there is no need to have :affects-dirty-status-p T for any slot.

Yes, it would require new major release number but let it be as it is.

I think :affects-dirty-status-p set to T by default is a mistake. It helps to avoid some (mark-dirty ...) statements but gives restrictions for cases where some widgets should not be rendered.

Thanks for response.

суббота, 3 января 2015 г., 11:10:06 UTC+2 пользователь Daniel Borchmann написал:

[Daniel Borchmann]

ghItlhpu' o_z <olex...@gmail.com>:

> Hi, I've met situation where too many widgets were rendered only because
> their slots values were changed.
> For any widget there are only few slots on which widget state depends. So
> there is no need to have :affects-dirty-status-p T for any slot.
> Yes, it would require new major release number but let it be as it is.
> I think :affects-dirty-status-p set to T by default is a mistake. It helps
> to avoid some (mark-dirty ...) statements but gives restrictions for cases
> where some widgets should not be rendered.

Ok, I see your point. Are you now going through all the widgets built
in in weblocks and see where `MARK-DIRTY` or `:AFFECTS-DIRTY-STATUS-P T`
is neccessary? I had some trouble with gridedit now updating
automatically when I wanted to add a new item (but recent commits in the
main repository may have fixed that; I will it out soon).

Best,

[o_z]

I will fix widgets bugs if they appear but not going to check all widgets. Weblocks selenium tests work now, that's enough for me at this point.

суббота, 10 января 2015 г., 12:34:37 UTC+2 пользователь Daniel Borchmann написал:

[Jason Miller]

Quicklisp just updated the version it has, and after updating, I noticed that the gridedit "Add" doesn't force a redraw, so has no effect.  Is that related to this change?

[Jason Miller]

I forgot to mention that adding a mark-dirty call to the end of dataedit-add-items-flow does fix the problem, though I'm not sure if that's the correct place to put it.

[o_z]

I suppose all is ok at this moment, bug is fixed.

среда, 21 января 2015 г., 18:52:22 UTC+2 пользователь Jason Miller написал: