tdf#92064 sw: Improve Tibetan layout performance

This change includes the following scalability improvements for
documents containing extremely long paragraphs:

- Reduces the size of layout contexts to account for line breaks.

- Disables a misbehaving glyph cache performance optimization for long
strings.

Change-Id: Ie9a3365076c0d112a7a655988d672a9f4609b42b
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/170420
Tested-by: Jenkins
Reviewed-by: Jonathan Clark <jonathan@libreoffice.org>
Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org>
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/172858
Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
Reviewed-by: Ilmari Lauhakangas <ilmari.lauhakangas@libreoffice.org>

3 files changed, 21 insertions, 4 deletions

diff --git a/sw/source/core/text/itrform2.cxx b/sw/source/core/text/itrform2.cxx
index 71b78bc0914d..3673b46ba247 100644
--- a/sw/source/core/text/itrform2.cxx
+++ b/sw/source/core/text/itrform2.cxx
@@ -1418,8 +1418,11 @@ SwTextPortion *SwTextFormatter::NewTextPortion( SwTextFormatInfo &rInf )
     if (!nCharWidthGuess)
         nCharWidthGuess = 1;
     auto nExpect = rInf.GetIdx() + TextFrameIndex(rInf.GetLineWidth() / nCharWidthGuess);
-    if (nExpect > rInf.GetIdx() && nNextChg > nExpect)
-        nNextChg = nExpect;
+    if (nExpect > rInf.GetIdx())
+    {
+        nNextChg = std::min(nNextChg, nExpect);
+        nNextContext = std::min(nNextContext, nExpect);
+    }
 
     // we keep an invariant during method calls:
     // there are no portion ending characters like hard spaces
@@ -1444,7 +1447,8 @@ SwTextPortion *SwTextFormatter::NewTextPortion( SwTextFormatInfo &rInf )
     // for the first text portion in a paragraph, or for any successive
     // portions that are outside of the bounds of the previous context.
     if (!rInf.GetLayoutContext().has_value()
-        || rInf.GetLayoutContext()->m_nEnd <= rInf.GetIdx().get())
+        || rInf.GetLayoutContext()->m_nBegin < rInf.GetLineStart().get()
+        || rInf.GetLayoutContext()->m_nEnd < nNextChg.get())
     {
         // The layout context must terminate at special characters
         sal_Int32 nEnd = rInf.GetIdx().get();
diff --git a/sw/source/core/text/portxt.cxx b/sw/source/core/text/portxt.cxx
index ee39f156c24a..ec0303c0665c 100644
--- a/sw/source/core/text/portxt.cxx
+++ b/sw/source/core/text/portxt.cxx
@@ -445,6 +445,12 @@ bool SwTextPortion::Format_( SwTextFormatInfo &rInf )
 
             SetLen( pGuess->BreakPos() - rInf.GetIdx() );
 
+            // Clamp layout context to the end of the line
+            if(auto stClampedContext = GetLayoutContext(); stClampedContext.has_value()) {
+                stClampedContext->m_nEnd = pGuess->BreakPos().get();
+                SetLayoutContext(stClampedContext);
+            }
+
             OSL_ENSURE( pGuess->BreakStart() >= pGuess->FieldDiff(),
                     "Trouble with expanded field portions during line break" );
             TextFrameIndex const nRealStart = pGuess->BreakStart() - pGuess->FieldDiff();
diff --git a/vcl/source/gdi/impglyphitem.cxx b/vcl/source/gdi/impglyphitem.cxx
index 1b40291736b7..946230a51762 100644
--- a/vcl/source/gdi/impglyphitem.cxx
+++ b/vcl/source/gdi/impglyphitem.cxx
@@ -364,6 +364,13 @@ const SalLayoutGlyphs* SalLayoutGlyphsCache::GetLayoutGlyphs(
             // part being underlined. Doing this for any two segments allows this optimization
             // even when the prefix of the string would use a different font.
             // TODO: Can those font differences be ignored?
+
+            // Shaping performance seems to scale poorly with respect to string length. Certain
+            // writing systems involve extremely long strings (for example, Tibetan: tdf#92064).
+            // In such cases, this optimization would be a net loss, and must be disabled.
+            constexpr sal_Int32 nOptLengthThreshold = 20000;
+            bool bEnableOptimization = (text.getLength() < nOptLengthThreshold);
+
             // Writer layouts tests enable SAL_NON_APPLICATION_FONT_USE=abort in order
             // to make PrintFontManager::Substitute() abort if font fallback happens. When
             // laying out the entire string the chance this happens increases (e.g. testAbi11870
@@ -374,7 +381,7 @@ const SalLayoutGlyphs* SalLayoutGlyphsCache::GetLayoutGlyphs(
                 const char* pEnv = getenv("SAL_NON_APPLICATION_FONT_USE");
                 return pEnv && strcmp(pEnv, "abort") == 0;
             }();
-            if (mLastSubstringKey.has_value() && !bAbortOnFontSubstitute)
+            if (bEnableOptimization && mLastSubstringKey.has_value() && !bAbortOnFontSubstitute)
             {
                 sal_Int32 pos = nIndex;
                 if (mLastSubstringKey->len < pos && text[pos - 1] == nbSpace)