Beta 2 of TiddlyWiki 2.4.0 now posted

[Martin Budden]

This is basically the same as Beta 1, with a few minor bug fixes.
Importantly we fixed the problem with the upgrade mechanism, so you
should now be able to test upgrading.

The upgrade mechanism is new in 2.4.0, so any testing (and the
resultant comments or bug reports) is very welcome.

Martin

[FND]

> This is basically the same as Beta 1, with a few minor bug fixes.

For the sake of completeness, here's the URL:
http://www.tiddlywiki.com/beta/

Interesting observation:
With the nightlies, I'm still experiencing the "excessive padding"
issue* (tested on Firefox 2.0.0.13 under Ubuntu Gutsy). However, the
beta2 file does *not* seem to have that problem.

I have no idea what might cause this bug - but I think it's not just a
CSS issue, as it also affects functionality; sometimes hitting save
(whether via clicking or by using the respective access key) does not
work on the first attempt.

-- F.

* Some controls - e.g. tabs and buttons - have excessive horizontal
and/or vertical padding, which disappears after clicking the respective
control once. This issue had been discovered in RippleRap (BlogTalk),
but it's been there before (since v2.3.0?).

[FND]

> With the nightlies, I'm still experiencing the "excessive padding"
> issue* (tested on Firefox 2.0.0.13 under Ubuntu Gutsy). However, the
> beta2 file does *not* seem to have that problem.

Correction: After downloading the beta file, the padding issue is
occurring again - I'm at a loss here.

-- F.

[Eric Shulman]

On Apr 21, 2:14 am, FND <F...@gmx.net> wrote:
> > With the nightlies, I'm still experiencing the "excessivepadding"
> > issue* (tested on Firefox 2.0.0.13 under Ubuntu Gutsy). However, the
> > beta2 file does *not* seem to have that problem.
>

> Correction: After downloading the beta file, thepaddingissue is

> occurring again - I'm at a loss here.

This is apparently caused by the way in which the display is refreshed
during startup, i.e., in main()...

Here's what I've tracked down:

1) in config.notifyTiddlers[] there is an entry for "refreshDisplay":
{name: null, notify: refreshDisplay}

2) in main(), this array is loaded into the store.namedNotifications[]

3) in main(), store.notifyAll() is invoked

4) However, notifyAll() contains this bit of code:
------------------
for(var t=0; t<this.namedNotifications.length; t++) {
var n = this.namedNotifications[t];
if(n.name)
n.notify(n.name);
}
------------------
Thus, notifyAll() will *never* invoke refreshDisplay() because n.name
is 'null' for that entry in namedNotifcations[]

An easy fix for this problem appears to be to add an explicit call to
refreshDisplay() directly into main(), like this:
------------------
...
store.notifyAll();
t7 = new Date();
restart();
refreshDisplay();
...
------------------
I've tested with with an empty TW240b3, and it succeeds in eliminating
the 'excessive padding' problem, with no apparent side effects!

Curiously, if the call to refreshDisplay() occurs *before* restart(),
which renders the default tiddlers, it doesn't seem to have the
desired effect.... but, if invoked *after* restart(), the 'excessive
padding' is properly eliminated.

I'm sure there are other ways to resolve this particular problem, by
tweaking the logic in notifyAll(). However, without digging deep into
the entire notification/refresh architecture, it is unclear what other
effects that might have.

HTH,
-e

[Simon Baird]

Eric, well done finding this. This one has been bugging me for ages.

Simon.

--
simon...@gmail.com

[FND]

> Eric, well done finding this. This one has been bugging me for ages.

I emphatically second that - great job tracking this down!

Why hasn't this issue ever come up before 2.3 though - that description
makes me think it should have been there all along!?

I've raised a ticket and created a patch as per Eric's instructions:
http://trac.tiddlywiki.org/ticket/612

> An easy fix for this problem appears to be to add an explicit call to
> refreshDisplay() directly into main()

My only concern here is that this might have a negative impact on
startup time. A quick test shows that this does seem to be the case -
refreshDisplay() adds about 60 ms on my machine.
However, that fix is definitely worth it - and we can always aim for a
more efficient solution in v2.5.

-- F.

[JayFresh - http://jayfresh.wordpress.com]

Hi,

Nice one Eric, really pleased you've tracked this down!

I've done a little poking and found out that it's the height of the
sidebarOptions -> 'a' tag (on my browser) drops from 14.73333...px to
11px. Not sure what's setting that yet, though.



J.

[FND]

Kamilla found a workaround in another thread*:
#contentWrapper {
background-attachment: fixed;
}

I can confirm this works on Firefox - no idea why though...

-- F.

* http://groups.google.com/group/TiddlyWikiDev/t/58778b7b65049ead