[Snowball-discuss] More patches

Olly Betts olly at survex.com
Mon Feb 12 08:30:28 GMT 2007


OK, here are some more:

This changes the generated C code to avoid shadowed local variables.  In
particular it generates int c1, c2, c3 instead of int c and the same for
m.  Also, the existing use of m3 is renamed to mlimit (which is clearer
anyway).  The motivations are that this makes the generated code easier
to understand to the human reader, and allows it to be compiled without
warnings with the GCC option -Wshadow, which avoids some special casing
in Xapian's build system:

http://oligarchy.co.uk/xapian/patches/snowball-no-shadowed-variables.patch

This patch eliminates a shadowed variable in the snowball compiler
itself, and 2 cases in the runtime:

http://oligarchy.co.uk/xapian/patches/snowball-rename-shadowed-variables-in-compiler-and-runtime.patch

This fixes various typos in the website (the change to r1r2.html
reflects that some of the stemmers now shipped don't actually use r1 and
r2 - lovins is the most obvious example):

http://oligarchy.co.uk/xapian/patches/snowball-website-typos.patch

This adds a "make check" rule which verifies that the UTF-8 and
ISO-8859-1 versions of the stemmers actually produce the expected
output on the test vocabulary.  To simplify the implementation
of this, the patch also converts all the voc.txt and output.txt
files to UTF-8 (the romanian ones were already) - I just ran them
through iconv with suitable options to do this:

http://oligarchy.co.uk/xapian/patches/snowball-make-check.patch

Cheers,
    Olly



More information about the Snowball-discuss mailing list