added signal handler code

pdf2htmlEX/CMakeLists.txt

@@ -149,6 +149,8 @@ set(PDF2HTMLEX_SRC ${PDF2HTMLEX_SRC}
     src/util/unicode.cc
     src/util/mingw.h
     src/util/mingw.cc
+    src/util/SignalHandler.h
+    src/util/SignalHandler.cc
     src/ArgParser.h
     src/ArgParser.cc
     src/Base64Stream.h

pdf2htmlEX/src/pdf2htmlEX.cc

@@ -25,6 +25,8 @@
 
 #include "pdf2htmlEX-config.h"
 
+#include "util/SignalHandler.h"
+
 #if ENABLE_SVG
 #include <cairo.h>
 #endif
@@ -55,11 +57,15 @@ void show_usage_and_exit(const char * dummy = nullptr)
 
 void show_version_and_exit(const char * dummy = nullptr)
 {
+    const FFWVersionInfo* ffwVersionInfo = ffw_get_version_info();
+
     cerr << "pdf2htmlEX version " << PDF2HTMLEX_VERSION << endl;
     cerr << "Copyright 2012-2015 Lu Wang <coolwanglu@gmail.com> and other contributors" << endl;
     cerr << "Libraries: " << endl;
     cerr << "  poppler " << POPPLER_VERSION << endl;
-    cerr << "  libfontforge " << ffw_get_version() << endl;
+    cerr << "  libfontforge " << ffwVersionInfo->majorVersion << "." <<
+      ffwVersionInfo->minorVersion << "." << ffwVersionInfo->gitVersion << endl;
+    cerr << "  libfontforge (date) " << ffwVersionInfo->versionDate << endl;
 #if ENABLE_SVG
     cerr << "  cairo " << cairo_version_string() << endl;
 #endif
@@ -394,6 +400,12 @@ int main(int argc, char **argv)
         exit(EXIT_FAILURE);
     }
 
+    // setup the signal handler
+    setupSignalHandler(argc, (const char**)argv,
+      param.data_dir.c_str(),
+      param.poppler_data_dir.c_str(),
+      param.tmp_dir.c_str());
+
     bool finished = false;
     // read config file
     globalParams = new GlobalParams(!param.poppler_data_dir.empty() ? param.poppler_data_dir.c_str() : NULL);

pdf2htmlEX/src/util/SignalHandler.cc

@@ -0,0 +1,220 @@
+/*
+ * SignalHandler.cc
+ *
+ *  Created on: 2019-Nov-14
+ *      Author: Stephen Gaito
+ */
+
+/**********
+
+This collection of basic ANSI-C functions implement a crude SIG-SEGV
+handler for when things go horribly wrong ;-(
+
+We implement them in as simple a maner as possible so that it will be as
+portable as possible and less likely to fall into a reentrant deadlock.
+
+At the moment the most critical SIG-SEGVs are happening inside the
+FontForge library. So we start by focusing upon adding more helpful
+messages when FontForge can't save a file for us.
+
+**********/
+
+
+#include <stdio.h>
+#include <string.h>
+#include <string>
+#include <unistd.h>
+#include <signal.h>
+
+#include "pdf2htmlEX-config.h"
+#include "util/ffw.h"
+#include <cairo/cairo.h>
+#include <poppler/poppler-config.h>
+
+const char* oopsMessage  =
+  "\n"
+  "\n"
+  "Oops! Something went horribly wrong.... sorry!\n"
+  "\n";
+
+const char* ffwMessage_1 =
+  "I was in the middle of asking FontForge to ";
+
+const char* ffwMessage_2 =
+  " a font.\n"
+  "\n"
+  "I suspect that the problem is that FontForge is having trouble\n"
+  "working with one of the fonts embedded in your pdf file.\n"
+  "\n"
+  "The best way to verify this is to install FontForge and to try\n"
+  "loading and then (re)generating each font in the\n"
+  "\n"
+  "    ";
+
+const char* ffwMessage_3 =
+  "\n"
+  "\n"
+  "directory. See:\n"
+  "\n"
+  "    https://github.com/pdf2htmlEX/pdf2htmlEX/wiki/Troubleshooting-Crashes\n"
+  "\n"
+  "for details.\n"
+  "\n"
+  "If one of these fonts crashes FontForge, then it is certainly not a\n"
+  "pdf2htmlEX problem.... sorry.\n"
+  "\n"
+  "IF it *is* a FontForge problem then you have the following options:\n"
+  "\n"
+  "  1) You can send a copy of the font you found which crashed FontForge to\n"
+  "     the FontForge developers and see if they can find the problem.\n"
+  "     (Be warned FontForge is a complex application, so it might be\n"
+  "      very hard for the developers to isolate your problem.... )\n"
+  "\n"
+  "  2) You can try re-creating your PDF with a different font.\n"
+  "\n"
+  "  3) Or you can do both of these things...\n"
+  "\n"
+  "IF you were not able to crash FontForge using the fonts in the above\n"
+  "directory... then and only then please raise an issue with the pdf2htmlEX\n"
+  "team... make sure you provide them with your pdf, as well as the following\n"
+  "details:\n";
+
+const char* noFfwMessage =
+  "Please raise an issue with the pdf2hmtlEX team so they can fix this...\n"
+  "\n"
+  "Please make sure you provide them with your pdf as well as the following\n"
+  "details:\n";
+
+const char* detailsHeader =
+  "\n"
+  "--------------------------------------------------------------------------\n"
+  "\n";
+
+const char* detailsBody = "no details recorded\n";
+
+const char* pdf2htmlEXTmpDir = NULL;
+const char* ffwAction        = NULL;
+
+struct sigaction act;
+
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wunused-result" 
+void signalHandler(int sigInt) {
+  write(STDERR_FILENO, oopsMessage,        strlen(oopsMessage) );
+
+  if (ffwAction) {
+    write(STDERR_FILENO, ffwMessage_1,     strlen(ffwMessage_1) );
+    write(STDERR_FILENO, ffwAction,        strlen(ffwAction) );
+    write(STDERR_FILENO, ffwMessage_2,     strlen(ffwMessage_2) );
+    write(STDERR_FILENO, pdf2htmlEXTmpDir, strlen(pdf2htmlEXTmpDir) );
+    write(STDERR_FILENO, ffwMessage_3,     strlen(ffwMessage_3) );
+  } else {
+    write(STDERR_FILENO, noFfwMessage,     strlen(noFfwMessage) );
+  }
+
+  write(STDERR_FILENO, detailsHeader,      strlen(detailsHeader) );
+  write(STDERR_FILENO, detailsBody,        strlen(detailsBody) );
+  exit(-1);
+
+}
+#pragma GCC diagnostic pop
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+void ffwSetAction(const char* anAction) {    ffwAction = anAction; }
+void ffwClearAction(const char* anAction) {  ffwAction = NULL; }
+
+#ifdef __cplusplus
+}
+#endif
+
+void setupSignalHandler(
+  int argc, const char* argv[],
+  const char* data_dir,
+  const char* poppler_data_dir,
+  const char* tmp_dir) {
+  // THIS MUST BE CALLED AFTER ARGUMENT PARSING
+
+  // Start by setting up the messages
+  //
+  // Construct the command line information
+  std::string detailInfo;
+  detailInfo = "Command line:\n";
+  for (int i = 0 ; i < argc ; i++ ){
+    detailInfo = detailInfo + "  " + std::to_string(i) + ": [" + argv[i] + "]\n" ;
+  }
+  detailInfo = detailInfo + "\n";
+
+  // Construct the version information
+  const pdf2htmlEX::FFWVersionInfo* ffwVersionInfo =
+    pdf2htmlEX::ffw_get_version_info();
+
+  detailInfo = detailInfo + "Version information:\n";
+  detailInfo = detailInfo + "  pdf2htmlEX version " + pdf2htmlEX::PDF2HTMLEX_VERSION + "\n";
+  detailInfo = detailInfo + "  Copyright 2012-2015 Lu Wang <coolwanglu@gmail.com> and other contributors\n";
+  detailInfo = detailInfo + "\n";
+  detailInfo = detailInfo + "Libraries:\n" ;
+  detailInfo = detailInfo + "  poppler " + POPPLER_VERSION + "\n";
+  detailInfo = detailInfo + "  libfontforge " + ffwVersionInfo->majorVersion + "." + ffwVersionInfo->minorVersion + "." + ffwVersionInfo->gitVersion + "\n";
+  detailInfo = detailInfo + "  libfontforge (date) " + ffwVersionInfo->versionDate + "\n";
+#if ENABLE_SVG
+  detailInfo = detailInfo + "  cairo " + cairo_version_string() + "\n";
+#endif
+  detailInfo = detailInfo + "\n";
+  detailInfo = detailInfo + "Default data-dir: " + data_dir + "\n";
+  detailInfo = detailInfo + "Default poppler-data-dir: " + poppler_data_dir + "\n";
+  detailInfo = detailInfo + "Default tmp-dir: " + tmp_dir + "\n";
+  detailInfo = detailInfo + "\n";
+  detailInfo = detailInfo + "Supported image formats:";
+#ifdef ENABLE_LIBPNG
+  detailInfo = detailInfo + " png";
+#endif
+#ifdef ENABLE_LIBJPEG
+  detailInfo = detailInfo + " jpg";
+#endif
+#if ENABLE_SVG
+  detailInfo = detailInfo + " svg";
+#endif
+  detailInfo = detailInfo + "\n\n";
+
+  detailsBody = strdup(detailInfo.c_str());
+
+  pdf2htmlEXTmpDir = strdup(tmp_dir);
+
+  // Now setup the signal handler
+  //
+  memset(&act, 0, sizeof(act));
+  act.sa_handler = signalHandler;
+  //
+  sigaction(SIGILL,  &act, NULL);  // Illegal Instruction
+  sigaction(SIGFPE,  &act, NULL);  // Floating-point exception
+  sigaction(SIGSEGV, &act, NULL);  // Invalid memory reference
+  sigaction(SIGBUS,  &act, NULL);  // Bus error (bad memory access)
+  sigaction(SIGSYS,  &act, NULL);  // Bad system call (SVr4)
+
+  // All done
+}
+
+
+/***
+int main(int argc, const char* argv[]) {
+
+  setupSignalHandler(argc, argv, "aTmpDir");
+
+  ffwSetAction("save");
+
+  for (int i = 0; i < 5; i++) {
+    printf("sleeping.... \n");
+    sleep(1);
+  }
+
+  int *foo = NULL;
+
+  *foo = 1;
+
+  printf("We should not reach here!\n");
+
+}
+***/

pdf2htmlEX/src/util/SignalHandler.h

@@ -0,0 +1,21 @@
+#ifndef SIGNAL_HANDLER_H
+#define SIGNAL_HANDLER_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+void ffwSetAction(const char* anAction);
+void ffwClearAction(void);
+
+#ifdef __cplusplus
+}
+#endif
+
+void setupSignalHandler(
+  int argc, const char* argv[],
+  const char* data_dir,
+  const char* poppler_data_dir,
+  const char* tmp_dir);
+
+#endif

pdf2htmlEX/src/util/ffw.c

@@ -16,6 +16,8 @@
 #include <fontforge.h>
 #include <baseviews.h>
 
+#include "SignalHandler.h"
+
 #include "ffw.h"
 #include "fontforge-2.0.20170731/autowidth.h"
 #include "fontforge-2.0.20170731/bitmapchar.h"
@@ -66,6 +68,7 @@ static void dumb_post_error(const char * title, const char * error, ...) { }
 
 void ffw_init(int debug)
 {
+    ffwSetAction("initialize");
     InitSimpleStuff();
     if ( default_encoding==NULL )
         default_encoding=FindOrMakeEncoding("ISO8859-1");
@@ -88,10 +91,12 @@ void ffw_init(int debug)
         v.u.ival = 1;
         SetPrefs("DetectDiagonalStems", &v, NULL);
     }
+    ffwClearAction();
 }
 
 void ffw_finalize(void)
 {
+    ffwSetAction("finalize");
     while(enc_head)
     {
         Encoding * next = enc_head->next;
@@ -107,21 +112,36 @@ void ffw_finalize(void)
         free(enc_head);
         enc_head = next;
     }
+    ffwClearAction();
 }
 
-long ffw_get_version(void)
+// see: https://stackoverflow.com/a/2653351
+#define xstr(a) str(a)
+#define str(a) #a
+
+static FFWVersionInfo ffwVersionInfo;
+
+const FFWVersionInfo* ffw_get_version_info(void)
 {
-    return FONTFORGE_VERSIONDATE_RAW;
+    ffwVersionInfo.gitVersion   = FONTFORGE_GIT_VERSION;
+    ffwVersionInfo.majorVersion = xstr(FONTFORGE_VERSION_MAJOR);
+    ffwVersionInfo.minorVersion = xstr(FONTFORGE_VERSION_MINOR);
+    ffwVersionInfo.versionDate  = FONTFORGE_VERSIONDATE;
+
+    return &ffwVersionInfo;
 }
 
 void ffw_new_font()
 {
+    ffwSetAction("create");
     assert((cur_fv == NULL) && "Previous font is not destroyed");
     cur_fv = FVAppend(_FontViewCreate(SplineFontNew()));
+    ffwClearAction();
 }
 
 void ffw_load_font(const char * filename)
 {
+    ffwSetAction("load");
     assert((cur_fv == NULL) && "Previous font is not destroyed");
 
     char * _filename = strcopy(filename);
@@ -147,6 +167,7 @@ void ffw_load_font(const char * filename)
         cur_fv->cidmaster->ascent = cur_fv->sf->ascent;
         cur_fv->cidmaster->descent = cur_fv->sf->descent;
     }
+    ffwClearAction();
 }
 
 /*
@@ -154,6 +175,7 @@ void ffw_load_font(const char * filename)
  */
 void ffw_prepare_font(void)
 {
+    ffwSetAction("prepare");
     memset(cur_fv->selected, 1, cur_fv->map->enccount);
     // remove kern
     FVRemoveKerns(cur_fv);
@@ -185,10 +207,12 @@ void ffw_prepare_font(void)
      */
     free(sf->fontname);
     sf->fontname = strcopy("");
+    ffwClearAction();
 }
 
 void ffw_save(const char * filename)
 {
+    ffwSetAction("save");
     char * _filename = strcopy(filename);
     char * _ = strcopy("");
 
@@ -200,11 +224,14 @@ void ffw_save(const char * filename)
 
     if(!r)
         err("Cannot save font to %s\n", filename);
+    ffwClearAction();
 }
 void ffw_close(void)
 {
+    ffwSetAction("close");
     FontViewClose(cur_fv);
     cur_fv = NULL;
+    ffwClearAction();
 }
 
 static void ffw_do_reencode(Encoding * encoding, int force)
@@ -234,25 +261,32 @@ static void ffw_do_reencode(Encoding * encoding, int force)
 
 void ffw_reencode_glyph_order(void)
 {
+    ffwSetAction("re-encode the glyph order in");
     ffw_do_reencode(original_enc, 0);
+    ffwClearAction();
 }
 
 void ffw_reencode_unicode_full(void)
 {
+    ffwSetAction("re-encode to unicode");
     ffw_do_reencode(unicodefull_enc, 0);
+    ffwClearAction();
 }
 
 void ffw_reencode(const char * encname, int force)
 {
+    ffwSetAction("re-encode");
     Encoding * enc = FindOrMakeEncoding(encname);
     if(!enc)
         err("Unknown encoding %s\n", encname);
 
     ffw_do_reencode(enc, force);
+    ffwClearAction();
 }
 
 void ffw_reencode_raw(int32 * mapping, int mapping_len, int force)
 {
+    ffwSetAction("re-encode (raw1)");
     Encoding * enc = calloc(1, sizeof(Encoding));
     enc->only_1byte = enc->has_1byte = true;
 
@@ -273,10 +307,12 @@ void ffw_reencode_raw(int32 * mapping, int mapping_len, int force)
     enc_head = enc;
 
     ffw_do_reencode(enc, force);
+    ffwClearAction();
 }
 
 void ffw_reencode_raw2(const char ** mapping, int mapping_len, int force)
 {
+    ffwSetAction("re-encode (raw2)");
     Encoding * enc = calloc(1, sizeof(Encoding));
     enc->enc_name = strcopy("");
     enc->char_cnt = mapping_len;
@@ -300,6 +336,7 @@ void ffw_reencode_raw2(const char ** mapping, int mapping_len, int force)
     enc_head = enc;
 
     ffw_do_reencode(enc, force);
+    ffwClearAction();
 }
 
 void ffw_cidflatten(void)
@@ -309,7 +346,9 @@ void ffw_cidflatten(void)
         fprintf(stderr, "Cannot flatten a non-CID font\n");
         return;
     }
+    ffwSetAction("flatten the cid in");
     SFFlatten(cur_fv->sf->cidmaster);
+    ffwClearAction();
 }
 
 /*
@@ -318,6 +357,7 @@ void ffw_cidflatten(void)
  */
 void ffw_add_empty_char(int32_t unicode, int width)
 {
+    ffwSetAction("add an empty character to");
     SplineChar * sc = SFMakeChar(cur_fv->sf, cur_fv->map, cur_fv->map->enccount);
     char buffer[400];
     SCSetMetaData(sc,
@@ -325,22 +365,29 @@ void ffw_add_empty_char(int32_t unicode, int width)
                 cur_fv->sf->uni_interp, cur_fv->sf->for_new_glyphs)),
         unicode, sc->comment);
     SCSynchronizeWidth(sc, width, sc->width, cur_fv);
+    ffwClearAction();
 }
 
 int ffw_get_em_size(void)
 {
-    return cur_fv->sf->ascent + cur_fv->sf->descent;
+    ffwSetAction("get the em size of");
+    int emSize = cur_fv->sf->ascent + cur_fv->sf->descent;
+    ffwClearAction();
+    return emSize;
 }
 
 void ffw_fix_metric()
 {
+    ffwSetAction("fix the metric of");
     double ascent, descent;
     ffw_get_metric(&ascent, &descent);
     ffw_set_metric(ascent, descent);
+    ffwClearAction();
 }
 
 void ffw_get_metric(double * ascent, double * descent)
 {
+    ffwSetAction("get the metric of");
     SplineFont * sf = cur_fv->sf;
 
     DBounds bb;
@@ -357,10 +404,12 @@ void ffw_get_metric(double * ascent, double * descent)
     {
         *ascent = *descent = 0;
     }
+    ffwClearAction();
 }
 
 void ffw_set_metric(double ascent, double descent)
 {
+    ffwSetAction("set the metric of");
     SplineFont * sf = cur_fv->sf;
     struct pfminfo * info = &sf->pfminfo;
 
@@ -402,6 +451,7 @@ void ffw_set_metric(double ascent, double descent)
 
     info->os2_typolinegap = 0;
     info->linegap = 0;
+    ffwClearAction();
 }
 
 /*
@@ -410,6 +460,7 @@ void ffw_set_metric(double ascent, double descent)
 void ffw_set_widths(int * width_list, int mapping_len,
         int stretch_narrow, int squeeze_wide)
 {
+    ffwSetAction("set the widths of");
     SplineFont * sf = cur_fv->sf;
 
     if(sf->onlybitmaps
@@ -453,13 +504,17 @@ void ffw_set_widths(int * width_list, int mapping_len,
 
         SCSynchronizeWidth(sc, width_list[i], sc->width, cur_fv);
     }
+    ffwClearAction();
 }
 
 void ffw_import_svg_glyph(int code, const char * filename, double ox, double oy, double width)
 {
+    ffwSetAction("import the glyphs from");
     int enc = SFFindSlot(cur_fv->sf, cur_fv->map, code, "");
-    if(enc == -1)
+    if(enc == -1) {
+        ffwClearAction();
         return;
+    }
 
     SplineChar * sc = SFMakeChar(cur_fv->sf, cur_fv->map, enc);
 
@@ -483,10 +538,12 @@ void ffw_import_svg_glyph(int code, const char * filename, double ox, double oy,
 
         SCSynchronizeWidth(sc, floor(width * (a+d) + 0.5), sc->width, cur_fv);
     }
+    ffwClearAction();
 }
 
 void ffw_auto_hint(void)
 {
+    ffwSetAction("automatically hint");
     // convert to quadratic
     if(!(cur_fv->sf->layers[ly_fore].order2))
     {
@@ -496,11 +553,14 @@ void ffw_auto_hint(void)
     memset(cur_fv->selected, 1, cur_fv->map->enccount);
     FVAutoHint(cur_fv);
     FVAutoInstr(cur_fv);
+    ffwClearAction();
 }
 
 void ffw_override_fstype(void)
 {
+    ffwSetAction("override the fstype of");
     *(int16 *)(&cur_fv->sf->pfminfo.fstype) = 0;
     cur_fv->sf->pfminfo.pfmset = true;
     cur_fv->sf->changed = true;
+    ffwClearAction();
 }

pdf2htmlEX/src/util/ffw.h

@@ -23,7 +23,15 @@ extern "C" {
 // global
 void ffw_init(int debug);
 void ffw_finalize(void);
-long ffw_get_version(void);
+
+typedef struct ffw_version_info {
+  const char* gitVersion;
+  const char* majorVersion;
+  const char* minorVersion;
+  const char* versionDate;
+} FFWVersionInfo ;
+
+const FFWVersionInfo* ffw_get_version_info(void);
 
 ////////////////////////
 // load & save