mirror/tesseract
1
0
Fork 0
mirror of https://github.com/tesseract-ocr/tesseract.git synced 2024-11-23 18:49:08 +08:00
Code Issues Actions 10 Packages Projects Releases Wiki Activity

Remove blanks at line endings

Browse Source 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
This commit is contained in:
Stefan Weil 2021-02-26 10:56:42 +01:00
parent 394c56ab15
commit ea446b1eae
13 changed files with 79 additions and 80 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
.github/workflows/autotools.yml vendored
View File

@ -172,12 +172,12 @@ jobs:
if: startsWith(matrix.config.cxx, 'clang')
run: |
make check
- name: Make and run Unit Tests (unset LANG needed for g++-8, g++-9, g++-10 on macOS)
if: startsWith(matrix.config.cxx, 'g')
shell: bash
run: |
unset LANG LC_ALL LC_CTYPE
unset LANG LC_ALL LC_CTYPE
locale
make check
@ -220,7 +220,7 @@ jobs:
run: |
cat test-suite.log
if: always()
# ============================================================================================
ports:
@ -299,7 +299,7 @@ jobs:
tesseract test/testing/hebrew.png - -l heb --oem 1 --tessdata-dir ../tessdata
tesseract test/testing/eurotext.tif - -l fra --oem 1 --tessdata-dir ../tessdata_best
tesseract test/testing/arabic.tif - -l ara --oem 1 --psm 6 --tessdata-dir ../tessdata
- name: Run Tesseract basicapitest
run: |
export "PKG_CONFIG_PATH=/usr/local/lib/pkgconfig"
@ -317,4 +317,4 @@ jobs:
run: |
cat test-suite.log
if: always()

42
.github/workflows/cmake.yml vendored
View File

@ -14,40 +14,40 @@ jobs:
fail-fast: false
matrix:
config:
- { name: macos-10.15-clang-12-cmake, os: macos-10.15, cxx: clang++ } # defualt
- { name: macos-10.15-clang-11-cmake, os: macos-10.15, cxx: '$(brew --prefix llvm)/bin/clang++' } #installed
- { name: macos-10.15-gcc-8-cmake, os: macos-10.15, cxx: g++-8 } #installed
- { name: macos-10.15-gcc-9-cmake, os: macos-10.15, cxx: g++-9 } #installed
- { name: macos-10.15-gcc-10-cmake, os: macos-10.15, cxx: g++-10 } #installed
- { name: ubuntu-18.04-clang-7-cmake, os: ubuntu-18.04, cxx: clang++-7 }
- { name: ubuntu-18.04-clang-8-cmake, os: ubuntu-18.04, cxx: clang++-8 } #installed
- { name: ubuntu-18.04-clang-9-cmake, os: ubuntu-18.04, cxx: clang++-9 } #installed
- { name: ubuntu-18.04-clang-10-cmake, os: ubuntu-18.04, cxx: clang++-10 }
- { name: ubuntu-18.04-gcc-7-cmake, os: ubuntu-18.04, cxx: g++-7 } #installed
- { name: ubuntu-18.04-gcc-8-cmake, os: ubuntu-18.04, cxx: g++-8 } #installed
- { name: ubuntu-18.04-gcc-9-cmake, os: ubuntu-18.04, cxx: g++-9 } #installed
- { name: ubuntu-18.04-gcc-10-cmake, os: ubuntu-18.04, cxx: g++-10 } #installed
- { name: ubuntu-20.04-clang-7-cmake, os: ubuntu-20.04, cxx: clang++-7 }
- { name: ubuntu-20.04-clang-8-cmake, os: ubuntu-20.04, cxx: clang++-8 } #installed
- { name: ubuntu-20.04-clang-9-cmake, os: ubuntu-20.04, cxx: clang++-9 } #installed
- { name: ubuntu-20.04-clang-10-cmake, os: ubuntu-20.04, cxx: clang++-10 } #installed
- { name: ubuntu-20.04-gcc-7-cmake, os: ubuntu-20.04, cxx: g++-7 } #installed
- { name: ubuntu-20.04-gcc-8-cmake, os: ubuntu-20.04, cxx: g++-8 } #installed
- { name: ubuntu-20.04-gcc-9-cmake, os: ubuntu-20.04, cxx: g++-9 } #installed
- { name: ubuntu-20.04-gcc-10-cmake, os: ubuntu-20.04, cxx: g++-10 } #installed
steps:
- name: Install compilers on Linux
run: |
sudo apt-get install ${{ matrix.config.cxx }} -y
if: runner.os == 'Linux'
# sudo apt-get install libarchive-dev libcurl4-openssl-dev libcurl4 curl -y
# sudo apt-get install libarchive-dev libcurl4-openssl-dev libcurl4 curl -y
- name: Install dependencies on Linux
run: |
sudo apt-get install autoconf-archive libleptonica-dev -y
@ -66,12 +66,12 @@ jobs:
ninja --version
cmake --version
if: runner.os == 'macOS'
- name: Checkout Source
uses: actions/checkout@v2
with:
submodules: recursive
- name: Configure Tesseract (Linux)
run: |
mkdir build
@ -102,32 +102,32 @@ jobs:
-DCMAKE_CXX_COMPILER=${{ matrix.config.cxx }} \
-DCMAKE_INSTALL_PREFIX:PATH=inst
if: runner.os == 'macOS'
- name: Build Tesseract
run: |
cmake --build build --config Release --target install
- name: Display Tesseract Version
run: |
build/inst/bin/tesseract -v
- name: Display Training Tools Version
run: |
build/inst/bin/lstmtraining -v
build/inst/bin/text2image -v
- name: Download fonts, tessdata and langdata required for tests
run: |
git clone https://github.com/egorpugin/tessdata tessdata_unittest
cp tessdata_unittest/fonts/* test/testing/
cp tessdata_unittest/fonts/* test/testing/
mv tessdata_unittest/* ../
- name: List languages in different tessdata-dir
run: |
build/inst/bin/tesseract --list-langs --tessdata-dir ../tessdata
build/inst/bin/tesseract --list-langs --tessdata-dir ../tessdata_best
build/inst/bin/tesseract --list-langs --tessdata-dir ../tessdata_fast
- name: Run Tesseract on test images in different languages
run: |
build/inst/bin/tesseract test/testing/phototest.tif - --oem 1 --tessdata-dir ../tessdata
@ -136,17 +136,17 @@ jobs:
build/inst/bin/tesseract test/testing/hebrew.png - -l heb --oem 1 --tessdata-dir ../tessdata
build/inst/bin/tesseract test/testing/eurotext.tif - -l fra --oem 1 --tessdata-dir ../tessdata_best
build/inst/bin/tesseract test/testing/arabic.tif - -l ara --oem 1 --psm 6 --tessdata-dir ../tessdata
- name: Build and run basicapitest
run: |
export "PKG_CONFIG_PATH=$GITHUB_WORKSPACE/build/inst/lib/pkgconfig/:$PKG_CONFIG_PATH"
cd test
${{ matrix.config.cxx }} -o basicapitest testing/basicapitest.cpp "-I$GITHUB_WORKSPACE/build/inst/include" "-L$GITHUB_WORKSPACE/build/inst/lib" `pkg-config --cflags --libs tesseract lept ` -pthread -std=c++11
${{ matrix.config.cxx }} -o basicapitest testing/basicapitest.cpp "-I$GITHUB_WORKSPACE/build/inst/include" "-L$GITHUB_WORKSPACE/build/inst/lib" `pkg-config --cflags --libs tesseract lept ` -pthread -std=c++11
./basicapitest
- name: Display Compiler Version
run: |
${{ matrix.config.cxx }} --version
git log -3 --pretty=format:'%h %ad %s | %an'
git log -3 --pretty=format:'%h %ad %s | %an'
if: always()

10
.github/workflows/sw.yml vendored
View File

@ -15,7 +15,7 @@ jobs:
matrix:
os: [windows-latest, ubuntu-20.04, macOS-latest]
steps:
steps:
- uses: actions/checkout@v2
with:
submodules: recursive
@ -31,21 +31,21 @@ jobs:
- name: download test data
run: git clone https://github.com/egorpugin/tessdata tessdata_unittest
- name: copy fonts
if: matrix.os != 'windows-latest'
run: cp tessdata_unittest/fonts/* test/testing/
- name: copy fonts
if: matrix.os == 'windows-latest'
run: Copy-Item -Path "tessdata_unittest\fonts\*" -Destination "test\testing" -Recurse
shell: pwsh
- name: test
if: matrix.os != 'windows-latest'
run: ./sw -static -shared -config "d,r" test -Dwith-tests=1 "-Dskip-tests=lstm,lstm_recode"
continue-on-error: true
- name: test-nightly
if: matrix.os != 'windows-latest' && github.event.schedule=='0 0 * * *'
run: ./sw -static -shared -config "d,r" test -Dwith-tests=1

16
.github/workflows/unittest-disablelegacy.yml vendored
View File

@ -20,7 +20,7 @@ jobs:
- uses: actions/checkout@v2
with:
submodules: recursive
- name: Install dependencies
run: |
sudo apt-get install autoconf-archive libleptonica-dev libpango1.0-dev -y
@ -34,7 +34,7 @@ jobs:
- name: Configure
run: |
./configure '--disable-shared' '--disable-legacy' 'CXX=${{ matrix.compiler }}'
- name: Make and Install Tesseract
run: |
make -j 8
@ -44,7 +44,7 @@ jobs:
run: |
make training -j 8
sudo make install training-install
- name: Display Version
run: |
${{ matrix.compiler }} --version
@ -52,13 +52,13 @@ jobs:
lstmtraining -v
text2image -v
if: success() || failure()
- name: Download fonts, tessdata and langdata required for tests
run: |
git clone https://github.com/egorpugin/tessdata tessdata_unittest
cp tessdata_unittest/fonts/* test/testing/
cp tessdata_unittest/fonts/* test/testing/
mv tessdata_unittest/* ../
- name: Run Tesseract on phototest.tif and devatest.png
run: |
tesseract test/testing/phototest.tif - --tessdata-dir ../tessdata
@ -67,10 +67,10 @@ jobs:
- name: Make and run Unit Tests
run: |
make check -j 4
- name: Display Unit Tests Report
run: |
git log -3
${{ matrix.compiler }} --version
${{ matrix.compiler }} --version
cat test-suite.log
if: always()

14
.github/workflows/unittest.yml vendored
View File

@ -38,7 +38,7 @@ jobs:
brew install leptonica cairo pango
brew install cabextract abseil
if: runner.os == 'macOS'
- name: Setup
run: |
mkdir -p m4
@ -68,7 +68,7 @@ jobs:
run: |
make training -j 8
sudo make training-install
- name: Display Version
run: |
${{ matrix.config.cxx }} --version
@ -76,13 +76,13 @@ jobs:
lstmtraining -v
text2image -v
if: success() || failure()
- name: Download fonts, tessdata and langdata required for tests
run: |
git clone https://github.com/egorpugin/tessdata tessdata_unittest
cp tessdata_unittest/fonts/* test/testing/
cp tessdata_unittest/fonts/* test/testing/
mv tessdata_unittest/* ../
- name: Run Tesseract on phototest.tif and devatest.png
run: |
tesseract test/testing/phototest.tif - --tessdata-dir ../tessdata
@ -91,10 +91,10 @@ jobs:
- name: Make and run Unit Tests
run: |
make check -j 4
- name: Display Unit Tests Report
run: |
cat test-suite.log
${{ matrix.config.cxx }} --version
git log -3 --pretty=format:'%h %ad %s | %an'
git log -3 --pretty=format:'%h %ad %s | %an'
if: always()

26
.github/workflows/vcpkg.yml vendored
View File

@ -4,26 +4,26 @@ on:
#push:
schedule:
- cron: 0 23 * * *
jobs:
jobs:
build:
runs-on: ${{ matrix.os }}
strategy:
fail-fast: false
matrix:
os: [windows-2019]
steps:
- name: Checkout Tesseract Source (--head from master branch)
uses: actions/checkout@v2
with:
submodules: recursive
- name: Visual Studio Setup
- name: Visual Studio Setup
shell: cmd
run: |
call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvars64.bat"
- name: Install vcpkg
run: |
git clone https://github.com/microsoft/vcpkg
@ -33,13 +33,13 @@ jobs:
- name: Build and Install Leptonica and image libraries using vcpkg
run: |
vcpkg/vcpkg install leptonica:x64-windows
- name: Configure and Build Tesseract (--head from master branch) with cmake
run: |
cmake . -B build -DCMAKE_BUILD_TYPE=Release -DSW_BUILD=OFF -DOPENMP_BUILD=OFF -DBUILD_TRAINING_TOOLS=OFF "-DCMAKE_TOOLCHAIN_FILE=${env:GITHUB_WORKSPACE}/vcpkg/scripts/buildsystems/vcpkg.cmake"
cmake . -B build -DCMAKE_BUILD_TYPE=Release -DSW_BUILD=OFF -DOPENMP_BUILD=OFF -DBUILD_TRAINING_TOOLS=OFF "-DCMAKE_TOOLCHAIN_FILE=${env:GITHUB_WORKSPACE}/vcpkg/scripts/buildsystems/vcpkg.cmake"
cmake --build build --config Release --target install
- name: Display Tesseract Version
- name: Display Tesseract Version
run: |
D:\a\tesseract\tesseract\build\bin\Release\tesseract.exe --version
@ -62,17 +62,17 @@ jobs:
target_link_libraries(basicapitest libtesseract)
EOF
cat CMakeLists.txt
- name: Configure basicapitest
run: |
cd test
cmake . "-DCMAKE_TOOLCHAIN_FILE=${env:GITHUB_WORKSPACE}/vcpkg/scripts/buildsystems/vcpkg.cmake"
- name: Build basicapitest
run: |
cd test
cmake --build . --config Release
- name: Download tessdata and image files used for tests
run: |
git clone https://github.com/egorpugin/tessdata tessdata_unittest
@ -91,7 +91,7 @@ jobs:
D:\a\tesseract\tesseract\build\bin\Release\tesseract.exe test\testing\hebrew.png - -l heb --oem 1 --tessdata-dir ..\tessdata
D:\a\tesseract\tesseract\build\bin\Release\tesseract.exe test\testing\eurotext.tif - -l fra --oem 1 --tessdata-dir ..\tessdata_best
D:\a\tesseract\tesseract\build\bin\Release\tesseract.exe test\testing\arabic.tif - -l ara --oem 1 --psm 6 --tessdata-dir ..\tessdata
- name: List languages in different test tessdata-dir
run: |
D:\a\tesseract\tesseract\build\bin\Release\tesseract.exe --list-langs --tessdata-dir ..\tessdata

5
.travis.yml
View File

@ -17,7 +17,7 @@ matrix:
ccache: true
directories:
- $HOME/Library/Caches/Homebrew
before_install:
before_install:
- brew install ccache
install:
- brew install ninja libtiff leptonica libarchive pango cairo icu4c
@ -33,7 +33,7 @@ matrix:
-G Ninja \
-DCMAKE_INSTALL_PREFIX:PATH=inst
cmake --build build --config Release --target install
- os: linux
dist: bionic
addons:
@ -57,4 +57,3 @@ matrix:
- cmake .. -DLeptonica_DIR=leptonica-$LEPT_VER/build -DSW_BUILD=OFF
- make
- sudo make install

6
doc/tesseract.natvis
View File

@ -1,4 +1,4 @@
<?xml version="1.0" encoding="utf-8"?>
<?xml version="1.0" encoding="utf-8"?>
<AutoVisualizer xmlns="http://schemas.microsoft.com/vstudio/debugger/natvis/2010">
<Type Name="STRING">
<DisplayString>{(char*)data_+sizeof(int)*2,s8}</DisplayString>
@ -15,7 +15,7 @@
</ArrayItems>
</Expand>
</Type>
<Type Name="tesseract::IntParam">
<DisplayString>{value_}</DisplayString>
</Type>
@ -30,5 +30,5 @@
<Type Name="tesseract::DoubleParam">
<DisplayString>{value_}</DisplayString>
</Type>
</AutoVisualizer>

2
src/api/hocrrenderer.cpp
View File

@ -403,7 +403,7 @@ char* TessBaseAPI::GetHOCRText(ETEXT_DESC* monitor, int page_number) {
tcnt++;
}
}
}
}
// Close ocrx_word.
if (hocr_boxes || lstm_choice_mode > 0) {
hocr_str << "\n ";

8
src/api/tesseractmain.cpp
View File

@ -76,13 +76,13 @@ static void Win32WarningHandler(const char* module, const char* fmt,
class AutoWin32ConsoleOutputCP {
public:
explicit AutoWin32ConsoleOutputCP(UINT codeCP) {
oldCP_ = GetConsoleOutputCP();
oldCP_ = GetConsoleOutputCP();
SetConsoleOutputCP(codeCP);
}
~AutoWin32ConsoleOutputCP() {
SetConsoleOutputCP(oldCP_);
~AutoWin32ConsoleOutputCP() {
SetConsoleOutputCP(oldCP_);
}
private:
private:
UINT oldCP_;
};

2
src/ccmain/ltrresultiterator.cpp
View File

@ -385,7 +385,7 @@ ChoiceIterator::ChoiceIterator(const LTRResultIterator& result_it) {
bool lstm_choice_mode = word_res_->tesseract->lstm_choice_mode;
rating_coefficient_ = word_res_->tesseract->lstm_rating_coefficient;
blanks_before_word_ = result_it.BlanksBeforeWord();
BLOB_CHOICE_LIST* choices = nullptr;
BLOB_CHOICE_LIST* choices = nullptr;
tstep_index_ = &result_it.blob_index_;
if (oemLSTM_ && !word_res_->CTC_symbol_choices.empty()) {
if (!word_res_->CTC_symbol_choices[0].empty() &&

10
unittest/third_party/utf/rune.c vendored
View File

@ -23,7 +23,7 @@ enum
Bit2 = 5,
Bit3 = 4,
Bit4 = 3,
Bit5 = 2,
Bit5 = 2,
T1 = ((1<<(Bit1+1))-1) ^ 0xFF, /* 0000 0000 */
Tx = ((1<<(Bitx+1))-1) ^ 0xFF, /* 1000 0000 */
@ -46,9 +46,9 @@ enum
/*
* Modified by Wei-Hwa Huang, Google Inc., on 2004-09-24
* This is a slower but "safe" version of the old chartorune
* This is a slower but "safe" version of the old chartorune
* that works on strings that are not necessarily null-terminated.
*
*
* If you know for sure that your string is null-terminated,
* chartorune will be a bit faster.
*
@ -162,7 +162,7 @@ badlen:
/*
* This is the older "unsafe" version, which works fine on
* This is the older "unsafe" version, which works fine on
* null-terminated strings.
*/
int
@ -248,7 +248,7 @@ isvalidcharntorune(const char* str, int length, Rune* rune, int* consumed) {
*consumed = charntorune(rune, str, length);
return *rune != Runeerror || *consumed == 3;
}
int
runetochar(char *str, const Rune *rune)
{

8
unittest/third_party/utf/utf.h vendored
View File

@ -66,7 +66,7 @@ int chartorune(Rune* r, const char* s);
// n bytes of s. If the UTF sequence is incomplete within n bytes,
// charntorune will set *r to Runeerror and return 0. If it is complete
// but not in UTF format, it will set *r to Runeerror and return 1.
//
//
// Added 2004-09-24 by Wei-Hwa Huang
int charntorune(Rune* r, const char* s, int n);
@ -150,7 +150,7 @@ char* utfecpy(char *s1, char *es1, const char *s2);
// These functions are rune-string analogues of the corresponding
// functions in strcat (3).
//
//
// These routines first appeared in Plan 9.
// SEE ALSO
// memmove (3)
@ -203,8 +203,8 @@ Rune totitlerune(Rune r);
// isupperrune tests for upper case characters, including Unicode
// upper case letters and targets of the toupper mapping. islowerrune
// and istitlerune are defined analogously.
// and istitlerune are defined analogously.
int isupperrune(Rune r);
int islowerrune(Rune r);
int istitlerune(Rune r);