Replaced Unicode non-ASCII codepoints from x86.csv

[coruus]

The presence of Unicode non-ASCII codepoints caused building to fail on Ubuntu 13.10 with default package versions. (That is, system built according to instructions in repo.)

All of these characters appear to be in the descriptions of the instructions, so this shouldn't have any impact on correctness.

Here's the script I used to fixup x64.csv; licensed CC0 if you want to incorporate in ODS-extraction process.

#!/usr/bin/python
"""Replace some Unicode characters in x64.csv"""
import sys

REPLACEMENTS = [
  (u'\u03c0', u'pi'),
  (u'\u2013', u'-'),
  (u'\u2014', u'-'),
  (u'\u2217', u'*'),
  (u'\u201d', u'"'),
  (u'\u201c', u'"')]

def do(ifn, ofn):
  with open(ifn, 'rb') as f:
    u = f.read().decode('utf-8')

  for c, repl in REPLACEMENTS:
    u = u.replace(c, repl)

  s = u.encode('ascii')

  with open(ofn, 'wb') as f:
    f.write(s)

if __name__ == '__main__':
  do(sys.argv[1], sys.argv[2])

[bchurchill]

Hi Coruus,

I think we used to build this on Ubuntu 13.10 without problems. When we get a chance we'll investigate this (I'll add an item in the tracker); it's possible there's something missing from our build documentation. @eschkufz will have the final say in the matter; we'll keep you posted.

[eschkufz]

It makes sense that this is an issue. The text descriptions of the instructions were copy-pasted from a pdf version of the intel manual. I'm totally fine merging this in, since it sounds like all it doesn't affect functionality at all and will help more people use our code.

[eschkufz]

There is a larger issue here, which is that this is just a stop-gap, and that everytime we edit the spreadsheet, we end up exporting it to .csv by hand. This means that whatever fixes we end up merging right now will have to be re-merged the next time around as well.

A longer-term solution is to transition over to using just the csv file and ditching the spreadsheet all together. We can change the tab characters so that the whole thing is more readable/editable. I'll add this as a longer term issue.

Berkeley, can you merge this request? You probably have a better idea than I do right now what branches are in what state and for some reason I can't find the network graph on github right now. Feel free to close whenever you want.

[bchurchill]

All things said and done, I've merged this into all the branches. I've also created a separate 'develop' branch for features or things that need more testing. The 'berkeley' branch is really just for my experiments and is becoming pretty hacky.