i18nutil/source/utility/unicode.cxx | 33 ++++++++++++++-------------------
include/i18nutil/unicode.hxx | 10 ++--------
sw/qa/extras/uiwriter/uiwriter7.cxx | 24 ++++++++++++++++++++++--
sw/source/uibase/shells/textsh1.cxx | 10 ++++++----
4 files changed, 44 insertions(+), 33 deletions(-)
New commits:
commit b54f188b38b2fdb7e3224e5826162d7186c23288
Author: Mike Kaganski <mike.kagan...@collabora.com>
AuthorDate: Wed Aug 28 18:23:55 2024 +0500
Commit: Mike Kaganski <mike.kagan...@collabora.com>
CommitDate: Wed Aug 28 20:52:12 2024 +0200
Drop newly unused ToggleUnicodeCodepoint::CharsToDelete
Change-Id: I25654d911b1cc11f8c0e7d0ef3e87898418c3408
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/172535
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kagan...@collabora.com>
diff --git a/i18nutil/source/utility/unicode.cxx
b/i18nutil/source/utility/unicode.cxx
index ae40fa421c0a..a1e03bd5b41e 100644
--- a/i18nutil/source/utility/unicode.cxx
+++ b/i18nutil/source/utility/unicode.cxx
@@ -1270,19 +1270,6 @@ OUString ToggleUnicodeCodepoint::StringToReplace()
return maInput.toString();
}
-sal_uInt32 ToggleUnicodeCodepoint::CharsToDelete()
-{
- OUString sIn = StringToReplace();
- sal_Int32 nPos = 0;
- sal_uInt32 counter = 0;
- while( nPos < sIn.getLength() )
- {
- sIn.iterateCodePoints(&nPos);
- ++counter;
- }
- return counter;
-}
-
OUString ToggleUnicodeCodepoint::ReplacementString()
{
OUString sIn = StringToReplace();
diff --git a/include/i18nutil/unicode.hxx b/include/i18nutil/unicode.hxx
index b7722fd77196..ddb75784784f 100644
--- a/include/i18nutil/unicode.hxx
+++ b/include/i18nutil/unicode.hxx
@@ -115,12 +115,6 @@ public:
*/
OUString StringToReplace();
OUString ReplacementString();
-
- /**
- While sInput.getLength() returns the number of utf16 units to delete,
- this function returns the number of "characters" to delete -
potentially a smaller number
- */
- sal_uInt32 CharsToDelete();
};
#endif
commit ac0117713e6ce248c7725b57d709390f27dc4e62
Author: Mike Kaganski <mike.kagan...@collabora.com>
AuthorDate: Wed Aug 28 18:07:56 2024 +0500
Commit: Mike Kaganski <mike.kagan...@collabora.com>
CommitDate: Wed Aug 28 20:52:04 2024 +0200
tdf#162657: re-check the result of combining surrogates
They may constitute a combining mark, and require more characters
to the left to be processed.
Change-Id: Ie626badf66edaf44c56e0ba2d294304f139d2bca
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/172533
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kagan...@collabora.com>
diff --git a/i18nutil/source/utility/unicode.cxx
b/i18nutil/source/utility/unicode.cxx
index b0e1cd4d85ef..ae40fa421c0a 100644
--- a/i18nutil/source/utility/unicode.cxx
+++ b/i18nutil/source/utility/unicode.cxx
@@ -1068,7 +1068,7 @@ OUString unicode::formatPercent(double dNumber,
return aRet;
}
-bool ToggleUnicodeCodepoint::AllowMoreInput(sal_Unicode uChar)
+bool ToggleUnicodeCodepoint::AllowMoreInput(sal_uInt32 uChar)
{
//arbitrarily chosen maximum length allowed - normal max usage would be
around 30.
if( maInput.getLength() > 255 )
@@ -1092,12 +1092,20 @@ bool ToggleUnicodeCodepoint::AllowMoreInput(sal_Unicode
uChar)
if( rtl::isLowSurrogate(uChar) && maUtf16.isEmpty() &&
maInput.isEmpty() )
{
- maUtf16.append(uChar);
+ maUtf16.append(sal_Unicode(uChar));
return true;
}
if( rtl::isHighSurrogate(uChar) && maInput.isEmpty() )
- maUtf16.insert(0, uChar );
- //end of hex strings, or unexpected order of high/low, so don't
accept more
+ maUtf16.insert(0, sal_Unicode(uChar));
+ if (maUtf16.getLength() == 2)
+ {
+ assert(rtl::isHighSurrogate(maUtf16[0]) &&
rtl::isLowSurrogate(maUtf16[1]));
+ // The resulting codepoint may itself be combining, so may
allow more
+ sal_uInt32 nUCS4 = rtl::combineSurrogates(maUtf16[0],
maUtf16[1]);
+ maUtf16.setLength(0);
+ return AllowMoreInput(nUCS4);
+ }
+ // unexpected order of high/low, so don't accept more
if( !maUtf16.isEmpty() )
maInput.append(maUtf16);
if( !maCombining.isEmpty() )
@@ -1122,7 +1130,7 @@ bool ToggleUnicodeCodepoint::AllowMoreInput(sal_Unicode
uChar)
mbAllowMoreChars = false;
return false;
}
- maCombining.insert(0, uChar);
+ maCombining.insertUtf32(0, uChar);
break;
default:
@@ -1138,7 +1146,7 @@ bool ToggleUnicodeCodepoint::AllowMoreInput(sal_Unicode
uChar)
if( !maCombining.isEmpty() )
{
- maCombining.insert(0, uChar);
+ maCombining.insertUtf32(0, uChar);
maInput = maCombining;
mbAllowMoreChars = false;
return false;
diff --git a/include/i18nutil/unicode.hxx b/include/i18nutil/unicode.hxx
index be08595e0b10..b7722fd77196 100644
--- a/include/i18nutil/unicode.hxx
+++ b/include/i18nutil/unicode.hxx
@@ -102,11 +102,11 @@ private:
public:
/**
- Build an input string of valid UTF16 units to toggle.
+ Build an input string of valid UTF16/UCS4 units to toggle.
-do not call the other functions until the input process is complete
-build string from Right to Left. (Start from the character to the
left of the cursor: move left.)
*/
- bool AllowMoreInput(sal_Unicode uChar);
+ bool AllowMoreInput(sal_uInt32 uChar);
/**
Validates (and potentially modifies) the input string.
diff --git a/sw/qa/extras/uiwriter/uiwriter7.cxx
b/sw/qa/extras/uiwriter/uiwriter7.cxx
index 4b3f1c9c533f..18c7fefd3248 100644
--- a/sw/qa/extras/uiwriter/uiwriter7.cxx
+++ b/sw/qa/extras/uiwriter/uiwriter7.cxx
@@ -2405,8 +2405,8 @@ CPPUNIT_TEST_FIXTURE(SwUiWriterTest7,
testUnicodeNotationToggle)
sDocString =
pWrtShell->GetCursor()->GetPointNode().GetTextNode()->GetText();
CPPUNIT_ASSERT_EQUAL(sOriginalDocString, sDocString);
- constexpr OUString sWithCombiningSMPName = u"xyzU+e0101"_ustr;
- constexpr OUString sWithCombiningSMP = u"xyz\U000e0101"_ustr;
+ constexpr OUString sWithCombiningSMPName = u"xyzU+4faeU+e0101"_ustr;
+ constexpr OUString sWithCombiningSMP = u"xyz\U00004fae\U000e0101"_ustr;
pWrtShell->SplitNode();
pWrtShell->Insert2(sWithCombiningSMPName);
dispatchCommand(mxComponent, u".uno:UnicodeNotationToggle"_ustr,
aPropertyValues);
@@ -2416,9 +2416,13 @@ CPPUNIT_TEST_FIXTURE(SwUiWriterTest7,
testUnicodeNotationToggle)
dispatchCommand(mxComponent, u".uno:UnicodeNotationToggle"_ustr,
aPropertyValues);
sDocString =
pWrtShell->GetCursor()->GetPointNode().GetTextNode()->GetText();
// Before tdf#162656 fix, this failed with
- // - Expected: xyzU+e0101
- // - Actual : xyU+e0101
+ // - Expected: xyzU+4faeU+e0101
+ // - Actual : xyxU+e0101
// i.e., one codepoint to the left of the combining codepoint was removed
+ // Before tdf#162657 fix, this failed with
+ // - Expected: xyzU+4faeU+e0101
+ // - Actual : xyz侮U+e0101
+ // i.e., one codepoint to the left of the combining codepoint was not
converted
CPPUNIT_ASSERT_EQUAL(sWithCombiningSMPName, sDocString);
}
commit 5f1579c6fe56960f529e96f629d9a37352752345
Author: Mike Kaganski <mike.kagan...@collabora.com>
AuthorDate: Wed Aug 28 17:18:09 2024 +0500
Commit: Mike Kaganski <mike.kagan...@collabora.com>
CommitDate: Wed Aug 28 20:51:59 2024 +0200
tdf#162656: make sure to replace correct range
ToggleUnicodeCodepoint::CharsToDelete returns number of codepoints,
while SwWrtShell::DelLeft operates glyphs, which may consist of a
number of combining codepoints.
Change-Id: I394ee76c576f2e6f921cb30d125cce2f83559608
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/172532
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kagan...@collabora.com>
diff --git a/sw/qa/extras/uiwriter/uiwriter7.cxx
b/sw/qa/extras/uiwriter/uiwriter7.cxx
index 1070f5f26652..4b3f1c9c533f 100644
--- a/sw/qa/extras/uiwriter/uiwriter7.cxx
+++ b/sw/qa/extras/uiwriter/uiwriter7.cxx
@@ -2399,11 +2399,27 @@ CPPUNIT_TEST_FIXTURE(SwUiWriterTest7,
testUnicodeNotationToggle)
dispatchCommand(mxComponent, u".uno:UnicodeNotationToggle"_ustr,
aPropertyValues);
sExpectedString = "u+";
sDocString =
pWrtShell->GetCursor()->GetPointNode().GetTextNode()->GetText();
- CPPUNIT_ASSERT_EQUAL(sDocString, sExpectedString);
+ CPPUNIT_ASSERT_EQUAL(sExpectedString, sDocString);
dispatchCommand(mxComponent, u".uno:UnicodeNotationToggle"_ustr,
aPropertyValues);
sDocString =
pWrtShell->GetCursor()->GetPointNode().GetTextNode()->GetText();
- CPPUNIT_ASSERT_EQUAL(sDocString, sOriginalDocString);
+ CPPUNIT_ASSERT_EQUAL(sOriginalDocString, sDocString);
+
+ constexpr OUString sWithCombiningSMPName = u"xyzU+e0101"_ustr;
+ constexpr OUString sWithCombiningSMP = u"xyz\U000e0101"_ustr;
+ pWrtShell->SplitNode();
+ pWrtShell->Insert2(sWithCombiningSMPName);
+ dispatchCommand(mxComponent, u".uno:UnicodeNotationToggle"_ustr,
aPropertyValues);
+ sDocString =
pWrtShell->GetCursor()->GetPointNode().GetTextNode()->GetText();
+ CPPUNIT_ASSERT_EQUAL(sWithCombiningSMP, sDocString);
+
+ dispatchCommand(mxComponent, u".uno:UnicodeNotationToggle"_ustr,
aPropertyValues);
+ sDocString =
pWrtShell->GetCursor()->GetPointNode().GetTextNode()->GetText();
+ // Before tdf#162656 fix, this failed with
+ // - Expected: xyzU+e0101
+ // - Actual : xyU+e0101
+ // i.e., one codepoint to the left of the combining codepoint was removed
+ CPPUNIT_ASSERT_EQUAL(sWithCombiningSMPName, sDocString);
}
CPPUNIT_TEST_FIXTURE(SwUiWriterTest7, testTdf34957)
diff --git a/sw/source/uibase/shells/textsh1.cxx
b/sw/source/uibase/shells/textsh1.cxx
index 9b11e6f1b4ef..fa0a4fad018f 100644
--- a/sw/source/uibase/shells/textsh1.cxx
+++ b/sw/source/uibase/shells/textsh1.cxx
@@ -822,8 +822,9 @@ void SwTextShell::Execute(SfxRequest &rReq)
rWrtSh.InfoReadOnlyDialog(false);
break;
}
+ OUString stringToReplace = aToggle.StringToReplace();
SwRewriter aRewriter;
- aRewriter.AddRule( UndoArg1, aToggle.StringToReplace() );
+ aRewriter.AddRule( UndoArg1, stringToReplace );
aRewriter.AddRule( UndoArg2, SwResId(STR_YIELDS) );
aRewriter.AddRule( UndoArg3, sReplacement );
rWrtSh.StartUndo(SwUndoId::REPLACE, &aRewriter);
@@ -832,9 +833,10 @@ void SwTextShell::Execute(SfxRequest &rReq)
rWrtSh.ClearMark();
if( rWrtSh.IsInSelect() ) // cancel any in-progress keyboard
selection as well
rWrtSh.EndSelect();
-
- for( sal_uInt32 i=aToggle.CharsToDelete(); i > 0; --i )
- rWrtSh.DelLeft();
+ // Select exactly what was chosen for replacement
+ rWrtSh.GetCursor()->SetMark();
+
rWrtSh.GetCursor()->GetPoint()->AdjustContent(-stringToReplace.getLength());
+ rWrtSh.DelLeft();
rWrtSh.Insert2( sReplacement );
rWrtSh.EndUndo(SwUndoId::REPLACE, &aRewriter);
}
