git

Status/Intentions

• Read that C++ 2017 standard should support UTF-16!? Decided to try to build htdig with it...
• I was able to pull my fork up-to-date for the icu repo, but not for the git one. So, removed, and cloned again (my fork) setting up the remote upstream. I was probably wrong: it would have worked. I have to pull, commit, and push —to my own fork. Problem: that's an endless task...
• Restored this page to June 16th version... Noted that I had already created a fork of icu, not cloned it? Looks like what I cloned as ~/git/unicode-org/icu is the master... So now, cloning my fork to ~/git/mgirod/icu
• Berkeley DB 6.2.32
• In git/hddev, a clone on the dev branch. The clones connect to each other: one can use git pull.
• In git/htdig, my development version, compiled with modern g++, with most warnings eliminated (although a few new ones popped up!?).
  The current code doesn't fail anymore to find non unicode search words in the db (tested with default). This shows up from the command line and thus under the debugger (shows previous status).
  My understanding for now is that before taking care of unicode, either with icu, or even without it, if support for it has already been integrated into the C++ library, I have to switch to standard string and list, etc...
• Under gdb, one can give command such as:

  gdb ./htsearch/htsearch
  handle SIGALRM ignore
  run restrict=foo+bar;words=default;format=builtin-short

• Build with: make CXXFLAGS="-g -O0"
• Working on strlist: public list<string> to replace StringList, e.g. in the context of htsearch.cc. Replacing it required to extract and add a create (parser) and a join functions.
• Download and commit to git the latest temporary version, in case it would have be dealing with the DB error, assuming this is not related to my port to recent gcc.
• Fix the DB error (it disappeared by itself on May 28?)
• Import to git the icu tree.
• Fix the shared library issue.
• Try to replace the native htdig regexp handling with the icu one.
  Dumped wchar_t, i.e. utf-8. Using utf-16 as does icu internally. So far not colliding against C++ specializations, although may in later stand versions (2014, 2018...?)