OCR - tesseract needs work :-)

[Tom Horsley]

For curiosity, I downloaded and tried to build the tesseract OCR
app that appeared in the original OCR message thread earlier.

It may have worked with the 2.96 gcc on a 32 bit system, but
the 4.1 gcc on my 64 bit FC6 system has fits.

Looking at some of the fits it is throwing, I can see that the
code simply has 64 bit problems - I don't think it will build
on a 64 bit system without a lot of work.

using CC="gcc -m32" and CXX="g++ -m32" env vars to force it
to build 32 bit, I get a lot farther, but run into more problems
with the new compiler complaining about gcc/g++ conflicting
declarations.

Probably not hard to fix, but tesseract definitely isn't as
easy to build as ocrad or gocr.