libjpeg-turbo

android varient fails make test

Bug #823960 reported by Tom Gall
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
libjpeg-turbo
Fix Released
High
Tom Gall
libjpeg-turbo 1.1.90-2011.09

Bug Description

When using the upstream code from : http://git.linaro.org/gitweb?p=people/tomgall/libjpeg-turbo/libjpeg-turbo.git;a=summary

The 1.1.1-linaro-android branch.

Make test fails.

Tom Gall (tom-gall)
Changed in libjpeg-turbo:
assignee: nobody → Tom Gall (tom-gall)
importance: Undecided → High
status: New → Confirmed
Revision history for this message
Tom Gall (tom-gall) wrote :
Download full text (216.8 KiB)

tgall@nereid:~/libjpeg/bld/android-mru/libjpeg-turbo$ make test
rm -f testout*
rm -f *_GRAYQ[0-9]*.bmp
rm -f *_GRAYQ[0-9]*.ppm
rm -f *_GRAYQ[0-9]*.jpg
rm -f *_GRAY.yuv
rm -f *_420Q[0-9]*.bmp
rm -f *_420Q[0-9]*.ppm
rm -f *_420Q[0-9]*.jpg
rm -f *_420.yuv
rm -f *_422Q[0-9]*.bmp
rm -f *_422Q[0-9]*.ppm
rm -f *_422Q[0-9]*.jpg
rm -f *_422.yuv
rm -f *_444Q[0-9]*.bmp
rm -f *_444Q[0-9]*.ppm
rm -f *_444Q[0-9]*.jpg
rm -f *_444.yuv
make all-recursive
make[1]: Entering directory `/bld/home/tgall/libjpeg/bld/android-mru/libjpeg-turbo'
Making all in simd
make[2]: Entering directory `/bld/home/tgall/libjpeg/bld/android-mru/libjpeg-turbo/simd'
make all-am
make[3]: Entering directory `/bld/home/tgall/libjpeg/bld/android-mru/libjpeg-turbo/simd'
make[3]: Nothing to be done for `all-am'.
make[3]: Leaving directory `/bld/home/tgall/libjpeg/bld/android-mru/libjpeg-turbo/simd'
make[2]: Leaving directory `/bld/home/tgall/libjpeg/bld/android-mru/libjpeg-turbo/simd'
make[2]: Entering directory `/bld/home/tgall/libjpeg/bld/android-mru/libjpeg-turbo'
make[2]: Leaving directory `/bld/home/tgall/libjpeg/bld/android-mru/libjpeg-turbo'
make[1]: Leaving directory `/bld/home/tgall/libjpeg/bld/android-mru/libjpeg-turbo'
./jpegut
RGB Top-Down -> 4:4:4 Q100 ... Done. 6.195068 ms
  Result in test_enc_RGB_TD_444Q100.jpg
JPEG -> RGB Top-Down ... FAILED!
255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 255/255/255 255/255/255 255/255/255
255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 255/255/255 255/255/255 255/255/255
255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 255/255/255 255/255/255 255/255/255
255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 255/255/255 255/255/255 255/255/255
255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 255/255/255 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/254 000/000/2...

Revision history for this message
Tom Gall (tom-gall) wrote :

Some narrowing information from the past couple of days that I've looked into this.

1.1.1 vanilla (with no linaro changes) this does not fail.
1.1.1 with linaro changes fails
1.1.2 (under review for release) with linaro changes fail
1.2 upstream does not fail

Under gdb, it would appear that the linaro yvup2bgr888_venu that was added is not quite able to deal with BGR, BGRX encoded jpeg images correctly.

Revision history for this message
Tom Gall (tom-gall) wrote :

And indeed if one comments out the use of jsimd_ycc_rgb_convert in jdcolor.c at line 375 which in turn calls yvup2bgr888_venum the bug as reported goes away.

Revision history for this message
Tom Gall (tom-gall) wrote :

jsimd_ycc_rgb_convert is basically wrong. Looking at back porting the 1.1.90 code which is correct.

Changed in libjpeg-turbo:
status: Confirmed → Triaged
Fathi Boudra (fboudra)
Changed in libjpeg-turbo:
milestone: none → 1.1.90
status: Triaged → Fix Released
Revision history for this message
Fathi Boudra (fboudra) wrote :

the Android WG bugs #823960 and #826642 are not present in libjpeg-turbo 1.1.90.
see http://lists.linaro.org/pipermail/linaro-dev/2011-August/006747.html

Revision history for this message
Tom Gall (tom-gall) wrote :

Corrected status of the bug and assigned it to the correct milestone.

Changed in libjpeg-turbo:
milestone: 1.1.90 → 1.1.90-android
status: Fix Released → Triaged
Revision history for this message
Tom Gall (tom-gall) wrote :

On linux but building with the android options all on, make test now passes.

Further on my android nook install, jpegs appear to be correctly displaying.

Changed in libjpeg-turbo:
status: Triaged → Fix Committed
Tom Gall (tom-gall)
Changed in libjpeg-turbo:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.
Subscribing...

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.