Improve code snapshot generation for PDF docs (#1353)

CMake: execute `code_snapshot` separately for each file we want to
  scan so we can create timestamps and generate PNG images only if
  the source file (e.g. unicode.dox) was modified.

documentation/src/unicode.dox: remove output folder `generated`, use
  only the filename. CMake sets the current working directory as needed.

util/code_snapshot.cxx: format and improve comments, reset code buffer
  so we can generate multiple images per input file.

3 files changed, 66 insertions, 24 deletions

diff --git a/documentation/CMakeLists.txt b/documentation/CMakeLists.txt
index 6c8524ffe..ada291af2 100644
--- a/documentation/CMakeLists.txt
+++ b/documentation/CMakeLists.txt
@@ -202,11 +202,42 @@ if(FLTK_BUILD_PDF_DOCS)
     @ONLY
   )
 
-  # generate fltk.pdf
+  # Generate code snapshots (images with international characters).
+  # Note: File names (even from different folders) must be unique !
+
+  set(image_output_dir "${CMAKE_CURRENT_BINARY_DIR}/generated")
+  set(_deps)
+  set(image_input_sources
+      ${CMAKE_CURRENT_SOURCE_DIR}/src/unicode.dox
+      # ${CMAKE_CURRENT_SOURCE_DIR}/src/cmp.dox
+      # ${FLTK_SOURCE_DIR}/src/Fl.cxx # example source file
+  )
+
+  # Generate one custom command and one timestamp per input file
+  # so dependencies can be used to generate the images only if
+  # the source file was changes.
+  # Variable `_deps` is used to make the final PDF generation
+  # depend on all input files.
+
+  foreach(_infile ${image_input_sources})
+    get_filename_component(_name "${_infile}" NAME)
+    set(_timestamp ${image_output_dir}/${_name}.timestamp)
+    list(APPEND _deps ${_timestamp})
+    # create custom command
+    add_custom_command(
+      OUTPUT  ${_timestamp}
+      COMMAND touch ${_timestamp}
+      COMMAND code_snapshot ${_infile}
+      DEPENDS ${_infile}
+      WORKING_DIRECTORY ${image_output_dir}
+      COMMENT "Generating code snapshots (PNG) from '${_infile}'"
+    )
+  endforeach()
+
+  # Now generate the PDF file (fltk.pdf)
 
   add_custom_command(
     OUTPUT  ${CMAKE_CURRENT_BINARY_DIR}/fltk.pdf
-    COMMAND code_snapshot ${CMAKE_CURRENT_SOURCE_DIR}/src/unicode.dox
     COMMAND ${CMAKE_CURRENT_SOURCE_DIR}/make_header
             ${DOXYGEN_EXECUTABLE}
             ${CMAKE_CURRENT_BINARY_DIR}/fltk-title.tex
@@ -214,8 +245,11 @@ if(FLTK_BUILD_PDF_DOCS)
     COMMAND ${DOXYGEN_EXECUTABLE} ${DOXYFILE}
     COMMAND ${CMAKE_CURRENT_BINARY_DIR}/make_pdf
     COMMAND cp -f latex/refman.pdf fltk.pdf
-    DEPENDS ${CMAKE_CURRENT_BINARY_DIR}/${DOXYFILE}
+    DEPENDS ${CMAKE_CURRENT_SOURCE_DIR}/Doxyfile.in
+            ${CMAKE_CURRENT_BINARY_DIR}/${DOXYFILE}.in
+            ${CMAKE_CURRENT_BINARY_DIR}/${DOXYFILE}
             ${CMAKE_CURRENT_BINARY_DIR}/fltk-title.tex
+            ${_deps} # source files processed to generate PNG's
     WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}
     COMMENT "Generating PDF documentation" VERBATIM
   )
diff --git a/documentation/src/unicode.dox b/documentation/src/unicode.dox
index b640e170e..c11308ca3 100644
--- a/documentation/src/unicode.dox
+++ b/documentation/src/unicode.dox
@@ -22,7 +22,7 @@ are not yet implemented.
 
 International code test for HTML and PDF:
 
-\code_international{"generated/unicode_about.png"}
+\code_international{"unicode_about.png"}
 // This is a test
 // 日本語テストテキスト
 // 中文测试文本
@@ -70,11 +70,11 @@ doesn't break legacy ASCII code.
 Functions to check and analyze UTF-8 text:
 
 fl_utf8test() - Check if a string contains valid UTF-8
-\code
+\code_international{"unicode_hello.png"}
 const char* text = "Hello 世界";
 int result = fl_utf8test(text, strlen(text));
 // Returns: 0=invalid, 1=ASCII, 2=2-byte chars, 3=3-byte chars, 4=4-byte chars
-\endcode
+\endcode_international
 
 fl_utf8len() - Get the length in bytes of a UTF-8 octet
 \code
diff --git a/util/code_snapshot.cxx b/util/code_snapshot.cxx
index 6613d5b51..907c6ae37 100644
--- a/util/code_snapshot.cxx
+++ b/util/code_snapshot.cxx
@@ -18,11 +18,20 @@
 //
 // Our documentation for the FLTK unicode contains international characters
 // to illustrate use of non ASCII characters in the GUI. To generate PDF
-// output, Doxygen uses LaTeX which can not easily handle UTF-8 characters in
+// output, Doxygen uses LaTeX which can not easily handle UTF-8 characters
 // beyond Western encoding. This tool generates PNG images from code segments
 // containing international characters so that they can be included in the
 // PDF documentation instead of the code segments with UTF-8 characters.
 //
+// Notes:
+// - This program is work in progress...
+// - The PDF generation process (CMake) calls it once for each source file
+//   that needs image generation. The loop over all commandline arguments is
+//   currently not necessary but kept for potential extensions.
+// - The program exits silently if no commandline is given.
+// - If the terminating "\endcode_international" line is missing or misspelled,
+//   the program reads the "code" until the end of the file is reached and
+//   terminates w/o error message. This could be improved...
 
 #include <FL/Fl_Window.H>
 #include <FL/Fl_Group.H>
@@ -54,7 +63,7 @@ void create_window() {
   code_viewer->box(FL_FLAT_BOX);
   code_viewer->color(0xf7f7ff00);
   code_viewer->textsize(30);
-  //code_viewer->cursor_style(CARET_CURSOR);
+  // code_viewer->cursor_style(CARET_CURSOR);
 
   window->resizable(group);
   group->resizable(code_viewer);
@@ -69,14 +78,13 @@ void create_window() {
   line_height = std::max(line_height, fl_height(FL_COURIER_BOLD_ITALIC, code_viewer->textsize()));
 }
 
-void save_snapshot(const char* code, const char* filename)
-{
-//  fprintf(stderr, "\\code\n%s\n\\endcode\n", code);
+void save_snapshot(const char* code, const char* filename) {
+  // fprintf(stderr, "\\code\n%s\n\\endcode\n", code);
 
   code_viewer->buffer()->text(code);
   int n_lines = 1;
   for (const char* s=code; *s; ++s) if (*s == '\n') n_lines++;
-  // 300 dpi for 7 inches = 2000 pixels
+  // 300 dpi for 7 inches = 2100 pixels
   window->size(2100, (line_height * n_lines) + 18 );
 
   // Generate the Image Surface
@@ -88,7 +96,7 @@ void save_snapshot(const char* code, const char* filename)
   fl_rect(0, 0, window->w(), window->h(), 0xccccff00);
   Fl_Image_Surface::pop_current();
 
-//  fprintf(stderr, "  Saving to \"%s\".\n", filename);
+  // fprintf(stderr, "  Saving to \"%s\".\n", filename);
 
   // Write the generated image
   Fl_RGB_Image *img = srfc->image();
@@ -102,25 +110,24 @@ void save_snapshot(const char* code, const char* filename)
 /**
   Main entry point for the PDF documentation helper tool.
 
-  The app scans the input file for the `\\code_international{"filename"}`
+  The app scans the input file for the `\code_international{"filename"}`
   directive, reads the following code segment until
-  `\\endcode_international`, and generates a PNG image file with the given
+  `\endcode_international`, and generates a PNG image file with the given
   filename containing the code segment rendered with FLTK's
   code rendering capabilities.
 
-  \param argc Argument count
-  \param argv a list of input files with documentation in Doxygen format
+  \param[in] argc Argument count
+  \param[in] argv a list of input files with documentation in Doxygen format
   \return Exit code (0 for success, non-zero for failure)
 */
-int main(int argc, char *argv[])
-{
+int main(int argc, char *argv[]) {
   int ret = 0;
   char line[1024];
   char cwd[FL_PATH_MAX];
 
-//  fl_getcwd(cwd, FL_PATH_MAX-1);
-//  fprintf(stderr, "code_snapshot:\n");
-//  fprintf(stderr, "Working directory is \"%s\".\n", cwd);
+  // fl_getcwd(cwd, FL_PATH_MAX-1);
+  // fprintf(stderr, "code_snapshot:\n");
+  // fprintf(stderr, "Working directory is \"%s\".\n", cwd);
 
   create_window();
 
@@ -134,7 +141,7 @@ int main(int argc, char *argv[])
       break;
     }
 
-//    fprintf(stderr, "Reading \"%s\".\n", argv[i]);
+    // fprintf(stderr, "Reading \"%s\".\n", argv[i]);
 
     std::string code;
     std::string filename;
@@ -145,10 +152,11 @@ int main(int argc, char *argv[])
       if (in_code_block) {
         if (strstr(line, "\\endcode_international")) {
           if (!code.empty()) {
-            code.resize( code.size()-1 );
+            code.resize(code.size() - 1);
             save_snapshot(code.c_str(), filename.c_str());
           }
           in_code_block = false;
+          code = "";
         } else {
           code += line;
         }