Closed catharsis71 closed 1 year ago
I tried this on Linux and couldn't reproduce the problem... Were you using v1.5.0? Does it happen if you compile the latest version from github (v1.5.1beta) ?
I tested on two systems:
I downloaded the 1.5.0 source for both but obviously had to compile separately since one is ARM and one is AMD.
Note that the number of garbled lines is proportionally small so you may only see this when processing a large number of files
On system 1, w=1 is clean, then I start seeing garbled lines when using w=2. Moving to w=3, the number of garbled lines actually decreases for some reason, then increases again when using w=4
root@uk2:/var/www/fanart/unsorted/1# jpegoptim-test -w 1 -n --nofix *.jpg | grep -v skipped | wc -l
0
root@uk2:/var/www/fanart/unsorted/1# jpegoptim-test -w 1 -n --nofix *.jpg | grep -v skipped | wc -l
0
root@uk2:/var/www/fanart/unsorted/1# jpegoptim-test -w 2 -n --nofix *.jpg | grep -v skipped | wc -l
85
root@uk2:/var/www/fanart/unsorted/1# jpegoptim-test -w 2 -n --nofix *.jpg | grep -v skipped | wc -l
86
root@uk2:/var/www/fanart/unsorted/1# jpegoptim-test -w 3 -n --nofix *.jpg | grep -v skipped | wc -l
44
root@uk2:/var/www/fanart/unsorted/1# jpegoptim-test -w 3 -n --nofix *.jpg | grep -v skipped | wc -l
45
root@uk2:/var/www/fanart/unsorted/1# jpegoptim-test -w 4 -n --nofix *.jpg | grep -v skipped | wc -l
91
root@uk2:/var/www/fanart/unsorted/1# jpegoptim-test -w 4 -n --nofix *.jpg | grep -v skipped | wc -l
90
root@uk2:/var/www/fanart/unsorted/1#
On system 2, I don't see garbled lines until w=4 after which the number of garbled lines varies unpredictably:
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 1 -n --nofix *.jpg | grep -v skipped | wc -l
0
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 1 -n --nofix *.jpg | grep -v skipped | wc -l
0
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 2 -n --nofix *.jpg | grep -v skipped | wc -l
0
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 2 -n --nofix *.jpg | grep -v skipped | wc -l
0
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 3 -n --nofix *.jpg | grep -v skipped | wc -l
0
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 3 -n --nofix *.jpg | grep -v skipped | wc -l
0
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 4 -n --nofix *.jpg | grep -v skipped | wc -l
44
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 4 -n --nofix *.jpg | grep -v skipped | wc -l
44
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 6 -n --nofix *.jpg | grep -v skipped | wc -l
93
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 6 -n --nofix *.jpg | grep -v skipped | wc -l
93
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 8 -n --nofix *.jpg | grep -v skipped | wc -l
46
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 8 -n --nofix *.jpg | grep -v skipped | wc -l
46
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 10 -n --nofix *.jpg | grep -v skipped | wc -l
46
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 10 -n --nofix *.jpg | grep -v skipped | wc -l
46
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 12 -n --nofix *.jpg | grep -v skipped | wc -l
91
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 12 -n --nofix *.jpg | grep -v skipped | wc -l
46
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 24 -n --nofix *.jpg | grep -v skipped | wc -l
90
:~/backup/www/fanart/unsorted/1$ jpegoptim-150 -w 24 -n --nofix *.jpg | grep -v skipped | wc -l
44
Even though these are two separate systems, in the examples above, they're running on the same set of files (1861 JPEGs totaling 329MB) as I synchronized the two systems with rsync before testing.
I will try compiling from master tomorrow
I recall there was some fixes to output since 1.5.0, that may have "fixed" this as well...
Single thread (no output because all files are already optimized and I'm using grep to exclude "skipped" messages):
Multi thread (same directory, same files, output varies randomly from run to run):
Leaving off the
| grep -v skipped
, here are a few examples of garbling issues in the full output: