... so the numbers are repainted, both in the document body and in
the footnotes.

Change-Id: I4164f1bace1b984a841f520f4832a324b811ab3f

Needed for FTNNUM_CHAPTER and FTNNUM_DOC, as well as end-notes.

It should be enough to update from the start node until the end.

Change-Id: I47414f93b9ad29bc4ab4ee0fc25ae40fae160427

Add layout parameter to SwFormatFootnote::GetViewNumStr().

Change-Id: Iff9c9ff31d1793b66ba7e3ebdfa5cfe413324471

Add a separate footnote number to SwFormatFootnote for the case when
the layout hides redlines.

If the footnote is in a delete redline, the value will be an arbitrary
number, let's use the same number as the last visible footnote for
simplicity.

Change-Id: Ic7fff9d1bb2c90f5ad811b859c6501a193b7ae70

The number is pretty useless: none of the callers know what the proper
auto-counted number is going to be, so they either pass in the number
that the footnote already has, or 0.

It is particularly funny if the SwPaM is actually a range containing
multiple footnotes; fortunately such a call is apparently not possible
currently.

Just remove the parameter and use the existing footnote number directly.

Change-Id: I45dee999b8b0b630eff4347a563ccca2869f84cc

If it is merged, it doesn't have GetDep().

Change-Id: I5a362c03fe7fcb0a6246aef0a0ac2824c48ede36

It doesn't have a layout, of course.

Change-Id: Iec4809ec4a14e1fa9852d09619a8b344419b29e7

This never called Rst() in the loops because the m_nPosition wasn't updated.

Change-Id: I5a9cf47d9fe6d92bb7fccf255acbbd22f04b7f47

... check, which should accept the past-the-last-char index, because
as the comments say it returns the hard attributes of the SwTextNode
as well, so it should work for empty paragraphs too.

(similar fix was done in getCharacterAttributes() in IA2)

Change-Id: I01e1a8f66f103b63584e88ce785ac2044c5db2b9

It doesn't actually work with a similar logic to the other iterators,
because it iterates ByEnd but forwards, so the hints and the extents
don't come in a matching order.

To prevent complicating this further, replace it with a new
implementation that does only what the one client expects, and put it
directly in SwTextFormatter replacing the previous integer iterator
m_nHintEndIndex, so that it is created only once.

Change-Id: I144bfcf7e837a4fb0e7ec279edfba4732d0ae897

Was using the wrong node when iterating; also the tricky case of empty
or without-end hints at the start of an extent was wrong.

MergedAttrIter also shouldn't include wihout-end hints and non-empty
hints at the end of an extent.

Change-Id: Ia0776c1d3043cbd6d76fa04905b4937ebba53398

Change-Id: Ieee07d149d045b7953ea8d5489f2fb6aed13e5e4

Change-Id: Ib401e6d453519890badf13320e6f2123d1a3b11b

There is an inconsistency between getHyperLinkCount and getHyperLink,
as the latter uses index 0 as a no-op for some ToC link that was never
implemented, so all the real hyperlinks have their index offset by 1.

Also it's pointless to call getHyperLinkCount once per loop iteration.

(regression from 76c549eb)

Change-Id: Iec83b6680abbdff5d372ca40358bd1306f3bd991

BuildSelectedParas() needs a little tweak to create idempotent map
entries while it iterates over the nodes of a merged paragraph.

InvalidateShapeInParaSelection() has some dodgy code to check if flys
are selected; probably should check for the start pos. of the frame
for AT_PARA anchor.

Change-Id: I2abe5dff1fb7fd0b95cc6c78b756f250db5f0872

This looks like it wants to deduplicate paragraphs; just use a std::set
for the frames, and retrieve them with a particular SwPosition, which
should deduplicate fine.

Change-Id: I0d0bd5cc4eb8d41d0a2f64979ccfde62ebc3533a

Change-Id: Ie46d8846887e9282ef40d63255f8a70e22f304e9

Just fix the TextFrameIndex / GetTextNode usages.

This does not make the layout cache actually work if IsHideRedlines is
enabled, hence the layout cache remains disabled in that case.

Change-Id: I9aa1fb490c3dd4fd26165f2e60513ec267857d78

Change-Id: I734e27c95e20373f8de32f8db2183275c94144e5

Change-Id: Id625ee27e69800780c04ea80d888f606cde0f308

Change-Id: I37624c92549b4b3cd4d9616a4952163bf51526ef

lcl_SetWrong works on model positions.

Change-Id: I34353a09260e336c15696a7bc48a5922bfd92e4a

... and finally remove GetTextNode().

Change-Id: I7ad91bf709b705afcf7ada79fd2a97335329e8b7

Tricky usage of SwWrongList here, let's hope this works...

Also, there is a similar bug here, the nIndex is in a11y coordinates
but it was used directly as input to SwWrongList; better to convert
first.

Change-Id: I9b49b4dc84089c03df0e1302512d13d289d16161

There appears to be a bug here, in that the nIndex is an index into
the accessiblity string, but it was used without conversion as a
model position in SwPosition; let's try to fix that.

Change-Id: I7a43ceacfe59102577f39ab4b8fec3b495db345d

Change-Id: I33079154b4775b1df55693bf6046adbf27b0c1e3

... functions in SwAccessibleParagraph, with the new WrongListIterator.

Change-Id: Ie401dd867a06a1963e49fa54afa978ad1f1346f5

This is not very efficient but should be good enough for a11y use.

Change-Id: Ibb00cf4ae18effb09673f3f7d9b9b2e1d72413b1

Delicious copypasta!

Change-Id: I95ab5a1ef58b01f775065a970a0d7eb2b6bf6394

Do i smell copypasta?

Change-Id: I5450118b58540eee5ac8ff2e941d005fda4572c3

(regression from 1397e7dd)

Change-Id: I9daa478092209b2e7c168f45280bd5f873f674f2

Use MergedAttrIter for the iteration.  Fortunately all use of the
iterator is transient, i.e. it doesn't live past a UNO method return.

Change-Id: I4292a4ecafe94396a9d4cc64c2a26c5f28429ef3

There is an annoying problem here: how does a SwAccessibleHyperlink
check whether its corresponding model hint still exists or not?

The existing check via the hint-position appears not ideal, and with the
MergedPara the obvious extension is to add a SwTextNode pointer (or
node index?) which would make it even less ideal...

Instead, do as a lot of classes in unocore do, and register as a
listener on the pool item.

There is another aspect there in how the SwAccessibleHyperTextData
map gets disposed quite eagerly when anything changes, but the
SwAccessibleHyperlink can live longer than this map.

Change-Id: I7d0780cf28794b4ef68cff5d640190694f67530e

The SwAccessibleParagraph is a client of SwTextNode, just so that it can
dispose its mpParaChangeTrackInfo on any and all editing operations.

This won't work with merged paragraphs, but on the other hand it's not
necessary with merged paragraphs because the
initChangeTrackTextMarkupLists() does nothing if IsHideRedlines().

However there is the problem that the node and the frame might become
mis-matched if the frame is "moved around", so to avoid problems due to
that, listen on the SwTextFrame instead of the node, and have the
SwTextFrame relay any events it gets.

Though the frame is only SwClient, this is now possible thanks to
the SfxBroadcaster on the SwFrame class, so convert to SfxListener;
also the Accessible lifetime is tied to SwTextFrame anyway.

In UnHideRedlines(), send a dummy event too so that toggling Show/Hide
disposes the mpParaChangeTrackInfo too, which previously happened
as a side effect of moving the redlines around in the nodes-array.

Change-Id: I4e5a7107cb55d408266ee42463f9fa2a4a8b721b

Change-Id: I0f6d2c09bc06a6de1b46627b205f3c7edd9e594c

Remove this IA2 appendix, it's also dead in aoo/trunk...

Change-Id: Id13835fb20e13f5f84a289f6a94c99c3b55409df

Remove the branch that handes the text frame that isn't a text frame.

Call GetLineCount() on the correct frame in the IsInFly case.

Change-Id: I9378722740ea6d3a915f25c35c35c90742dc779c

Remove() will remove the node including its children, but
SvTreeList::Next() will actually return the first child, so the pChild
points to an entry that has been deleted.

Change-Id: Ia4bd75d64c8436ea03c0727a8d49ee0c34fda16f

This calls the same function as hyper_link_get_uri:
getAccessibleActionObject(), which is documented to return an URL
string, not an "object".  Since IA2 it even *does* return an URL string,
and none of the 3 implementations return an XAccessible, and there is
no obvious other method to call, so just remove the non-working
implementation so it doesn't print misleading exception warnings.

Change-Id: Ia32aef56439a8455698011ac636be264244656b3

To avoid conficts; TODO re-apply.

Change-Id: I697a943504857e41978f936d21beb038c4a0e010