xgettext: don't strip by default, ignore empty strings and all whitespace strings, process special characters such as the "bell character"

I have some experience with building large databases with strings extracted from source code. Some of my findings:

* ignore empty strings: you will find that many strings will be the empty string. These are quite useless for anything related to matching.
* some strings will be white space only (before stripping). These tend to be useless as well.
* there are quite a few characters that cannot be printed, such as the ASCII bell character. You might want to remove these. A test example would be the file `libbb/lineedit.c` from a recent version of BusyBox. The whole list of characters that I am currently removing:

```
['\a', '\b', '\v', '\f', '\x01', '\x02', '\x03', '\x04',
'\x05', '\x06', '\x0e', '\x0f', '\x10', '\x11', '\x12',
'\x13', '\x14', '\x15', '\x16', '\x17', '\x18', '\x19',
'\x1a', '\x1b', '\x1c', '\x1d', '\x1e', '\x1f', '\x7f']
```

Currently you are not doing those clean ups. On the other hand you *are* stripping regular strings, where (I think) whitespace *could* be relevant. If you want to clean up, then at least you should be consistent :-)

My advise: do not strip strings, ignore empty strings or whitespace only strings, remove non-printable characters.