October 2012 - DVDnav-discuss - lists.mplayerhq.hu

[PATCH] "dvdnav_jump_to_sector" as an alternative to "dvdnav_time_search" (REV 6: add mode argument)
by gnosygnu 16 Dec '12

16 Dec '12

This revision features one minor change. It adds a new "mode" parameter to "dvdnav_jump_to_sector". As per the included comment: /* NOTE: Mode is currently unimplemented. Only 0 should be passed. */ /* 1 and -1 are for future implementation */ /* 0: Default. Jump to a time which may be either <> time_in_pts_ticks */ /* 1: After. Always jump to a time that is > time_in_pts_ticks */ /* -1: Before. Always jump to a time that is < time_in_pts_ticks */ dvdnav_status_t dvdnav_jump_to_sector_by_time(dvdnav_t *this, uint64_t time_in_pts_ticks, int32_t mode) ... The corresponding .h interface was also updated. My apologies for not implementing the new parameter, but as I detailed in another email, the change looks quite significant. As far as I know, there are no other known pending changes/requests. I am hoping that this patch is at a final state. Please let me know what else should be done to get this patch accepted. Thanks.

4 6

[PATCH] Fix out of array pointer access
by Bryce Harrington 10 Dec '12

10 Dec '12

This patch fixes a segmentation fault hit when reading the DVD 'The Express'. It prevents a read/write beyond end of an array due to using a length value taken from the DVD, which can exceed the allocated size. https://bugs.launchpad.net/ubuntu/+source/libdvdread/+bug/894170 The patch was originally written by rickyrockrat (sorry, I don't have his email address) for 4.1.3. I got the DVD and reproduced the segfault using 4.2.0 and verified the patch stops the segfault from happening. We're not confident this is the best fix though, so are posting it here for review. Signed-off-by: Bryce Harrington <bryce(a)canonical.com> diff -Nrup libdvdread-4.2.0//src/ifo_read.c libdvdread-4.2.0-working//src/ifo_read.c --- libdvdread-4.2.0//src/ifo_read.c 2011-12-13 20:24:16.000000000 -0800 +++ libdvdread-4.2.0-working//src/ifo_read.c 2011-12-13 21:21:53.915978130 -0800 @@ -1082,6 +1082,12 @@ int ifoRead_TT_SRPT(ifo_handle_t *ifofil return 0; } + if(tt_srpt->nr_of_srpts>info_length/sizeof(title_info_t)){ + fprintf(stderr,"libdvdread: data mismatch: info_length (%ld)!= nr_of_srpts (%d). Truncating.\n", + info_length/sizeof(title_info_t),tt_srpt->nr_of_srpts); + tt_srpt->nr_of_srpts=info_length/sizeof(title_info_t); + } + for(i = 0; i < tt_srpt->nr_of_srpts; i++) { B2N_16(tt_srpt->title[i].nr_of_ptts); B2N_16(tt_srpt->title[i].parental_id);

3 3

[PATCH] fix C++ compatibility broken in r1208
by Anssi Hannula 09 Dec '12

09 Dec '12

Revision 1208 ("Add dvdnav_program_play & dvdnav_current_title_program") added dvdnav_program_play function to dvdnav.h. However, unlike the other API functions, its first parameter is named 'this' instead of 'self'. 'this' is a C++ keyword and prevents compilation of C++ projects using libdvdnav. Restore C++ compatibility by renaming the parameter to 'self' like in the other functions. -- Anssi Hannula

3 2

[PATCH] Make sure we get nav packets for all cells in multi angle
by Joakim Plate 09 Dec '12

09 Dec '12

Hi, This is a resend of an old patch that we still keep in xbmc's version of libdvdnav. The problem is that libdvdnav jumps over NAV packets when playing multi angle features. I don't remember all the exact details on this patch anymore since it was many years since i tracked it down. But it is very obvious in xbmc since we use nav packet end/start times to correct timestamps. Without the patch we git gaps in nav packets that doesn's exist in the streams. I think i still have a dvd sample that shows the issue. /Joakim

3 5

libdvdnav 4.1.4
by Jean-Jacques Sarton 09 Dec '12

09 Dec '12

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello, On Fedora 14 totem and vlc end both with folowing message: libdvdread: Found 1 VTS's libdvdread: Elapsed time 0 libdvdnav: Language 'en' not found, using '' instead libdvdnav: Menu Languages available: libdvdnav: Language 'en' not found, using '' instead libdvdnav: Menu Languages available: libdvdnav: Language 'en' not found, using '' instead libdvdnav: Menu Languages available: libdvdnav: *** pgci_ut handle is NULL *** totem: /builddir/build/BUILD/libdvdnav-4.1.4/src/vm/vm.c:1772: get_ID: Assertion `pgcit != ((void *)0)' failed. Of course tje line totem:... is to be replaced withz vlc:... according to the used player The dvd contain no menu and contain a privat video. If I rename libdvdnav totem/vlc works. Regards, Jean-Jacques Sarton -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) iEYEARECAAYFAk0h3bAACgkQM9JbiR3CwQsEEgCggN7qJrKQXspNYBJbCGRh237R +joAn1pQaRWHFN+nASDqRbL6Jgj3CM5S =iNgb -----END PGP SIGNATURE-----

4 6

[PATCH] consuming all memory when scanning dvd "Up"
by John Stebbins 09 Dec '12

09 Dec '12

First a disclaimer. This is a big invasive patch. If you don't like it, my feelings won't be hurt. I will be using it in HandBrake. So if you see problems with it, please let me know. Many dvd's are showing up that have titles that have numerous repeated language unit tables and pgc's. The new release "Up" is an example. I believe "Dark Knight" also exhibits this. When scanning such discs with lsdvd (and other programs that scan all titles), libdvdread will consume many GB of data for these repeated elements and gets very slow (e.g hours to read all titles) on OS X due to reading the duplicate data from uncached raw devices. This patch detects duplicates and reference counts pgc's and pgcit's. When a duplicate is detected, a reference count is incremented instead of allocating new memory and re-reading the data. I opted for reference counting instead of simply returning a failure on duplicate detection because it is perfectly valid to have a title with such duplicates. I could easily imagine titles adding duplicates, then navigating around them.

3 5

[PATCH] re-enable builtin dlopen on mingw
by John Stebbins 05 Nov '12

05 Nov '12

Not all mingw build environments have libdl. And those that do often only have it in dll form which would have to be distributed with the application. So allowing use of the builtin version of dlopen in dvdread is useful. This patch does the right thing if dlopen is not found on the system and it adds a configure option to force use of the builtin dlopen. It also seems that building dlopen broke at some point. #includes pointing to the wrong places.

3 5

cross/mingw compilation (was assertion failure with certain DVD's)
by Roger Pack 18 Oct '12

18 Oct '12

>> $./configure2 >> ./configure2: line 95: cd: /usr/local/: No such file or directory # >> needs to be run in bash I believe > > Nope, you need to set the install prefix with --prefix=blabla Sweet that fixed it, as long as the prefix exists, otherwise same error message different path :) >> Checking if we're big-endian... ./configure2: line 113: mktemp: command not found This meant I didn't have msys installed, only mingw (my bad). >> Done, type 'make install' to build and install >> (appears to set HAVE_DLFCN_H in error) > > Without using autotools, dlfcn.h presence is not detected Is there a way to use autotools with libdvdread? I'm a bit lost, as is evident. though I can work around it for now. This actually makes me wonder how anybody is compiling it for their windows distros today. I know mplayer just uses the src directory and its own configs (which work), dunno about VLC/others. It seems like both msys and cross compiling don't work, or can somebody give me an example of how they do work, somehow? > ./msvc/config.h:#define HAVE_DLFCN_H 1 > > And it is assumed to exist (is this file made for Visual Studio?) I guess so, since it's under the msvc dir. >> Also as a note when it runs strip and ranlib, if it's being cross >> compiled, it would be helpful to call their equivalent: >> $ ./configure2 --cc=i586-mingw32msvc-gcc >> should eventually call i586-mingw32msvc-ranlib >> >> maybe it could use a new configure option for a prefix or what not. > > ./configure --host=i586-mingw32-msvc should work Yes I think it needs an option like that added (configure2 doesn't appear to have one as of yet). -roger-

2 11

[PATCH] detect if path is physical dvd
by John Stebbins 16 Oct '12

16 Oct '12

Hi Erik, We are evaluating using your github versions of dvdread and dvdnav in handbrake. So I've been having a close look at what patches you've accepted and what's still missing. I've got a few patches I'm going to through at you. The first was part of another patch that I submitted. You merged the rest but left this part out. It's purpose is to detect if an input path is a physical drive (on windows), and if so attempt to open it as an image. The current code will only attempt to open the path as an image if the path contains no directory components (e.g. "X:" or "X:\"). So if the path instead is "X:\VIDEO_TS", it will fail to open it as an image and instead open it as a directory tree and css decryption will not work. This makes the behavior more similar to what happens on other platforms.

2 1

[PATCH] dvdnav configure option to disable examples
by John Stebbins 16 Oct '12

16 Oct '12

There was a time where dvdnav examples failed to build on mingw. Rather than fix them, we just disabled examples since we had no need for them. Although the examples seem to build fine now, this option may be useful in general.

2 1