sw/qa/extras/htmlexport/data/char_border_and_font_color.fodt | 14
sw/qa/extras/htmlexport/htmlexport.cxx | 24
sw/source/filter/html/css1atr.cxx | 6
sw/source/filter/html/htmlatr.cxx | 416 +++++------
4 files changed, 233 insertions(+), 227 deletions(-)
New commits:
commit b13693d88f9a3f2a0ded26651b36d5c9f80673ac
Author: Mike Kaganski <mike.kagan...@collabora.com>
AuthorDate: Mon Mar 4 12:20:13 2024 +0600
Commit: Mike Kaganski <mike.kagan...@collabora.com>
CommitDate: Tue Mar 5 00:16:17 2024 +0600
tdf160017: make sure to emit the closing tags in correct order
This reimplements how the starts and ends of attributes are stored in
HTMLEndPosLst. Instead of a plain list, now it is a sorted map, with
positions as keys, and a vector of HTMLStartEndPos* as values.
In commit b94b1fe936ddc4a9b86fbeb9c9c6ab0fca52f0bc (CharBrd 9.1: HTML
filters, 2013-09-08), the character borders attributes started to be
set in a special order, in front of the position's other attributes,
to allow merging them. But that created a problem of knowing in which
order to close respective tags.
The change here sorts the closing tags for the current node only when
writing them. At this point, it is possible to consider the opening
positions correctly.
Change-Id: I466ffa1c0eb28874ded003035e0cf772e31585b3
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/164325
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kagan...@collabora.com>
diff --git a/sw/qa/extras/htmlexport/data/char_border_and_font_color.fodt
b/sw/qa/extras/htmlexport/data/char_border_and_font_color.fodt
new file mode 100644
index 000000000000..bda2ec63133c
--- /dev/null
+++ b/sw/qa/extras/htmlexport/data/char_border_and_font_color.fodt
@@ -0,0 +1,14 @@
+<?xml version="1.0" encoding="UTF-8"?>
+
+<office:document
xmlns:office="urn:oasis:names:tc:opendocument:xmlns:office:1.0"
xmlns:fo="urn:oasis:names:tc:opendocument:xmlns:xsl-fo-compatible:1.0"
xmlns:style="urn:oasis:names:tc:opendocument:xmlns:style:1.0"
xmlns:text="urn:oasis:names:tc:opendocument:xmlns:text:1.0"
xmlns:loext="urn:org:documentfoundation:names:experimental:office:xmlns:loext:1.0"
office:version="1.3" office:mimetype="application/vnd.oasis.opendocument.text">
+ <office:automatic-styles>
+ <style:style style:name="P1" style:family="paragraph">
+ <style:text-properties fo:color="#000000" loext:border="none"/>
+ </style:style>
+ </office:automatic-styles>
+ <office:body>
+ <office:text>
+ <text:p text:style-name="P1">foo</text:p>
+ </office:text>
+ </office:body>
+</office:document>
\ No newline at end of file
diff --git a/sw/qa/extras/htmlexport/htmlexport.cxx
b/sw/qa/extras/htmlexport/htmlexport.cxx
index 0e2396c39fd5..4afef3cccc96 100644
--- a/sw/qa/extras/htmlexport/htmlexport.cxx
+++ b/sw/qa/extras/htmlexport/htmlexport.cxx
@@ -3023,6 +3023,30 @@ CPPUNIT_TEST_FIXTURE(SwHtmlDomExportTest,
testReqIF_NoBrClearForImageWrap)
0);
}
+CPPUNIT_TEST_FIXTURE(SwHtmlDomExportTest, testReqIF_Tdf160017_spanClosingOrder)
+{
+ // Given a document with a paragraph having explicit font color and
character border properties:
+ createSwDoc("char_border_and_font_color.fodt");
+ // When exporting to reqif:
+ ExportToReqif();
+ SvMemoryStream aStream;
+ WrapReqifFromTempFile(aStream);
+ xmlDocUniquePtr pXmlDoc = parseXmlStream(&aStream);
+ // Without the fix, this would fail, because there was an extra closing
</reqif-xhtml:span>
+ CPPUNIT_ASSERT(pXmlDoc);
+}
+
+CPPUNIT_TEST_FIXTURE(SwHtmlDomExportTest, testHTML_Tdf160017_spanClosingOrder)
+{
+ // Given a document with a paragraph having explicit font color and
character border properties:
+ createSwDoc("char_border_and_font_color.fodt");
+ // When exporting to HTML:
+ save("HTML (StarWriter)");
+ // Parse it as XML (strict!)
+ // Without the fix, this would fail, because span and font elements closed
in wrong order
+ CPPUNIT_ASSERT(parseXml(maTempFile));
+}
+
CPPUNIT_PLUGIN_IMPLEMENT();
/* vim:set shiftwidth=4 softtabstop=4 expandtab: */
diff --git a/sw/source/filter/html/css1atr.cxx
b/sw/source/filter/html/css1atr.cxx
index 27d0c645ad56..af2fe301c96c 100644
--- a/sw/source/filter/html/css1atr.cxx
+++ b/sw/source/filter/html/css1atr.cxx
@@ -3335,14 +3335,16 @@ Writer& OutCSS1_SvxBox( Writer& rWrt, const
SfxPoolItem& rHt )
if( rHt.Which() == RES_CHRATR_BOX )
{
+ constexpr std::string_view inline_block("inline-block");
if( rHTMLWrt.m_bTagOn )
{
// Inline-block to make the line height changing correspond to the
character border
- rHTMLWrt.OutCSS1_PropertyAscii(sCSS1_P_display, "inline-block");
+ rHTMLWrt.OutCSS1_PropertyAscii(sCSS1_P_display, inline_block);
}
else
{
- HTMLOutFuncs::Out_AsciiTag( rWrt.Strm(),
Concat2View(rHTMLWrt.GetNamespace() + OOO_STRING_SVTOOLS_HTML_span), false );
+ if (!IgnorePropertyForReqIF(rHTMLWrt.mbReqIF, sCSS1_P_display,
inline_block))
+ HTMLOutFuncs::Out_AsciiTag( rWrt.Strm(),
Concat2View(rHTMLWrt.GetNamespace() + OOO_STRING_SVTOOLS_HTML_span), false );
return rWrt;
}
}
diff --git a/sw/source/filter/html/htmlatr.cxx
b/sw/source/filter/html/htmlatr.cxx
index a99140564301..247b22e5ff2f 100644
--- a/sw/source/filter/html/htmlatr.cxx
+++ b/sw/source/filter/html/htmlatr.cxx
@@ -1056,7 +1056,7 @@ public:
HTMLStartEndPos( const SfxPoolItem& rItem, sal_Int32 nStt, sal_Int32 nE );
- const SfxPoolItem* GetItem() const { return m_pItem.get(); }
+ const SfxPoolItem& GetItem() const { return *m_pItem; }
void SetStart(sal_Int32 nStt) { m_nStart = nStt; }
sal_Int32 GetStart() const { return m_nStart; }
@@ -1073,7 +1073,7 @@ HTMLStartEndPos::HTMLStartEndPos(const SfxPoolItem&
rItem, sal_Int32 nStt, sal_I
, m_pItem(rItem.Clone())
{}
-typedef std::vector<HTMLStartEndPos *> HTMLStartEndPositions;
+typedef std::map<sal_Int32, std::vector<HTMLStartEndPos*>>
HTMLStartEndPositions;
namespace {
@@ -1089,8 +1089,8 @@ enum HTMLOnOffState { HTML_NOT_SUPPORTED, //
unsupported Attribute
class HTMLEndPosLst
{
- HTMLStartEndPositions m_aStartLst; // list, sorted for start positions
- HTMLStartEndPositions m_aEndLst; // list, sorted for end positions
+ HTMLStartEndPositions m_aStartLst; // list, each position's elements
sorted by appearance order
+ HTMLStartEndPositions m_aEndLst; // list, no sort of elements in position
std::deque<sal_Int32> m_aScriptChgLst; // positions where script changes
// 0 is not contained in this list,
// but the text length
@@ -1108,8 +1108,7 @@ class HTMLEndPosLst
// Insert/remove a SttEndPos in/from the Start and End lists.
// The end position is known.
- void InsertItem_( HTMLStartEndPos *pPos, HTMLStartEndPositions::size_type
nEndPos );
- void RemoveItem_( HTMLStartEndPositions::size_type nEndPos );
+ void InsertItem_(HTMLStartEndPos* pPos);
// determine the 'type' of the attribute
HTMLOnOffState GetHTMLItemState( const SfxPoolItem& rItem );
@@ -1123,8 +1122,7 @@ class HTMLEndPosLst
sal_Int32 nEndPos );
// adapt the end of a split item
- void FixSplittedItem( HTMLStartEndPos *pPos, sal_Int32 nNewEnd,
- HTMLStartEndPositions::size_type nStartPos );
+ void FixSplittedItem(HTMLStartEndPos* pPos, sal_Int32 nNewEnd);
// insert an attribute in the lists and, if necessary, split it
void InsertItem( const SfxPoolItem& rItem, sal_Int32 nStart,
@@ -1142,6 +1140,8 @@ class HTMLEndPosLst
const SwHTMLFormatInfo *GetFormatInfo( const SwFormat& rFormat,
SwHTMLFormatInfos& rFormatInfos );
+ void OutEndAttrs(SwHTMLWriter& rWrt, std::vector<HTMLStartEndPos*>&
posItems);
+
public:
HTMLEndPosLst( SwDoc *pDoc, SwDoc* pTemplate, std::optional<Color>
xDfltColor,
@@ -1167,36 +1167,46 @@ public:
bool IsHTMLMode(sal_uLong nMode) const { return (m_nHTMLMode & nMode) !=
0; }
};
-}
-
-void HTMLEndPosLst::InsertItem_( HTMLStartEndPos *pPos,
HTMLStartEndPositions::size_type nEndPos )
+struct SortEnds
{
- // Insert the attribute in the Start list behind all attributes that
- // were started before, or at the same position.
- sal_Int32 nStart = pPos->GetStart();
- HTMLStartEndPositions::size_type i {0};
+ HTMLStartEndPositions& m_startList;
+ SortEnds(HTMLStartEndPositions& startList) : m_startList(startList) {}
+ bool operator()(const HTMLStartEndPos* p1, const HTMLStartEndPos* p2)
+ {
+ // if p1 start after p2, then it ends before
+ if (p1->GetStart() > p2->GetStart())
+ return true;
+ if (p1->GetStart() < p2->GetStart())
+ return false;
+ for (const auto p : m_startList[p1->GetStart()])
+ {
+ if (p == p1)
+ return false;
+ if (p == p2)
+ return true;
+ }
+ assert(!"Neither p1 nor p2 found in their start list");
+ return false;
+ }
+};
- while (i < m_aStartLst.size() && m_aStartLst[i]->GetStart() <= nStart)
- ++i;
- m_aStartLst.insert(m_aStartLst.begin() + i, pPos);
+#ifndef NDEBUG
+bool IsEmpty(const HTMLStartEndPositions& l)
+{
+ return std::find_if(l.begin(), l.end(), [](auto& i) { return
!i.second.empty(); }) == l.end();
+}
+#endif
- // the position in the End list was supplied
- m_aEndLst.insert(m_aEndLst.begin() + nEndPos, pPos);
}
-void HTMLEndPosLst::RemoveItem_( HTMLStartEndPositions::size_type nEndPos )
+void HTMLEndPosLst::InsertItem_(HTMLStartEndPos* pPos)
{
- HTMLStartEndPos* pPos = m_aEndLst[nEndPos];
+ // Character border attribute must be the first which is written out
because of border merge.
+ auto& posItems1 = m_aStartLst[pPos->GetStart()];
+ auto it = pPos->GetItem().Which() == RES_CHRATR_BOX ? posItems1.begin() :
posItems1.end();
+ posItems1.insert(it, pPos);
- // now, we are looking for it in the Start list
- HTMLStartEndPositions::iterator it = std::find(m_aStartLst.begin(),
m_aStartLst.end(), pPos);
- OSL_ENSURE(it != m_aStartLst.end(), "Item not found in Start List!");
- if (it != m_aStartLst.end())
- m_aStartLst.erase(it);
-
- m_aEndLst.erase(m_aEndLst.begin() + nEndPos);
-
- delete pPos;
+ m_aEndLst[pPos->GetEnd()].push_back(pPos);
}
HTMLOnOffState HTMLEndPosLst::GetHTMLItemState( const SfxPoolItem& rItem )
@@ -1350,23 +1360,25 @@ HTMLOnOffState HTMLEndPosLst::GetHTMLItemState( const
SfxPoolItem& rItem )
bool HTMLEndPosLst::ExistsOnTagItem( sal_uInt16 nWhich, sal_Int32 nPos )
{
- for (auto pTest : m_aStartLst)
+ for (const auto& [startPos, items] : m_aStartLst)
{
- if( pTest->GetStart() > nPos )
+ if (startPos > nPos)
{
// this attribute, and all attributes that follow, start later
break;
}
- else if( pTest->GetEnd() > nPos )
+
+ for (const auto* pTest : items)
{
- // the attribute starts before, or at, the current position and
- // ends after it
- const SfxPoolItem *pItem = pTest->GetItem();
- if( pItem->Which() == nWhich &&
- HTML_ON_VALUE == GetHTMLItemState(*pItem) )
+ if (pTest->GetEnd() > nPos)
{
- // an OnTag attribute was found
- return true;
+ // the attribute starts before, or at, the current position
and ends after it
+ const SfxPoolItem& rItem = pTest->GetItem();
+ if (rItem.Which() == nWhich && HTML_ON_VALUE ==
GetHTMLItemState(rItem))
+ {
+ // an OnTag attribute was found
+ return true;
+ }
}
}
}
@@ -1384,24 +1396,17 @@ bool HTMLEndPosLst::ExistsOffTagItem( sal_uInt16
nWhich, sal_Int32 nStartPos,
return false;
}
- for (auto pTest : m_aStartLst)
+ for (const auto* pTest : m_aStartLst[nStartPos])
{
- if( pTest->GetStart() > nStartPos )
- {
- // this attribute, and all attributes that follow, start later
- break;
- }
- else if( pTest->GetStart()==nStartPos &&
- pTest->GetEnd()==nEndPos )
+ if (pTest->GetEnd() == nEndPos)
{
- // the attribute starts before or at the current position and
- // ends after it
- const SfxPoolItem *pItem = pTest->GetItem();
- sal_uInt16 nTstWhich = pItem->Which();
+ // the attribute starts before or at the current position and ends
after it
+ const SfxPoolItem& rItem = pTest->GetItem();
+ sal_uInt16 nTstWhich = rItem.Which();
if( (nTstWhich == RES_CHRATR_CROSSEDOUT ||
nTstWhich == RES_CHRATR_UNDERLINE ||
nTstWhich == RES_CHRATR_BLINK) &&
- HTML_OFF_VALUE == GetHTMLItemState(*pItem) )
+ HTML_OFF_VALUE == GetHTMLItemState(rItem) )
{
// an OffTag attribute was found that is exported the same
// way as the current item
@@ -1413,55 +1418,55 @@ bool HTMLEndPosLst::ExistsOffTagItem( sal_uInt16
nWhich, sal_Int32 nStartPos,
return false;
}
-void HTMLEndPosLst::FixSplittedItem( HTMLStartEndPos *pPos, sal_Int32 nNewEnd,
- HTMLStartEndPositions::size_type
nStartPos )
+void HTMLEndPosLst::FixSplittedItem(HTMLStartEndPos* pPos, sal_Int32 nNewEnd)
{
+ // remove the item from the End list
+ {
+ auto& posItems = m_aEndLst[pPos->GetEnd()];
+ posItems.erase(std::remove(posItems.begin(), posItems.end(), pPos),
posItems.end());
+ }
// fix the end position accordingly
pPos->SetEnd( nNewEnd );
-
- // remove the item from the End list
- HTMLStartEndPositions::iterator it = std::find(m_aEndLst.begin(),
m_aEndLst.end(), pPos);
- OSL_ENSURE(it != m_aEndLst.end(), "Item not found in End List!");
- if (it != m_aEndLst.end())
- m_aEndLst.erase(it);
-
- // from now on, it is closed as the last one at the corresponding position
- HTMLStartEndPositions::size_type nEndPos {0};
- while (nEndPos < m_aEndLst.size() && m_aEndLst[nEndPos]->GetEnd() <=
nNewEnd)
- ++nEndPos;
- m_aEndLst.insert(m_aEndLst.begin() + nEndPos, pPos);
+ // from now on, it is closed at the corresponding position
+ m_aEndLst[nNewEnd].push_back(pPos);
// now, adjust the attributes that got started afterwards
- for (HTMLStartEndPositions::size_type i = nStartPos + 1; i <
m_aStartLst.size(); ++i)
+ const sal_Int32 nPos = pPos->GetStart();
+ for (const auto& [startPos, items] : m_aStartLst)
{
- HTMLStartEndPos* pTest = m_aStartLst[i];
- sal_Int32 nTestEnd = pTest->GetEnd();
- if( pTest->GetStart() >= nNewEnd )
- {
- // the Test attribute and all the following ones start, after the
- // split attribute ends
+ if (startPos < nPos)
+ continue;
+
+ if (startPos >= nNewEnd)
break;
+
+ auto it = items.begin();
+ if (startPos == nPos)
+ {
+ it = std::find(items.begin(), items.end(), pPos);
+ if (it != items.end())
+ ++it;
}
- else if( nTestEnd > nNewEnd )
+ for (; it != items.end(); ++it)
{
+ HTMLStartEndPos* pTest = *it;
+ const sal_Int32 nTestEnd = pTest->GetEnd();
+ if (nTestEnd <= nNewEnd)
+ continue;
+
// the Test attribute starts before the split attribute
// ends, and ends afterwards, i.e., it must be split, as well
+ // remove the attribute from the End list
+ auto& posItems = m_aEndLst[pTest->GetEnd()];
+ posItems.erase(std::remove(posItems.begin(), posItems.end(),
pTest), posItems.end());
// set the new end
pTest->SetEnd( nNewEnd );
-
- // remove the attribute from the End list
- it = std::find(m_aEndLst.begin(), m_aEndLst.end(), pTest);
- OSL_ENSURE(it != m_aEndLst.end(), "Item not found in End List!");
- if (it != m_aEndLst.end())
- m_aEndLst.erase(it);
-
- // it now ends as the first attribute in the respective position.
- // We already know this position in the End list.
- m_aEndLst.insert(m_aEndLst.begin() + nEndPos, pTest);
+ // it now ends in the respective position.
+ m_aEndLst[nNewEnd].push_back(pTest);
// insert the 'rest' of the attribute
- InsertItem( *pTest->GetItem(), nNewEnd, nTestEnd );
+ InsertItem( pTest->GetItem(), nNewEnd, nTestEnd );
}
}
}
@@ -1469,36 +1474,38 @@ void HTMLEndPosLst::FixSplittedItem( HTMLStartEndPos
*pPos, sal_Int32 nNewEnd,
void HTMLEndPosLst::InsertItem( const SfxPoolItem& rItem, sal_Int32 nStart,
sal_Int32 nEnd )
{
- HTMLStartEndPositions::size_type i;
- for (i = 0; i < m_aEndLst.size(); i++)
+ assert(nStart < nEnd);
+
+ for (auto& [endPos, items] : m_aEndLst)
{
- HTMLStartEndPos* pTest = m_aEndLst[i];
- sal_Int32 nTestEnd = pTest->GetEnd();
- if( nTestEnd <= nStart )
+ if (endPos <= nStart)
{
// the Test attribute ends, before the new one starts
continue;
}
- else if( nTestEnd < nEnd )
+ if (endPos >= nEnd)
+ {
+ // the Test attribute (and all that follow) ends, before the new
+ // one ends
+ break;
+ }
+
+ std::sort(items.begin(), items.end(), SortEnds(m_aStartLst));
+
+ for (HTMLStartEndPos* pTest : items)
{
if( pTest->GetStart() < nStart )
{
// the Test attribute ends, before the new one ends. Thus, the
// new attribute must be split.
- InsertItem_( new HTMLStartEndPos( rItem, nStart, nTestEnd ), i
);
- nStart = nTestEnd;
+ InsertItem_(new HTMLStartEndPos(rItem, nStart, endPos));
+ nStart = endPos;
}
}
- else
- {
- // the Test attribute (and all that follow) ends, before the new
- // one ends
- break;
- }
}
// one attribute must still be inserted
- InsertItem_( new HTMLStartEndPos( rItem, nStart, nEnd ), i );
+ InsertItem_(new HTMLStartEndPos(rItem, nStart, nEnd));
}
void HTMLEndPosLst::SplitItem( const SfxPoolItem& rItem, sal_Int32 nStart,
@@ -1509,59 +1516,49 @@ void HTMLEndPosLst::SplitItem( const SfxPoolItem&
rItem, sal_Int32 nStart,
// first, we must search for the old items by using the start list and
// determine the new item range
- for (HTMLStartEndPositions::size_type i = 0; i < m_aStartLst.size(); ++i)
+ for (auto& [nTestStart, items] : m_aStartLst)
{
- HTMLStartEndPos* pTest = m_aStartLst[i];
- sal_Int32 nTestStart = pTest->GetStart();
- sal_Int32 nTestEnd = pTest->GetEnd();
-
if( nTestStart >= nEnd )
{
// this attribute, and all that follow, start later
break;
}
- else if( nTestEnd > nStart )
+
+ for (auto it = items.begin(); it != items.end();)
{
+ HTMLStartEndPos* pTest = *it;
+ sal_Int32 nTestEnd = pTest->GetEnd();
+ if (nTestEnd <= nStart)
+ continue;
+
// the Test attribute ends in the range that must be deleted
- const SfxPoolItem *pItem = pTest->GetItem();
+ const SfxPoolItem& rTestItem = pTest->GetItem();
// only the corresponding OnTag attributes have to be considered
- if( pItem->Which() == nWhich &&
- HTML_ON_VALUE == GetHTMLItemState( *pItem ) )
+ if (rTestItem.Which() == nWhich && HTML_ON_VALUE ==
GetHTMLItemState(rTestItem))
{
- bool bDelete = true;
+ // if necessary, insert the second part of the split
+ // attribute
+ if (nTestEnd > nEnd)
+ InsertItem(pTest->GetItem(), nEnd, nTestEnd);
- if( nTestStart < nStart )
- {
- // the start of the new attribute corresponds to the new
- // end of the attribute
- FixSplittedItem( pTest, nStart, i );
- bDelete = false;
- }
- else
+ if (nTestStart >= nStart)
{
// the Test item only starts after the new end of the
// attribute. Therefore, it can be completely erased.
- m_aStartLst.erase(m_aStartLst.begin() + i);
- i--;
-
- HTMLStartEndPositions::iterator it
- = std::find(m_aEndLst.begin(), m_aEndLst.end(), pTest);
- OSL_ENSURE(it != m_aEndLst.end(), "Item not found in End
List!");
- if (it != m_aEndLst.end())
- m_aEndLst.erase(it);
- }
-
- // if necessary, insert the second part of the split
- // attribute
- if( nTestEnd > nEnd )
- {
- InsertItem( *pTest->GetItem(), nEnd, nTestEnd );
+ it = items.erase(it);
+ auto& posItems = m_aEndLst[pTest->GetEnd()];
+ posItems.erase(std::remove(posItems.begin(),
posItems.end(), pTest),
+ posItems.end());
+ delete pTest;
+ continue;
}
- if( bDelete )
- delete pTest;
+ // the start of the new attribute corresponds to the new
+ // end of the attribute
+ FixSplittedItem(pTest, nStart);
}
+ ++it;
}
}
}
@@ -1609,8 +1606,8 @@ HTMLEndPosLst::HTMLEndPosLst(SwDoc* pD, SwDoc* pTempl,
std::optional<Color> xDfl
HTMLEndPosLst::~HTMLEndPosLst()
{
- OSL_ENSURE(m_aStartLst.empty(), "Start List not empty in destructor");
- OSL_ENSURE(m_aEndLst.empty(), "End List not empty in destructor");
+ assert(IsEmpty(m_aStartLst) && "Start List not empty in destructor");
+ assert(IsEmpty(m_aEndLst) && "End List not empty in destructor");
}
void HTMLEndPosLst::InsertNoScript( const SfxPoolItem& rItem,
@@ -1896,53 +1893,25 @@ void HTMLEndPosLst::OutStartAttrs( SwHTMLWriter& rHWrt,
sal_Int32 nPos )
{
rHWrt.m_bTagOn = true;
- // Character border attribute must be the first which is written out
- // because of border merge.
- HTMLStartEndPositions::size_type nCharBoxIndex = 0;
- while (nCharBoxIndex < m_aStartLst.size()
- && m_aStartLst[nCharBoxIndex]->GetItem()->Which() != RES_CHRATR_BOX)
- {
- ++nCharBoxIndex;
- }
-
+ auto it = m_aStartLst.find(nPos);
+ if (it == m_aStartLst.end())
+ return;
// the attributes of the start list are sorted in ascending order
- for (HTMLStartEndPositions::size_type i = 0; i < m_aStartLst.size(); ++i)
+ for (HTMLStartEndPos* pPos : it->second)
{
- HTMLStartEndPos *pPos = nullptr;
- if (nCharBoxIndex < m_aStartLst.size())
+ // output the attribute
+ sal_uInt16 nCSS1Script = rHWrt.m_nCSS1Script;
+ sal_uInt16 nWhich = pPos->GetItem().Which();
+ if( RES_TXTATR_CHARFMT == nWhich ||
+ RES_TXTATR_INETFMT == nWhich ||
+ RES_PARATR_DROP == nWhich )
{
- if( i == 0 )
- pPos = m_aStartLst[nCharBoxIndex];
- else if( i == nCharBoxIndex )
- pPos = m_aStartLst[0];
- else
- pPos = m_aStartLst[i];
- }
- else
- pPos = m_aStartLst[i];
-
- sal_Int32 nStart = pPos->GetStart();
- if( nStart > nPos )
- {
- // this attribute, and all that follow, will be opened later on
- break;
- }
- else if( nStart == nPos )
- {
- // output the attribute
- sal_uInt16 nCSS1Script = rHWrt.m_nCSS1Script;
- sal_uInt16 nWhich = pPos->GetItem()->Which();
- if( RES_TXTATR_CHARFMT == nWhich ||
- RES_TXTATR_INETFMT == nWhich ||
- RES_PARATR_DROP == nWhich )
- {
- rHWrt.m_nCSS1Script = GetScriptAtPos( nPos, nCSS1Script );
- }
- HTMLOutFuncs::FlushToAscii( rHWrt.Strm() ); // was one time only -
do we still need it?
- Out( aHTMLAttrFnTab, *pPos->GetItem(), rHWrt );
- rHWrt.maStartedAttributes[pPos->GetItem()->Which()]++;
- rHWrt.m_nCSS1Script = nCSS1Script;
+ rHWrt.m_nCSS1Script = GetScriptAtPos( nPos, nCSS1Script );
}
+ HTMLOutFuncs::FlushToAscii( rHWrt.Strm() ); // was one time only - do
we still need it?
+ Out( aHTMLAttrFnTab, pPos->GetItem(), rHWrt );
+ rHWrt.maStartedAttributes[pPos->GetItem().Which()]++;
+ rHWrt.m_nCSS1Script = nCSS1Script;
}
}
@@ -1950,59 +1919,56 @@ void HTMLEndPosLst::OutEndAttrs( SwHTMLWriter& rHWrt,
sal_Int32 nPos )
{
rHWrt.m_bTagOn = false;
- // the attributes in the End list are sorted in ascending order
- HTMLStartEndPositions::size_type i {0};
- while (i < m_aEndLst.size())
+ if (nPos == SAL_MAX_INT32)
+ {
+ for (auto& element : m_aEndLst)
+ OutEndAttrs(rHWrt, element.second);
+ }
+ else
{
- HTMLStartEndPos* pPos = m_aEndLst[i];
- sal_Int32 nEnd = pPos->GetEnd();
+ auto it = m_aEndLst.find(nPos);
+ if (it != m_aEndLst.end())
+ OutEndAttrs(rHWrt, it->second);
+ }
+}
- if( SAL_MAX_INT32 == nPos || nEnd == nPos )
+void HTMLEndPosLst::OutEndAttrs(SwHTMLWriter& rWrt,
std::vector<HTMLStartEndPos*>& posItems)
+{
+ std::sort(posItems.begin(), posItems.end(), SortEnds(m_aStartLst));
+ for (auto it = posItems.begin(); it != posItems.end(); it =
posItems.erase(it))
+ {
+ HTMLStartEndPos* pPos = *it;
+ HTMLOutFuncs::FlushToAscii( rWrt.Strm() ); // was one time only - do
we still need it?
+ // Skip closing span if next character span has the same border
(border merge)
+ bool bSkipOut = false;
+ if( pPos->GetItem().Which() == RES_CHRATR_BOX )
{
- HTMLOutFuncs::FlushToAscii( rHWrt.Strm() ); // was one time only -
do we still need it?
- // Skip closing span if next character span has the same border
(border merge)
- bool bSkipOut = false;
- if( pPos->GetItem()->Which() == RES_CHRATR_BOX )
+ auto& startPosItems = m_aStartLst[pPos->GetEnd()];
+ for (auto it2 = startPosItems.begin(); it2 != startPosItems.end();
++it2)
{
- HTMLStartEndPositions::iterator it
- = std::find(m_aStartLst.begin(), m_aStartLst.end(), pPos);
- OSL_ENSURE(it != m_aStartLst.end(), "Item not found in Start
List!");
- if (it != m_aStartLst.end())
- ++it;
- while (it != m_aStartLst.end())
+ HTMLStartEndPos* pEndPos = *it2;
+ if( pEndPos->GetItem().Which() == RES_CHRATR_BOX &&
+ static_cast<const SvxBoxItem&>(pEndPos->GetItem()) ==
+ static_cast<const SvxBoxItem&>(pPos->GetItem()) )
{
- HTMLStartEndPos *pEndPos = *it;
- if( pEndPos->GetItem()->Which() == RES_CHRATR_BOX &&
- *static_cast<const SvxBoxItem*>(pEndPos->GetItem()) ==
- *static_cast<const SvxBoxItem*>(pPos->GetItem()) )
- {
- pEndPos->SetStart(pPos->GetStart());
- bSkipOut = true;
- break;
- }
- ++it;
+ startPosItems.erase(it2);
+ pEndPos->SetStart(pPos->GetStart());
+ auto& oldStartPosItems = m_aStartLst[pEndPos->GetStart()];
+ oldStartPosItems.insert(oldStartPosItems.begin(), pEndPos);
+ bSkipOut = true;
+ break;
}
}
- if( !bSkipOut )
- {
- Out( aHTMLAttrFnTab, *pPos->GetItem(), rHWrt );
- rHWrt.maStartedAttributes[pPos->GetItem()->Which()]--;
- }
- RemoveItem_( i );
- }
- else if( nEnd > nPos )
- {
- // this attribute, and all that follow, are closed later on
- break;
}
- else
+ if( !bSkipOut )
{
- // The attribute is closed before the current position. This
- // is not allowed, but we can handle it anyway.
- OSL_ENSURE( nEnd >= nPos,
- "The attribute should've been closed a long time ago" );
- i++;
+ Out( aHTMLAttrFnTab, pPos->GetItem(), rWrt );
+ rWrt.maStartedAttributes[pPos->GetItem().Which()]--;
}
+
+ auto& posItems2 = m_aStartLst[pPos->GetStart()];
+ posItems2.erase(std::remove(posItems2.begin(), posItems2.end(), pPos),
posItems2.end());
+ delete pPos;
}
}
