| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
newsize is a long, but storage is an int. This means the allocation
could succeed but storage would overflow.
Closes #2300
|
|
|
|
|
|
| |
Close the C4244 warning while compiling under windows with cmake+msbuild.
Signed-off-by: Tristan Matthews <tmatth@videolan.org>
|
|
|
|
|
|
|
|
| |
oy.data memory is alloc by ogg_sync_buffer(), but does not call free()
before main() exit. After fixing it, I test test_framing by valgrind.
And no more memory leak in test_framing and test_bitwise right now.
Signed-off-by: Ralph Giles <giles@thaumas.net>
|
|
|
|
|
|
|
| |
We call ogg_stream_init() in main() of framing.c, but no
ogg_stream_clear() in corresponding. It will cause memory leak.
Signed-off-by: Ralph Giles <giles@thaumas.net>
|
|
|
|
|
| |
Signed-off-by: Erik de Castro Lopo <erikd@mega-nerd.com>
Signed-off-by: Ralph Giles <giles@thaumas.net>
|
|
|
|
|
|
| |
Suggested-By Mark Harris
Fixes #2297
|
|
|
|
|
|
| |
Found via `codespell -q 3`
Signed-off-by: Erik de Castro Lopo <erikd@mega-nerd.com>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Subversion, like cvs, had a mechanism for replacing inline template
text on checkout for representing things like 'last modified date'.
Git does not support this, so remove the template strings from
file header comments.
Also less aggressive language is os_types.h.
|
|
|
|
| |
Useful for fuzzing
|
|
|
|
|
|
|
|
|
|
|
| |
On 64-bit Windows, sizeof(unsigned char *) is 8, sizeof(long) is 4 and
alignof(ogg_int64_t) is 8. This results in a 4-byte hole in ogg_packet
after "long e_o_s", which means ogg_packet structs cannot be reliably
compared by memcmp. Compare ogg_packet structs member-by-member instead.
This fixes `make check` for me on mingw-w64/GCC.
Signed-off-by: Ralph Giles <giles@thaumas.net>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A false continued packet occurs if the previous page ended at the
end of a packet, with no sequence number gap and no continued data
on that page, while the current page has the continued packet flag
set.
Previously, if you drained all of the buffered packet data by
repeatedly calling ogg_stream_packetout() after submitting the
previous page but before submitting the current page, libogg would
discard the continued data at the beggining of the current page.
However, if you left the previous page's data in the buffer, then
libogg would happily return the continued data at the start of the
current page as a separate packet when you eventually did call
ogg_stream_packetout().
This patch makes libogg consistently discard this data.
|
|
|
|
|
|
|
| |
Their size could overflow a long (especially on, e.g., Win64, where
they could still fit in memory).
svn path=/trunk/ogg/; revision=18758
|
|
|
|
| |
svn path=/trunk/ogg/; revision=18052
|
|
|
|
|
|
| |
Remove trailing whitespace and re-indent some comments.
svn path=/trunk/ogg/; revision=18051
|
|
|
|
|
|
| |
ogg_stream_flush will produce multiple pages when >4096 bytes and >4 packets are available.
svn path=/trunk/ogg/; revision=18044
|
|
|
|
|
|
|
| |
4
svn path=/trunk/ogg/; revision=17592
|
|
|
|
| |
svn path=/trunk/ogg/; revision=17579
|
|
|
|
| |
svn path=/trunk/ogg/; revision=17571
|
|
|
|
|
|
|
| |
mathing the ogg_packet declaration.
svn path=/trunk/ogg/; revision=17269
|
|
|
|
| |
svn path=/trunk/ogg/; revision=17124
|
|
|
|
|
|
| |
delay sensitive flushing to produce big pages. Increase the default minimum fill amount to 8 based on latency measurements with actual files. These changes may be controversial but since we've recently had a release I thought there would be no harm in getting them into the repository for discussion.
svn path=/trunk/ogg/; revision=17098
|
|
|
|
|
|
| |
Warnings reported by clang (the llvm C compiler).
svn path=/trunk/ogg/; revision=17060
|
|
|
|
| |
svn path=/trunk/ogg/; revision=17039
|
|
|
|
|
|
|
|
|
|
| |
1) Don't unneccessarily span pages.
2) Unless necessary, don't flush pages if there are less than four packets on
them; this expands page size to reduce unneccessary overhead if incoming
packets are large.
svn path=/trunk/ogg/; revision=16993
|
|
|
|
| |
svn path=/trunk/ogg/; revision=16051
|
|
|
|
|
|
|
|
|
| |
makes it safe to ignore most return codes in applications and simply
check occasionally to see if something went wrong at some earlier
point.
svn path=/trunk/ogg/; revision=16047
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
so much, autofoo)
Update ogg to check the return of all allocs for those on embedded
platforms with guarantees non-ovecommit. Be aware that these checks
are useless on any modern desktop OS, but that embedded folks with no
MMU and a hard heap boundary will benefit.
Add one new call to bitpacker that will query a write-mode
oggpack_buffer to see if it ran out of memory at some point; there was
no other way to add the error reporting in an ABI-compatible manner
(and it also allows the writepacker to avoid extra checks; it will
just keep going after running out of memory without faulting and state
can be queried later).
svn path=/trunk/ogg/; revision=16016
|
|
|
|
| |
svn path=/trunk/ogg/; revision=14463
|
|
|
|
|
|
|
| |
review.
svn path=/trunk/ogg/; revision=12446
|
|
|
|
|
|
|
|
|
|
|
|
| |
no packet ends. The spec is clear about this but the unit tests assumed
the old behaviour (carry over the previous granulepos) was correct.
I did not verify the new CRC values by hand however, they are just copied
from the library output.
Also update changelog for the bug fix.
svn path=/trunk/ogg/; revision=9601
|
|
|
|
|
|
|
|
|
| |
packets ending on it (which commonly happens with video keyframes, for
instance) must have a granulepos of -1.
libogg never implemented this correctly; this patch ensures it does.
svn path=/trunk/ogg/; revision=9598
|
|
|
|
| |
svn path=/trunk/ogg/; revision=9371
|
|
|
|
|
|
| |
tests.
svn path=/trunk/ogg/; revision=7835
|
|
|
|
| |
svn path=/trunk/ogg/; revision=7824
|
|
|
|
|
|
| |
leak-checking environments
svn path=/trunk/ogg/; revision=7586
|
|
|
|
|
|
|
| |
(patch from Colin Ward)
+ tested on linux/gcc
svn path=/trunk/ogg/; revision=7525
|
|
|
|
|
|
| |
problem.
svn path=/trunk/ogg/; revision=3964
|
|
|
|
| |
svn path=/trunk/ogg/; revision=3577
|
|
|
|
|
|
|
|
|
| |
functionality before required a clear/init cycle that involved
free()/malloc().
Monty
svn path=/trunk/ogg/; revision=3125
|
|
|
|
| |
svn path=/trunk/ogg/; revision=2867
|
|
|
|
| |
svn path=/trunk/ogg/; revision=2405
|
|
|
|
|
|
|
|
| |
should have been.
Monty
svn path=/trunk/ogg/; revision=2402
|
|
|
|
| |
svn path=/trunk/ogg/; revision=2397
|
|
|
|
|
|
| |
Forgot to check libogg when I was checking for these yesterday...
svn path=/trunk/ogg/; revision=2372
|
|
|
|
| |
svn path=/trunk/ogg/; revision=2106
|
|
|
|
| |
svn path=/trunk/ogg/; revision=1455
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
packetpeek can report -1 indicating a hole in the data (not really an
error).
Previously, packetout would have to be called to clear the hole; now,
the hole is cleared by packetpeek. That is, the change is that
ogg_stream_packetpeek must be followed by a call to
ogg_stream_packetout(state,NULL) to continue advancing forward only if
packetpeek returned a packet.
svn path=/trunk/ogg/; revision=1454
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
added ogg_stream_packetpeek; returns the next packet from the stream
stste without removing it from the stream.
ogg_stream_packetout can now be passed a NULL packet pointer; in this
case it removes the packet at the head of the stream and drops it on
the floor.
Monty
svn path=/trunk/ogg/; revision=1453
|
|
|
|
|
|
|
| |
change to BSD license
version update
svn path=/trunk/ogg/; revision=1333
|