create empty layout in all text classes

[Georg Baum]

An empty layout is needed for docbook, fixed width table cells (see
toggleFixedWidth() in tabular.C, where "Standard" is used wrongly), ERT insets
and maybe more.
This should automatically be provided by LyXTextClass for all text classes.

[Georg Baum]

You can see one aspect of the missing empty layout in e.g. a g-brief-de
document. The default layout is Brieftext here, and it has a static label
"Text:". This is shown in ERT boxes, but it should not.

[Juergen Spitzmueller]

* #4037 has been marked as a duplicate of this bug. *

[Juergen Spitzmueller]

This also breaks g-brief documents with tables, see attachement 2032 from bug
4037.

This is a major issue. A user reported that he cannot upgrade from
LyX 1.3 because of this:
​http://marc.info/?l=lyx-devel&m=118399190607136&w=2

[Juergen Spitzmueller]

see attachement 2032

attachment 2032.

[Juergen Spitzmueller]

A quick fix for the g-breif issue would be to reintroduce the Standard layout in
g-brief. Almost any other layout has it (except for hollywood, I think).

Is it worth it, or shall we wait for a proper fix?

[Uwe Stöhr]

Subject: Re: create empty layout in all text classes

Is it worth it

Yes.

[Georg Baum]

Subject: Re: create empty layout in all text classes

Is it worth it

Yes.

But it would not fix the bug unless you made it the default style. And the
latter would be annoying for users. Remember that the style name "Standard"
has no special meaning at all. The only special style is the one marked as
default in the layout file, and it can have an arbitrary name.

[Juergen Spitzmueller]

But it would not fix the bug unless you made it the default style.

my proposal includes this change.

And the latter would be annoying for users. Remember that the style name
"Standard" has no special meaning at all. The only special style is the one
marked as default in the layout file, and it can have an arbitrary name.

Problem is that we have (really existing) users who are forced to stick with 1.3
because of this bug. Cf.
​http://marc.info/?l=lyx-devel&m=118399190607136&w=2

This is embarrasing.

[Juergen Spitzmueller]

anyway, this simple change wouldn't fix the case for existing documents (that
have been converted to 1.4/1.5 already). So it's not such a good idea anyway.

[Juergen Spitzmueller]

1.5.2 is frozen. This bug has to be postponed to 1.5.3.

[Juergen Spitzmueller]

1.5.3 is released.

[Richard Heck]

So what we really need to do here seems to be what Georg originally suggested:
We need an EMPTY layout that's used in certain cases, and is even forced to be
used in those cases. Right? It would not need to appear in the list of layouts,
and it would not need to be "Standard". Just forced in tables and such.

[Juergen Spitzmueller]

ad comment 12: yes, that's correct IMHO.

[Juergen Spitzmueller]

Soft freeze for 1.5.4: from now on, only critical bugs and regressions are
allowed.

[Richard Heck]

Hopefully fixed at 22966.

[Uwe Stöhr]

Not FIXED since the patch will not be 1.5.4.

[Juergen Spitzmueller]

not yet fixed in trunk. Try attachment 2032.

[Richard Heck]

Right. I missed this. If you create a table, you can see that (a) the layout is
actually set correctly, but (b) LyX gets confused because it can't tell that we
are in a table. Same problem on reading that attachment.

The fix is to change this code:
bool Paragraph::forceEmptyLayout() const
{

return inInset() && inInset()->forceEmptyLayout();

}

bool Paragraph::allowParagraphCustomization() const
{

return inInset() && inInset()->allowParagraphCustomization(0);

}

bool Paragraph::useEmptyLayout() const
{

return inInset() && inInset()->useEmptyLayout();

}
in Paragraph.cpp so that it treats table cells as requiring empty layout. A
better fix, and I seem to remember seeing a comment about this elsewhere in the
code, would be to have a real InsetTableCell-type inset.

[Juergen Spitzmueller]

The fix is to change this code:

could you dio this?

[Richard Heck]

I could add the same hack we have elsewhere, I suppose.

[Juergen Spitzmueller]

I fear this is too complicated and fragile for the 1.5 series.

[Richard Heck]

I agree. And we don't even have the table stuff working yet.

[Richard Heck]

Putative fix committed at 23662. This does work with attachment 2032.

[Juergen Spitzmueller]

excellent. Am I right that this has gotten too complicated for branch?

[Richard Heck]

At this late stage, I wouldn't try this in branch. It'll need extensive testing,
and by the time it gets it, we'll be at beta 2. It's too bad, but even porting
it to branch would be a lot of work at this point.

[Juergen Spitzmueller]

o.k.

[Dov Feldstern]

hmmm... the patch applied in r23662 seems to cause some new problems. See
​http://permalink.gmane.org/gmane.editors.lyx.devel/104068 for details.

[Richard Heck]

Additional fix committed at r23756. Perhaps this will do it.

[Richard Heck]

Another attempt at 23879. I hope very much that one will do. ;-)

[Juergen Spitzmueller]

LyX 1.6.0 is released.