>>>>> "Juergen" == Juergen Spitzmueller <[EMAIL PROTECTED]> writes:
Juergen> Jean-Marc Lasgouttes wrote:
>> Concerning this
>>
>> + // there's obviously no better way to omit indendation
>> in tabular cells :-( + && (par.ownerCode() !=
>> InsetBase::TEXT_CODE || isMainText())
>>
>> Do you mean "there is no obvious better way"?
Juergen> Read: *I* did not find any other way.
Your comment implies that there is no better way :)
>> What is the value of par.ownerCode() here?
Juergen> par.ownerCode() returns TEXT_CODE for tabular cells (not, as
Juergen> I had expected, TABULAR_CODE). The problem is that there
Juergen> seems to be no way (except the above, which strikes me rather
Juergen> vague) to tell leftMargin that a paragraph is part of a
Juergen> tabular cell.
So the test only works by chance, because a tabular cell is the only
insettext that is not subclassed, right? The right solution, then,
could be to create an InsetCell, which is a child of insettext and
returns CELL_CODE. But this is not something we want to do now.
We could move the test lyxCode == TEXT_CODE inside
InsetText::neverIndent, but this does not handle the isMainText()
case.
What about
bool InsetText::neverIndent{} const
{
// only a tabular cell does that.
return lyxCode() == TEXT_CODE;
}
and then in the patch, test for
+ // in some insets, paragraphs are never indented
+ && (isMainText() || !par.inInset()->neverIndent())
[inInset is always tru AFAICS]
Would that work?
JMarc
