[Ffmpeg-devel] [PATCH] reduce MAX_SCAN_PACKETS in mpegts.c

[Nico Sabbi]

Michael Niedermayer wrote:

>no objections assuming regression tests pass & the code works
>
>[...]
>  
>
well, the code works and as for regressions:
before patch:

--- ./data/ffmpeg.regression    2006-08-27 12:00:00.000000000 +0200
+++ /home/nico/ffmpeg2/tests/ffmpeg.regression.ref      2006-08-27 
11:53:59.000000000 +0200
@@ -27,9 +27,9 @@
 810646 ./data/a-mpeg2threadivlc.mpg
 6b820ffc74de50cc863b41d51f72c22b *./data/out.yuv
 stddev:  7.64 PSNR:30.45 bytes:7602176
-e372b4176b0dea357bf2e20029c8e0d6 *./data/a-mpeg2reuse.mpg
-2121253 ./data/a-mpeg2reuse.mpg
-0e5ab36ad369b7c584dbd360f5f86aab *./data/out.yuv
+edb3e58ea66de2e9b32713daf9b5c3f1 *./data/a-mpeg2reuse.mpg
+2119706 ./data/a-mpeg2reuse.mpg
+1dab2eaca5320fe01afbe9596b2bd6cc *./data/out.yuv
 stddev:  7.68 PSNR:30.41 bytes:7602176
 f24f317ad39c5f2d0195ce0675a828e7 *./data/a-msmpeg4v2.avi
 643702 ./data/a-msmpeg4v2.avi
@@ -75,10 +75,10 @@
 601276 ./data/a-mpeg4-adv.avi
 0b7935cee2070f3b838478cea6dc4b3c *./data/out.yuv
 stddev: 10.25 PSNR:27.90 bytes:7602176
-d19817fff546ac68331b2050a9d4e47a *./data/a-mpeg4-thread.avi
-768820 ./data/a-mpeg4-thread.avi
-fcf87bbceb80ac044a3ceda070fcd1fd *./data/out.yuv
-stddev: 12.32 PSNR:26.30 bytes:7602176
+729af3941305c27ce330e2afaebb5592 *./data/a-mpeg4-thread.avi
+765722 ./data/a-mpeg4-thread.avi
+252d1e8cc1ee88eea693cebd6b4e7e53 *./data/out.yuv
+stddev: 12.32 PSNR:26.31 bytes:7602176
 670f6e540bcf504ccc72d41952cc401e *./data/a-mpeg4-Q.avi
 900488 ./data/a-mpeg4-Q.avi
 af59c063ed644c7dc4151a0a2debbddd *./data/out.yuv


after patch:
--- ./data/ffmpeg.regression    2006-08-27 12:15:40.000000000 +0200
+++ /home/nico/ffmpeg2/tests/ffmpeg.regression.ref      2006-08-27 
11:53:59.000000000 +0200
@@ -27,9 +27,9 @@
 810646 ./data/a-mpeg2threadivlc.mpg
 6b820ffc74de50cc863b41d51f72c22b *./data/out.yuv
 stddev:  7.64 PSNR:30.45 bytes:7602176
-d2dacf5450a6d3514450ee1c1a75d061 *./data/a-mpeg2reuse.mpg
-2122606 ./data/a-mpeg2reuse.mpg
-28f42f44e1873d1ea44a37f7ad85fdf0 *./data/out.yuv
+edb3e58ea66de2e9b32713daf9b5c3f1 *./data/a-mpeg2reuse.mpg
+2119706 ./data/a-mpeg2reuse.mpg
+1dab2eaca5320fe01afbe9596b2bd6cc *./data/out.yuv
 stddev:  7.68 PSNR:30.41 bytes:7602176
 f24f317ad39c5f2d0195ce0675a828e7 *./data/a-msmpeg4v2.avi
 643702 ./data/a-msmpeg4v2.avi
@@ -75,10 +75,10 @@
 601276 ./data/a-mpeg4-adv.avi
 0b7935cee2070f3b838478cea6dc4b3c *./data/out.yuv
 stddev: 10.25 PSNR:27.90 bytes:7602176
-9900cc3b1da5f75fb19bab09fd4b0c90 *./data/a-mpeg4-thread.avi
-768554 ./data/a-mpeg4-thread.avi
-301e4d4c10e4872da200c763f2cf783f *./data/out.yuv
-stddev: 12.32 PSNR:26.30 bytes:7602176
+729af3941305c27ce330e2afaebb5592 *./data/a-mpeg4-thread.avi
+765722 ./data/a-mpeg4-thread.avi
+252d1e8cc1ee88eea693cebd6b4e7e53 *./data/out.yuv
+stddev: 12.32 PSNR:26.31 bytes:7602176
 670f6e540bcf504ccc72d41952cc401e *./data/a-mpeg4-Q.avi
 900488 ./data/a-mpeg4-Q.avi
 af59c063ed644c7dc4151a0a2debbddd *./data/out.yuv


Even running the tests on the unpatched code 2 times in a row give 
different checksums.
I don't think my patch intruduced any regression (after all the code is 
the same as before),
but these test results leave me with some doubt.