io/source/TextInputStream/TextInputStream.cxx | 114 +++++++++----------------- 1 file changed, 43 insertions(+), 71 deletions(-)
New commits: commit a9734683dfa8a36eafd6f32414408c9b8f4a221c Author: Mike Kaganski <[email protected]> AuthorDate: Sat Dec 7 18:33:17 2024 +0500 Commit: Mike Kaganski <[email protected]> CommitDate: Sat Dec 7 16:11:57 2024 +0100 Simplify OTextInputStream a bit Change-Id: I94c4432228c470ad31d15a765de87565b4ada9df Reviewed-on: https://gerrit.libreoffice.org/c/core/+/178039 Reviewed-by: Mike Kaganski <[email protected]> Tested-by: Jenkins diff --git a/io/source/TextInputStream/TextInputStream.cxx b/io/source/TextInputStream/TextInputStream.cxx index bcea56604e3c..94710fe6c8d5 100644 --- a/io/source/TextInputStream/TextInputStream.cxx +++ b/io/source/TextInputStream/TextInputStream.cxx @@ -169,28 +169,21 @@ OUString OTextInputStream::implReadString( const Sequence< sal_Unicode >& Delimi return aRetStr; // Only for bFindLineEnd - sal_Unicode cLineEndChar1 = 0x0D; - sal_Unicode cLineEndChar2 = 0x0A; + constexpr sal_Unicode cLineEndChar1 = ' '; + constexpr sal_Unicode cLineEndChar2 = ' '; sal_Int32 nBufferReadPos = 0; - sal_Int32 nCopyLen = 0; - bool bFound = false; - bool bFoundFirstLineEndChar = false; + sal_Int32 nCopyLen = -1; sal_Unicode cFirstLineEndChar = 0; - while( !bFound ) + while (true) { // Still characters available? if( nBufferReadPos == mnCharsInBuffer ) { // Already reached EOF? Then we can't read any more // Or no, so read new characters - if( mbReachedEOF || !implReadNext() ) { - if( bFoundFirstLineEndChar ) { - bFound = true; - nCopyLen = nBufferReadPos - 1; - } + if( !implReadNext() ) break; - } } // Now there should be characters available @@ -199,41 +192,35 @@ OUString OTextInputStream::implReadString( const Sequence< sal_Unicode >& Delimi if( bFindLineEnd ) { - if( bFoundFirstLineEndChar ) + if (cFirstLineEndChar != 0) { - bFound = true; - nCopyLen = nBufferReadPos - 2; - if( c == cLineEndChar1 || c == cLineEndChar2 ) + assert(nCopyLen >= 0); + // This is a check if the next character after a line end char is its second half + // Same line end char -> new line break; non-line-end char -> new line start + if ((c == cFirstLineEndChar) || (c != cLineEndChar1 && c != cLineEndChar2)) { - // Same line end char -> new line break - if( c == cFirstLineEndChar ) - { - nBufferReadPos--; - } - } - else - { - // No second line end char + // Not a two-char line end nBufferReadPos--; } + break; } else if( c == cLineEndChar1 || c == cLineEndChar2 ) { - bFoundFirstLineEndChar = true; - cFirstLineEndChar = c; + nCopyLen = nBufferReadPos - 1; // we know what to copy + cFirstLineEndChar = c; // take one more loop, to check if it's a two-char line end } } else if( comphelper::findValue(Delimiters, c) != -1 ) { - bFound = true; nCopyLen = nBufferReadPos; if( bRemoveDelimiter ) nCopyLen--; + break; } } // Nothing found? Return all - if( !nCopyLen && !bFound && mbReachedEOF ) + if (nCopyLen < 0) nCopyLen = nBufferReadPos; // Create string @@ -251,35 +238,40 @@ OUString OTextInputStream::implReadString( const Sequence< sal_Unicode >& Delimi sal_Int32 OTextInputStream::implReadNext() { - sal_Int32 nFreeBufferSize = mvBuffer.size() - mnCharsInBuffer; - if( nFreeBufferSize < READ_BYTE_COUNT ) - mvBuffer.resize(mvBuffer.size() * 2); - nFreeBufferSize = mvBuffer.size() - mnCharsInBuffer; + // Already reached EOF? Then we can't read any more + if (mbReachedEOF) + return 0; try { - sal_Int32 nRead = mxStream->readSomeBytes( mSeqSource, READ_BYTE_COUNT ); - sal_Int32 nTotalRead = nRead; - if( nRead == 0 ) + if (mxStream->readSomeBytes(mSeqSource, READ_BYTE_COUNT) == 0) + { mbReachedEOF = true; + return 0; + } // Try to convert - sal_uInt32 uiInfo; - sal_Size nSrcCvtBytes = 0; - sal_Size nTargetCount = 0; + sal_uInt32 uiInfo = mvBuffer.size() - mnCharsInBuffer < o3tl::make_unsigned(mSeqSource.getLength()) + ? RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOOSMALL + : 0; + const sal_Int32 nOldCharsInBuffer = mnCharsInBuffer; sal_Size nSourceCount = 0; while( true ) { - const sal_Int8 *pbSource = mSeqSource.getConstArray(); + if (uiInfo & RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOOSMALL) + { + mvBuffer.resize(mvBuffer.size() * 2); + } // All invalid characters are transformed to the unicode undefined char - nTargetCount += rtl_convertTextToUnicode( + sal_Size nSrcCvtBytes = 0; + mnCharsInBuffer += rtl_convertTextToUnicode( mConvText2Unicode, mContextText2Unicode, - reinterpret_cast<const char*>(&( pbSource[nSourceCount] )), - nTotalRead - nSourceCount, - mvBuffer.data() + mnCharsInBuffer + nTargetCount, - nFreeBufferSize - nTargetCount, + reinterpret_cast<const char*>(mSeqSource.getConstArray() + nSourceCount), + mSeqSource.getLength() - nSourceCount, + mvBuffer.data() + mnCharsInBuffer, + mvBuffer.size() - mnCharsInBuffer, RTL_TEXTTOUNICODE_FLAGS_UNDEFINED_DEFAULT | RTL_TEXTTOUNICODE_FLAGS_MBUNDEFINED_DEFAULT | RTL_TEXTTOUNICODE_FLAGS_INVALID_DEFAULT, @@ -287,41 +279,21 @@ sal_Int32 OTextInputStream::implReadNext() &nSrcCvtBytes ); nSourceCount += nSrcCvtBytes; - bool bCont = false; - if( uiInfo & RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOOSMALL ) - { - mvBuffer.resize(mvBuffer.size() * 2); - bCont = true; - } - if( uiInfo & RTL_TEXTTOUNICODE_INFO_SRCBUFFERTOOSMALL ) { // read next byte - static Sequence< sal_Int8 > aOneByteSeq( 1 ); - nRead = mxStream->readSomeBytes( aOneByteSeq, 1 ); - if( nRead == 0 ) + Sequence<sal_Int8> aOneByteSeq(1); + if (mxStream->readSomeBytes(aOneByteSeq, 1) == 0) { mbReachedEOF = true; - break; + return mnCharsInBuffer - nOldCharsInBuffer; } - sal_Int32 nOldLen = mSeqSource.getLength(); - nTotalRead++; - if( nTotalRead > nOldLen ) - { - mSeqSource.realloc( nTotalRead ); - } - mSeqSource.getArray()[ nOldLen ] = aOneByteSeq.getConstArray()[ 0 ]; - bCont = true; + mSeqSource = comphelper::concatSequences(mSeqSource, aOneByteSeq); } - - if( bCont ) - continue; - break; + else if (!(uiInfo & RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOOSMALL)) + return mnCharsInBuffer - nOldCharsInBuffer; // finished } - - mnCharsInBuffer += nTargetCount; - return nTargetCount; } catch( NotConnectedException& ) {
