Use TESS_API for every public symbol. Public symbol is exported from …

…the library. This also applies to unit test and training symbols. Users will be limited to public api, but set of exported symbols will be wider still. Remove TESS_LOCAL. Fix several symbol issues that made visible with these changes. All build systems must set -fvisibility-hidden for *nix systems.
karama300 · Dec 31, 2020 · c86325e · c86325e
1 parent 4d817d0
commit c86325e
Show file tree

Hide file tree

Showing 116 changed files with 324 additions and 223 deletions.
diff --git a/include/tesseract/baseapi.h b/include/tesseract/baseapi.h
@@ -736,7 +736,7 @@ class TESS_API TessBaseAPI {
  protected:
   /** Common code for setting the image. Returns true if Init has been called.
    */
-  TESS_LOCAL bool InternalSetImage();
+  bool InternalSetImage();
 
   /**
    * Run the thresholder to make the thresholded image. If pix is not nullptr,
@@ -748,7 +748,7 @@ class TESS_API TessBaseAPI {
    * Find lines from the image making the BLOCK_LIST.
    * @return 0 on success.
    */
-  TESS_LOCAL int FindLines();
+  int FindLines();
 
   /** Delete the pageres and block list ready for a new page. */
   void ClearResults();
@@ -758,20 +758,20 @@ class TESS_API TessBaseAPI {
    * to ignore all BiDi smarts at that point.
    * delete once you're done with it.
    */
-  TESS_LOCAL LTRResultIterator* GetLTRIterator();
+  LTRResultIterator* GetLTRIterator();
 
   /**
    * Return the length of the output text string, as UTF8, assuming
    * one newline per line and one per block, with a terminator,
    * and assuming a single character reject marker for each rejected character.
    * Also return the number of recognized blobs in blob_count.
    */
-  TESS_LOCAL int TextLength(int* blob_count);
+  int TextLength(int* blob_count);
 
   //// paragraphs.cpp ////////////////////////////////////////////////////
-  TESS_LOCAL void DetectParagraphs(bool after_text_recognition);
+  void DetectParagraphs(bool after_text_recognition);
 
-  TESS_LOCAL const PAGE_RES* GetPageRes() const {
+  const PAGE_RES* GetPageRes() const {
     return page_res_;
   }
 

diff --git a/include/tesseract/ltrresultiterator.h b/include/tesseract/ltrresultiterator.h
@@ -183,7 +183,7 @@ class TESS_API LTRResultIterator : public PageIterator {
 };
 
 // Class to iterate over the classifier choices for a single RIL_SYMBOL.
-class ChoiceIterator {
+class TESS_API ChoiceIterator {
  public:
   // Construction is from a LTRResultIterator that points to the symbol of
   // interest. The ChoiceIterator allows a one-shot iteration over the

diff --git a/include/tesseract/pageiterator.h b/include/tesseract/pageiterator.h
@@ -319,7 +319,7 @@ class TESS_API PageIterator {
    * Sets up the internal data for iterating the blobs of a new word, then
    * moves the iterator to the given offset.
    */
-  TESS_LOCAL void BeginWord(int offset);
+  void BeginWord(int offset);
 
   /** Pointer to the page_res owned by the API. */
   PAGE_RES* page_res_;

diff --git a/include/tesseract/platform.h b/include/tesseract/platform.h
@@ -15,31 +15,23 @@
 //
 ///////////////////////////////////////////////////////////////////////
 
-#ifndef TESSERACT_CCUTIL_PLATFORM_H_
-#define TESSERACT_CCUTIL_PLATFORM_H_
+#ifndef TESSERACT_PLATFORM_H_
+#define TESSERACT_PLATFORM_H_
 
-#if defined(_WIN32) || defined(__CYGWIN__)
+#ifndef TESS_API
+# if defined(_WIN32) || defined(__CYGWIN__)
 #  if defined(TESS_EXPORTS)
 #    define TESS_API __declspec(dllexport)
 #  elif defined(TESS_IMPORTS)
 #    define TESS_API __declspec(dllimport)
 #  else
 #    define TESS_API
 #  endif
-#  define TESS_LOCAL
-#else
-#  if __GNUC__ >= 4
-#    if defined(TESS_EXPORTS) || defined(TESS_IMPORTS)
+# else
+#  if defined(TESS_EXPORTS) || defined(TESS_IMPORTS)
 #      define TESS_API __attribute__((visibility("default")))
-#      define TESS_LOCAL __attribute__((visibility("hidden")))
-#    else
-#      define TESS_API
-#      define TESS_LOCAL
-#    endif
-#  else
-#    define TESS_API
-#    define TESS_LOCAL
 #  endif
+# endif
 #endif
 
-#endif  // TESSERACT_CCUTIL_PLATFORM_H_
+#endif  // TESSERACT_PLATFORM_H_
diff --git a/include/tesseract/resultiterator.h b/include/tesseract/resultiterator.h
@@ -144,7 +144,7 @@ class TESS_API ResultIterator : public LTRResultIterator {
    *   it resets to the beginning of the paragraph instead of staying wherever
    *   resit might have pointed.
    */
-  TESS_LOCAL explicit ResultIterator(const LTRResultIterator& resit);
+  explicit ResultIterator(const LTRResultIterator& resit);
 
  private:
   /**

diff --git a/include/tesseract/unichar.h b/include/tesseract/unichar.h
@@ -55,7 +55,7 @@ using char32 = signed int;
 // a single Unicode character (stored as between 1 and 4 utf8 bytes) or
 // multiple Unicode characters representing the NFKC expansion of a ligature
 // such as fi, ffl etc. These are also stored as utf8.
-class UNICHAR {
+class TESS_API UNICHAR {
  public:
   UNICHAR() {
     memset(chars, 0, UNICHAR_LEN);
@@ -105,7 +105,7 @@ class UNICHAR {
   //     int char_len = it.get_utf8(buf); buf[char_len] = '\0';
   //     tprintf("Char = %s\n", buf);
   //   }
-  class const_iterator {
+  class TESS_API const_iterator {
     using CI = const_iterator;
 
    public:

diff --git a/src/arch/intsimdmatrix.h b/src/arch/intsimdmatrix.h
@@ -61,7 +61,7 @@ class GenericVector;
 // NOTE that, although the subclasses execute on different SIMD hardware, no
 // virtual methods are needed, as the constructor sets up everything that
 // is required to allow the base class implementation to do all the work.
-struct IntSimdMatrix {
+struct TESS_API IntSimdMatrix {
   // Computes a reshaped copy of the weight matrix w.
   void Init(const GENERIC_2D_ARRAY<int8_t>& w,
             std::vector<int8_t>& shaped_w,
@@ -115,12 +115,12 @@ struct IntSimdMatrix {
   // Number of groups of inputs to be broadcast.
   // num_input_groups_ = num_inputs_per_register_ / num_inputs_per_group_
 
-  static TESS_API const IntSimdMatrix* intSimdMatrix;
+  static const IntSimdMatrix* intSimdMatrix;
   // Only available with NEON.
-  static TESS_API const IntSimdMatrix intSimdMatrixNEON;
+  static const IntSimdMatrix intSimdMatrixNEON;
   // Only available with AVX2 / SSE.
-  static TESS_API const IntSimdMatrix intSimdMatrixAVX2;
-  static TESS_API const IntSimdMatrix intSimdMatrixSSE;
+  static const IntSimdMatrix intSimdMatrixAVX2;
+  static const IntSimdMatrix intSimdMatrixSSE;
 };
 
 }  // namespace tesseract

diff --git a/src/ccmain/equationdetect.h b/src/ccmain/equationdetect.h
@@ -35,7 +35,7 @@ class ColPartition;
 class ColPartitionGrid;
 class ColPartitionSet;
 
-class EquationDetect : public EquationDetectBase {
+class TESS_API EquationDetect : public EquationDetectBase {
  public:
   EquationDetect(const char* equ_datapath,
                  const char* equ_language);

diff --git a/src/ccmain/mutableiterator.h b/src/ccmain/mutableiterator.h
@@ -40,7 +40,7 @@ class Tesseract;
 // ResultIterator adds text-specific methods for access to OCR output.
 // MutableIterator adds access to internal data structures.
 
-class MutableIterator : public ResultIterator {
+class TESS_API MutableIterator : public ResultIterator {
  public:
   // See argument descriptions in ResultIterator()
   MutableIterator(PAGE_RES* page_res, Tesseract* tesseract,

diff --git a/src/ccmain/paragraphs.h b/src/ccmain/paragraphs.h
@@ -87,6 +87,7 @@ class RowInfo {
 //   paragraphs - this is the actual list of PARA objects.
 //   models - the list of paragraph models referenced by the PARA objects.
 //            caller is responsible for deleting the models.
+TESS_API
 void DetectParagraphs(int debug_level,
                       std::vector<RowInfo> *row_infos,
                       GenericVector<PARA *> *row_owners,
@@ -98,6 +99,7 @@ void DetectParagraphs(int debug_level,
 // saving the ParagraphModels in models.  Caller owns the models.
 // We use unicharset during the function to answer questions such as "is the
 // first letter of this word upper case?"
+TESS_API
 void DetectParagraphs(int debug_level,
                       bool after_text_recognition,
                       const MutableIterator *block_start,

diff --git a/src/ccmain/paragraphs_internal.h b/src/ccmain/paragraphs_internal.h
@@ -31,18 +31,21 @@ class UNICHARSET;
 class WERD_CHOICE;
 
 // Return whether the given word is likely to be a list item start word.
+TESS_API
 bool AsciiLikelyListItem(const STRING &word);
 
 // Return the first Unicode Codepoint from werd[pos].
 int UnicodeFor(const UNICHARSET *u, const WERD_CHOICE *werd, int pos);
 
 // Set right word attributes given either a unicharset and werd or a utf8
 // string.
+TESS_API
 void RightWordAttributes(const UNICHARSET *unicharset, const WERD_CHOICE *werd,
                          const STRING &utf8,
                          bool *is_list, bool *starts_idea, bool *ends_idea);
 
 // Set left word attributes given either a unicharset and werd or a utf8 string.
+TESS_API
 void LeftWordAttributes(const UNICHARSET *unicharset, const WERD_CHOICE *werd,
                         const STRING &utf8,
                         bool *is_list, bool *starts_idea, bool *ends_idea);

diff --git a/src/ccmain/tesseractclass.h b/src/ccmain/tesseractclass.h
@@ -175,7 +175,7 @@ struct WordData {
 using WordRecognizer = void (Tesseract::*)(const WordData&, WERD_RES**,
                                            PointerVector<WERD_RES>*);
 
-class Tesseract : public Wordrec {
+class TESS_API Tesseract : public Wordrec {
  public:
   Tesseract();
   ~Tesseract() override;

diff --git a/src/ccstruct/boxread.h b/src/ccstruct/boxread.h
@@ -32,6 +32,7 @@ const int kBoxReadBufSize = 1024;
 
 // Open the boxfile based on the given image filename.
 // Returns nullptr if the box file cannot be opened.
+TESS_API
 FILE* OpenBoxFile(const char* filename);
 
 // Reads all boxes from the given filename.
@@ -51,6 +52,7 @@ bool ReadAllBoxes(int target_page, bool skip_blanks, const char* filename,
 // continue_on_failure allows reading to continue even if an invalid box is
 // encountered and will return true if it succeeds in reading some boxes.
 // It otherwise gives up and returns false on encountering an invalid box.
+TESS_API
 bool ReadMemBoxes(int target_page, bool skip_blanks, const char* box_data,
                   bool continue_on_failure,
                   std::vector<TBOX>* boxes,
@@ -66,20 +68,24 @@ bool ReadMemBoxes(int target_page, bool skip_blanks, const char* box_data,
 // for valid utf-8 and allows space or tab between fields.
 // utf8_str is set with the unichar string, and bounding box with the box.
 // If there are page numbers in the file, it reads them all.
+TESS_API
 bool ReadNextBox(int *line_number, FILE* box_file,
                  STRING* utf8_str, TBOX* bounding_box);
 // As ReadNextBox above, but get a specific page number. (0-based)
 // Use -1 to read any page number. Files without page number all
 // read as if they are page 0.
+TESS_API
 bool ReadNextBox(int target_page, int *line_number, FILE* box_file,
                  STRING* utf8_str, TBOX* bounding_box);
 
 // Parses the given box file string into a page_number, utf8_str, and
 // bounding_box. Returns true on a successful parse.
+TESS_API
 bool ParseBoxFileStr(const char* boxfile_str, int* page_number,
                      STRING* utf8_str, TBOX* bounding_box);
 
 // Creates a box file string from a unichar string, TBOX and page number.
+TESS_API
 void MakeBoxFileStr(const char* unichar_str, const TBOX& box, int page_num,
                     STRING* box_str);
 

diff --git a/src/ccstruct/ccstruct.h b/src/ccstruct/ccstruct.h
@@ -22,7 +22,7 @@
 #include "ccutil.h" // for CCUtil
 
 namespace tesseract {
-class CCStruct : public CCUtil {
+class TESS_API CCStruct : public CCUtil {
  public:
   CCStruct() = default;
   ~CCStruct() override;

diff --git a/src/ccstruct/fontinfo.h b/src/ccstruct/fontinfo.h
@@ -146,26 +146,34 @@ struct FontSet {
 // are replaced.
 class FontInfoTable : public GenericVector<FontInfo> {
  public:
+  TESS_API // when you remove inheritance from GenericVector, move this on class level
   FontInfoTable();
+  TESS_API
   ~FontInfoTable();
 
   // Writes to the given file. Returns false in case of error.
+  TESS_API
   bool Serialize(FILE* fp) const;
   // Reads from the given file. Returns false in case of error.
   // If swap is true, assumes a big/little-endian swap is needed.
+  TESS_API
   bool DeSerialize(TFile* fp);
 
   // Returns true if the given set of fonts includes one with the same
   // properties as font_id.
+  TESS_API
   bool SetContainsFontProperties(
       int font_id, const GenericVector<ScoredFont>& font_set) const;
   // Returns true if the given set of fonts includes multiple properties.
+  TESS_API
   bool SetContainsMultipleFontProperties(
       const GenericVector<ScoredFont>& font_set) const;
 
   // Moves any non-empty FontSpacingInfo entries from other to this.
+  TESS_API
   void MoveSpacingInfoFrom(FontInfoTable* other);
   // Moves this to the target unicity table.
+  TESS_API
   void MoveTo(UnicityTable<FontInfo>* target);
 };
 

diff --git a/src/ccstruct/imagedata.h b/src/ccstruct/imagedata.h
@@ -104,7 +104,7 @@ struct FloatWordFeature {
 // The text transcription is the ground truth UTF-8 text for the image.
 // Character boxes are optional and indicate the desired segmentation of
 // the text into recognition units.
-class ImageData {
+class TESS_API ImageData {
  public:
   ImageData();
   // Takes ownership of the pix.
@@ -213,19 +213,24 @@ class ImageData {
 // A collection of ImageData that knows roughly how much memory it is using.
 class DocumentData {
  public:
+  TESS_API
   explicit DocumentData(const STRING& name);
+  TESS_API
   ~DocumentData();
 
   // Reads all the pages in the given lstmf filename to the cache. The reader
   // is used to read the file.
+  TESS_API
   bool LoadDocument(const char* filename, int start_page, int64_t max_memory,
                     FileReader reader);
   // Sets up the document, without actually loading it.
   void SetDocument(const char* filename, int64_t max_memory, FileReader reader);
   // Writes all the pages to the given filename. Returns false on error.
+  TESS_API
   bool SaveDocument(const char* filename, FileWriter writer);
 
   // Adds the given page data to this document, counting up memory.
+  TESS_API
   void AddPageToDocument(ImageData* page);
 
   const STRING& document_name() const {
@@ -257,6 +262,7 @@ class DocumentData {
   void LoadPageInBackground(int index);
   // Returns a pointer to the page with the given index, modulo the total
   // number of pages. Blocks until the background load is completed.
+  TESS_API
   const ImageData* GetPage(int index);
   // Returns true if the requested page is available, and provides a pointer,
   // which may be nullptr if the document is empty. May block, even though it
@@ -325,7 +331,9 @@ class DocumentData {
 // content.
 class DocumentCache {
  public:
+  TESS_API
   explicit DocumentCache(int64_t max_memory);
+  TESS_API
   ~DocumentCache();
 
   // Deletes all existing documents from the cache.
@@ -335,6 +343,7 @@ class DocumentCache {
   }
   // Adds all the documents in the list of filenames, counting memory.
   // The reader is used to read the files.
+  TESS_API
   bool LoadDocuments(const std::vector<STRING>& filenames,
                      CachingStrategy cache_strategy, FileReader reader);
 
@@ -358,16 +367,19 @@ class DocumentCache {
   }
   // Returns the total number of pages in an epoch. For CS_ROUND_ROBIN cache
   // strategy, could take a long time.
+  TESS_API
   int TotalPages();
 
  private:
   // Returns a page by serial number, selecting them in a round-robin fashion
   // from all the documents. Highly disk-intensive, but doesn't need samples
   // to be shuffled between files to begin with.
+  TESS_API
   const ImageData* GetPageRoundRobin(int serial);
   // Returns a page by serial number, selecting them in sequence from each file.
   // Requires the samples to be shuffled between the files to give a random or
   // uniform distribution of data. Less disk-intensive than GetPageRoundRobin.
+  TESS_API
   const ImageData* GetPageSequential(int serial);
 
   // Helper counts the number of adjacent cached neighbour documents_ of index

diff --git a/src/ccstruct/linlsq.h b/src/ccstruct/linlsq.h
@@ -28,7 +28,7 @@ namespace tesseract {
 
 template <typename T> class GenericVector;
 
-class LLSQ {
+class TESS_API LLSQ {
  public:
   LLSQ() {  // constructor
     clear();  // set to zeros