The bugs fixed are:
- programs required for -o2 weren't checked for when using -o2
but were checked for lower -o levels.
- if no savings could be found the results printed at the end
were confusing and the script had a divide by zero error.
BUG=None
Review URL: https://chromiumcodereview.appspot.com/15970004
git-svn-id: http://src.chromium.org/svn/trunk/src/build@203051 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
There seems to be a bug in optipng's bit depth reduction when coverting to grayscale.
We don't know exact condition, but its savings is relatively small, so
we should just skip this process for grayscale images.
BUG=174505
TEST=none
NOTRY=true
Review URL: https://codereview.chromium.org/12224023
git-svn-id: http://src.chromium.org/svn/trunk/src/build@181170 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
default (=1) does less optimization but runs fast
-o (=2) does more optimization but slow.
-n just run pngcrush and skip other optimizations
default produces good enough results and runs a lot faster,
fast enough so that we can now ask people to run.
Changed to show the results for changed files only
NOTRY=true
BUG=150046
TEST=none
Review URL: https://chromiumcodereview.appspot.com/12095097
git-svn-id: http://src.chromium.org/svn/trunk/src/build@180254 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
The optimize code is based on pngslim. A few steps are omitted, but it runs all other processes. I'll check if it
worth fixing them (for example, advdef didn't make much difference)
Using -o option requires new third party program: optpng, pngout and advdef.
The script asks to install if they're not installed.
So for, I've got ~30% reduction in ash/resources (it's probably because pngslim was never run on this directory), I'd expect less
on other places where pngslim was run before.
Print results
BUG=150046
TEST=none
NOTRY=true
Review URL: https://chromiumcodereview.appspot.com/11905002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@177202 4ff67af0-8c30-449e-8e8b-ad334ec8d88c