Discard absurd values (and log warnings) when graphing bench results.

Discard absurd values (and log warnings) when graphing bench results.

BUG=http://code.google.com/p/skia/issues/detail?id=596
Committed: https://code.google.com/p/skia/source/detail?r=3914

[bungeman, 2012-05-11 18:02:15]

Not properly escaping makes my eye twitch. Other than that, LGTM.

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py
File bench/bench_graph_svg.py (right):

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:393: description += 'r%d: %s<br />\n' % (revision,
points_at_this_revision)
This is going into a textarea, do we need the <br />? I don't think so, but
there may be somewhere where it makes a difference.

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:403: print description
should probably be qe(description) (if we don't need the <br /> above). If we do
need the <br /> then points_at_this_revision should be qe'ed above where
description is set.

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:414: print '<title>%s</title>' % title
Since I saw this below, at some point this should be qe(title)

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:503: print '<table border="0" width="%s"><tr
valign="top"><td width="50%%">' % requested_width
%%?

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:553: title,
I know this is just moved, but while moving it, qe(title)

[epoger, 2012-05-11 18:25:11]

Thanks for the suggestions!  I took 'em all.  Will commit in a minute...

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py
File bench/bench_graph_svg.py (right):

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:393: description += 'r%d: %s<br />\n' % (revision,
points_at_this_revision)
On 2012/05/11 18:02:15, bungeman wrote:
> This is going into a textarea, do we need the <br />? I don't think so, but
> there may be somewhere where it makes a difference.

Right you are, we don't need it.

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:403: print description
On 2012/05/11 18:02:15, bungeman wrote:
> should probably be qe(description) (if we don't need the <br /> above). If we
do
> need the <br /> then points_at_this_revision should be qe'ed above where
> description is set.

Done.

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:414: print '<title>%s</title>' % title
On 2012/05/11 18:02:15, bungeman wrote:
> Since I saw this below, at some point this should be qe(title)

Done.

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:414: print '<title>%s</title>' % title
On 2012/05/11 18:02:15, bungeman wrote:
> Since I saw this below, at some point this should be qe(title)

Done.

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:503: print '<table border="0" width="%s"><tr
valign="top"><td width="50%%">' % requested_width
On 2012/05/11 18:02:15, bungeman wrote:
> %%?

Moved it into the triple-quoted section so just one % does the trick.

http://codereview.appspot.com/6197066/diff/2001/bench/bench_graph_svg.py#newc...
bench/bench_graph_svg.py:553: title,
On 2012/05/11 18:02:15, bungeman wrote:
> I know this is just moved, but while moving it, qe(title)

Done.