Proudly breaking your toolchain

LibreOffice Conference 2014 (http://conference.libreoffice.org/2014/speaker/#miklos-vajna), Bern
Video: https://vimeo.com/143871602

Transcript

1. Proudly breaking
   your toolchain
   Miklos Vajna
   2014­09­04
   

   View Slide

2. 2 / 5
   LibreOffice Conference 2014, Bern | Miklos Vajna
   Android
   ●
   Gdb (README.Android):
   ●
   Ld (tinderbox output):
   Some versions of the NDK had a broken gdb in the way that it can see
   symbols only for shlibs that were already loaded when the debuggee was
   attached, so you need to carefully guess where to put:
   fprintf(stderr, "Sleeping NOW!\n"); ::sleep(20);
   into the code; and when you see that in logcat, you have time to run:
   ndk-gdb and it will attach the process.
   /home/android/android-ndk-r9d/toolchains/arm-linux-androideabi-
   4.8/prebuilt/linux-x86_64/bin/../lib/gcc/arm-linux-androideabi/4.8/../../../../arm-
   linux-androideabi/bin/ld.gold: internal error in global_symbol, at /s/ndk-
   toolchain/src/build/../binutils/binutils-2.23/gold/object.h:2041
   

   View Slide

3. 3 / 5
   LibreOffice Conference 2014, Bern | Miklos Vajna
   MSVC
   ●
   7aa7047eb48e81bc0b32448ff0487c993fe9a4db
   had to be reverted:
   fatal error C1061: compiler limit :
   blocks nested too deeply
   Apparently there can be much more than "just
   a few values" in some cases (186 in
   OOXMLFactory_dml_shapeGeometry::getList
   Value) and MSVC compiler cannot handle if
   statement that deeply nested.
   

   View Slide

4. 4 / 5
   LibreOffice Conference 2014, Bern | Miklos Vajna
   (Apple's) clang
   ●
   d7aef2407a0d9e8ff3bf72e198feb55a71307733
   #ifdef IOS
   static void dummy_function()
   {
   pid_t pid = getpid();
   (void) pid;
   }
   #endif
   #ifdef IOS
   // Work around what is
   // either a compiler bug in
   // Xcode 5.1.1,
   ...
   dummy_function();
   #endif
   

   View Slide

5. 5 / 5
   LibreOffice Conference 2014, Bern | Miklos Vajna
   GCC
   ●
   2b5953a19e36a02040f2ff08bc87efe4785f80bd
   ●
   There is hope! (Suspicious SbxDataType)
   https://gcc.gnu.org/bugzilla/show_bug.cgi?id=36187
   // gcc4.4 (and 4.3 and possibly older) have a problem
   with dynamic_cast directly to the target class,
   // so help it with an intermediate cast. I'm not sure
   what exactly the problem is, seems to be unrelated
   // to RTLD_GLOBAL, so most probably a gcc bug.
   FormulaExportBase* pBase =
   dynamic_cast(dynamic_cast<
   SfxBaseModel*>(xClosable.get()));
   

   View Slide