Commit graph

7691 commits

Author SHA1 Message Date
Jan Alexander Steffens (heftig)
286a3829b6 rtmp2: Improve handling incoming set chunk/window size
Reject out-of-spec sizes and warn about suspiciously small sizes.
2020-02-21 15:20:41 +00:00
Jan Alexander Steffens (heftig)
14fd7e0884 rtmp2: Lock self->lock before OBJECT_LOCK
OBJECT_LOCK is used to protect property access only. self->lock is
used to access the RtmpConnection, mostly between the streaming thread
and the loop thread.

To avoid deadlocks involving these two locks, we obey a lock order:
If both self->lock and OBJECT_LOCK are needed, self->lock must be locked
first. Clarify this.
2020-02-21 15:20:41 +00:00
Jan Alexander Steffens (heftig)
6583e00d50 rtmp2: Reject oversized messages
We only have 24 bits for the size, so reject anything larger.
2020-02-21 15:20:41 +00:00
Jan Alexander Steffens (heftig)
0044e7a1ba rtmp2: Count in_bytes_acked instead of in_bytes_unacked
This is nicer for statistics.
2020-02-21 15:20:41 +00:00
Jan Alexander Steffens (heftig)
11a1de0053 rtmp2: rtmpconnection: Use more appropriate size types
- guint32 for chunk size and window size
- guint64 for running counters
2020-02-21 15:20:41 +00:00
Jan Alexander Steffens (heftig)
279e3c333c rtmp2: Add a g_return_val_if_fail 2020-02-21 15:20:41 +00:00
Jan Alexander Steffens (heftig)
03c3257f0f rtmp2: Replace explicit unref with g_main_context_invoke_full 2020-02-21 15:20:41 +00:00
Jan Alexander Steffens (heftig)
baad4fd91b rtmp2: rtmpconnection: Use GST_*_OBJECT logging
GstRtmpConnection isn't a GstObject with a name or path, but we still
get the GObject's type and address.
2020-02-21 15:20:41 +00:00
Marc Leeman
424c593871 rist: fix two minor memory leaks 2020-02-21 12:16:31 +01:00
Marc Leeman
6da6b6f3f0 rtpmanagerbad: fix two minor memory leaks 2020-02-21 12:16:28 +01:00
Marc Leeman
a710fbc12b rtpmanagerbad: reduce lock in rtpsink 2020-02-21 12:16:21 +01:00
Marc Leeman
61b062a12e rtpmanagerbad: documentation comment fix 2020-02-21 12:16:17 +01:00
Vivia Nikolaidou
3df3c3c5f6 tsparse: Add split-on-rai property
If set, buffers sized smaller than the alignment will be sent so that
RAI packets are at the start of a new buffer.

Fixes: #1190
2020-02-11 10:56:54 +00:00
Thiago Santos
0a128155b3 sdpdemux: check if connections are available on media entry before get
Otherwise we trigger an assert.
2020-02-02 22:15:40 +00:00
Vivia Nikolaidou
8d522bf3e6 mpegtsparse: Moved dispose function into finalize
dispose can be called several times and would double-free the flow
combiner in that case.
2020-01-29 20:39:44 +00:00
Vivia Nikolaidou
0d2e908523 mpegtsparse: Added alignment property
alignment works like in mpegtsmux, joining several MpegTS packets into
one buffer. Default value of 0 joins as many as possible for each
incoming buffer, to optimise CPU usage.
2020-01-29 20:39:44 +00:00
Vivia Nikolaidou
2f946274d5 mpegtsparse: Set delta unit flag on non-random-access buffers
If they don't have the random access flag set, they cannot be decoded
independently.
2020-01-29 20:39:44 +00:00
Vivia Nikolaidou
34af8ed66a mpegtsparse: Packetize output on default srcpad
Align buffer boundaries with mpeg-ts packets, instead of keeping
whatever packetization we have from the source (network, file reading).
2020-01-29 20:39:44 +00:00
Vivia Nikolaidou
e44cbfb1da mpegtsparse: Factor common code into mpegts_packet_to_buffer
The same code was used twice for turning an MpegTSPacketizerPacket into
a GstBuffer.
2020-01-29 20:39:44 +00:00
Vivia Nikolaidou
68f69d419b mpegtspacketizer: Fix typo in flag name 2020-01-29 20:39:44 +00:00
Sebastian Dröge
287f9b18b0 mpegdemux: Update the last_ts correctly if we have no DTS
If we have no DTS but a PTS then this means both are the same, and we
should update the last_ts with the PTS. Only if both are unknown then we
don't know the current position and should not update it at all.

Previously we would always update the last_ts to GST_CLOCK_TIME_NONE if
the DTS is unknown, which caused the position to jump around and to
cause spurious gap events to be sent.
2020-01-21 23:50:52 +00:00
Sebastian Dröge
5f95a9ec61 mpegpsdemux: Send gap events for late streams whenever updating the SCR
Instead of doing it on each packet and doing it based on the distance to
the previous SCR instead of based on the DTS.

Previously we would send gap events for audio all the time if the SCR
distance was 400ms because the threshold for audio is 300ms and by only
ever updating the position when the SCR updates we would always be 100ms
above the threshold and send needless gap events.

This fixes audio glitches on various files caused by gap events.
2020-01-21 10:08:53 +00:00
Jan Schmidt
e2bdc0c48d yadif: Re-renable MMX asm on x86_64 with meson
The meson build doesn't automatically set HAVE_CPU_* defines
like autotools did, so the yadif plugin was being built without
the MMX assembler support
2020-01-19 08:50:19 +00:00
Jan Schmidt
1986d4f942 yadif: Only build inline Asm with gcc/clang 2020-01-19 08:50:19 +00:00
Tim-Philipp Müller
415c798b73 mxfdemux: add support for Apple ProRes 2020-01-15 11:51:20 +00:00
Sebastian Dröge
f72aaed9c7 timecodestamper: Add property to set the extra latency to introduce for waiting for LTC timecodes
Default to 150ms instead of 8 frames, which seems to work in the
majority of cases.
2020-01-13 18:17:23 +00:00
Sebastian Dröge
fdca7ebb4c timecodestamper: Add some more debug output 2020-01-13 18:17:23 +00:00
Josep Torra
bebf20c906 h264parse: do not push wrong PTS with some raw files
Some raw h264 encoded files trigger the assignment of wrong PTS to buffers
when some SEI data is provided. This change prevents it to happen.

Also ensure this behavior is being tested.
2020-01-10 15:03:38 +00:00
Sebastian Dröge
a4c925f694 timecodestamper: Skip over invalid LTC timecodes immediately 2020-01-10 15:59:27 +02:00
Sebastian Dröge
a1443518e0 timecodestamper: Clean up old LTC timecodes on LTC discontinuity
We might have some old timecodes that are in the future now and have to
drop those to make sure that our queue is correctly ordered and we don't
have multiple timecodes for the same running time.
2020-01-10 15:59:26 +02:00
Sebastian Dröge
bbdb392abe timecodestamper: Fix waiting for the first video frame in case of live video input 2020-01-10 15:59:25 +02:00
Sebastian Dröge
d7bb5b8a16 timecodestamper: Fix up handling/queueing of LTC timecodes
Directly read them out of the decoder as soon as we passed audio and
then store them in a queue that we handle internally together with their
timestamps. This cleans up memory management and gives us proper control
over the queue instead of guessing how the queue inside the LTC decoder
actually works and when it overflows.
2020-01-10 15:59:24 +02:00
Sebastian Dröge
0a53f6560a timecodestamper: Only allow requesting LTC audio pad in NULL/READY states
And don't introduce any latency at all if not LTC audio pad was
requested.
2020-01-10 15:59:21 +02:00
Sebastian Dröge
0a499242e9 timecodestamper: In live mode wait correctly for the latency to pass
And also introduce 6 instead of 2 frames of latency compared to the LTC
audio input as that seems to be an upper bound for how much the LTC
library is lagging behind.
2020-01-10 15:58:29 +02:00
Sebastian Dröge
31d7862051 timecodestamper: Use the internal LTC timecode tracker instead of the last one we retrieved
Otherwise we don't interpolate between LTC timecodes but only ever put
an LTC timecode on buffers once we actually received one.
2020-01-10 15:58:06 +02:00
Stéphane Cerveau
4b8c47ee37 h26xparse: Handle state change on IDR first slice
As the H265/H264 bitstream can support multiple slices,
mastering_display_info_state and content_light_level_state
should be changed only on first slice segment.

Fix #1152
2020-01-07 08:55:28 +00:00
Stéphane Cerveau
d414e90eff h265parse: use same algo for MDCV and CLL SEI management 2020-01-07 08:55:28 +00:00
Stéphane Cerveau
b481edd745 h264parser: add MDCV and CLL SEI message parsing
Allow to parse SEI message for:
- mastering display colour volume
- Light level infomation

Set to caps if necessary.

Fix #958
2020-01-07 08:55:28 +00:00
Thibault Saunier
1e5c117c7c fakevideosink: Use our pad template to create pad 2020-01-06 20:35:00 +00:00
Mark Nauwelaerts
42b60627fa mpegtsdemux: resurrect actual and efficient seeking of all kinds
... by seeking to target offset determined by new seek segment,
rather than that of the previous segment.  The latter would typically
seek back to start for a non-accurate seek, and lead to a lot
of skipping in case of an accurate seek.
2019-12-31 10:44:35 +01:00
Stéphane Cerveau
add7878e14 bad: use of g_value_dup_string
Use helper method to get string from GValue.
2019-12-30 14:13:03 +00:00
Sebastian Dröge
0dc783d719 timecodestamper: Refactor LTC audio waiting and properly handle live inputs
If one of the inputs is live, add a latency of 2 frames to the video
stream and wait on the clock for that much time to pass to allow for the
LTC audio to be ahead.

In case of live LTC, don't do any waiting but only ensure that we don't
overflow the LTC queue.

Also in non-live LTC audio mode, flush too old items from the LTC queue
if the video is actually ahead instead of potentially waiting forever.
This could've happened if there was a bigger gap in the video stream.
2019-12-30 09:36:23 +00:00
Yeongjin Jeong
3f2240498b h265parser: Add simple GstH265Profile/string public utilites
It makes more simplifies the conversion between GstH265Profile and string.
2019-12-20 15:43:55 +00:00
Nicolas Dufresne
416728f213 autoconvert: Fix lock-less exchange or free condition
Before this change, we would free the list we just have saved.

Fixes #1158
2019-12-19 22:35:18 +00:00
Stéphane Cerveau
c6eb17be6e h264parse: Align GST_H264_PROFILE_HIGH_422 to H264 standards
According to H264 ITU standards from 06/19, GST_H264_PROFILE_HIGH_422
(profile_idc = 122) with constraint_set1_flag = 0 and
constraint_set3_flag = 0 can be mapped to high-4:2:2 or high-4:4:4.
GST_H264_PROFILE_HIGH_422 with constraint_set1_flag = 0 and
constraint_set3_flag = 1 can be mapped to high-4:2:2, high-4:4:4,
high-4:2:2-intra or high-4:4:4-intra.
2019-12-18 03:03:40 +00:00
Olivier Crête
1f766a7145 Revert "videoparseutils: support two new EIA 608 closed caption formats"
This reverts commit f5c1c90122.
2019-12-17 16:44:10 -05:00
Aaron Boxer
f5c1c90122 videoparseutils: support two new EIA 608 closed caption formats 2019-12-17 18:26:35 +00:00
Stéphane Cerveau
6bc0e9527e remove various useless linefeed in logs 2019-12-11 10:51:29 +01:00
Alicia Boya García
f816903e65 gsttestsrcbin: Avoid not-linked errors when switching tracks
The previous implementation had a very high reproducibility race where
if after a track switch, the ex-active track pad completed a buffer
chain (now returning not-linked) the flow combiner had all their pads in
non-linked state, propagating it as an error and stopping the pipeline.

By resetting the flow combiner in response to RECONFIGURE events that
race is made impossible.
2019-12-09 18:12:29 +01:00
Sebastian Dröge
812d593c4e interlace: Store unsigned integers in unsigned integer types
And add some assertions to guard against overflows and out of bounds
reads.
2019-12-03 21:12:26 +00:00
Sebastian Dröge
c67146b27a interlace: Increment phase_index before checking if we're at the end of the phase
Incrementing it afterwards will always have to phase_index >= 1 and we
will never be at the beginning (0) of the phase again, and thus never
reset timestamp tracking accordingly.

This was broken in bea13ef43b in 2010, and
causes interlace to run into integer overflows after 2^31 frames or
about 5 hours at 29.97fps. Due to usage of wrong types for the integers
this then causes negative numbers to be used in calculations and all
calculations spectacularly fail, leading to all following buffers to
have the timestamp of the first buffer minus one nanosecond.
2019-12-03 21:12:26 +00:00
Jan Alexander Steffens (heftig)
fd6c51b2e7
rtmp2sink: Only apply @setDataFrame to onMetaData messages
Only the metadata needs to be made "sticky". Custom data messages should
be passed on unmodified.

https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/merge_requests/878
2019-12-03 14:11:47 +01:00
Jan Alexander Steffens (heftig)
042e439829
rtmp2: Add gst_rtmp_message_is_metadata
https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/merge_requests/878
2019-12-03 14:11:47 +01:00
Jan Alexander Steffens (heftig)
e07a1bb48f
rtmp2: Add gst_rtmp_connection_set_data_frame
https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/merge_requests/878
2019-12-03 14:11:47 +01:00
Jan Alexander Steffens (heftig)
8f1ae04ac5
rtmp2: Add single-value AMF0 parsing and serializing
https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/merge_requests/878
2019-12-03 14:11:47 +01:00
Jan Alexander Steffens (heftig)
f5b068b26c
rtmp2: Minor changes
- Remove an unneeded initialization to zero from AmfParser
- Add missing initialization to gst_amf_serialize_command_valist
- Add a g_return_if_fail to gst_rtmp_connection_request_window_size

https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/merge_requests/878
2019-12-03 14:11:46 +01:00
Edward Hervey
d8a51c6097 atscmux: Add missing break in switch
CID: 1455515
2019-11-27 15:41:26 +01:00
Aaron Boxer
e3297be433 h264parse: buffer mismatch in map/unmap 2019-11-26 13:07:47 -05:00
Marc Leeman
31861b095a rtpmanagerbad: allow setting multicast-iface
Allowing the UDP elements to bind on an interface is needed in more
complex networks where there are mutiple networks interfaces without
default gateway
2019-11-19 12:39:59 +00:00
Sebastian Dröge
c7393af0bf timecodestamper: Create LTC sink pad with the correct name according to the template
Should be "ltc_sink" and not just "ltc"
2019-11-19 12:21:30 +02:00
Vivia Nikolaidou
851682629e videoparsers: Disable gst_base_parse_set_infer_ts
From the documentation of gst_base_parse_set_infer_ts, it should be
disabled for non-audio data. Currently just disabling for all video
parsers that have reordered data: h264, h265, mpeg, mpeg4, vc1. Was
already disabled in h263.
2019-11-19 10:23:31 +02:00
Kyrylo Polezhaiev
6af38c6ffe tsmux: Fix copying of buffer region 2019-11-18 07:31:33 +00:00
Marc Leeman
1569c33f24 rtpmanagerbad: name the element children
As discussed with RIST, it is best to name the children of the elements
since these are now created at the element initialisation.
2019-11-17 16:00:19 +00:00
Vivia Nikolaidou
413c6ec57b tsdemux: Always issue a DTS even when it's equal to PTS
Currently tsdemux timestamps only the PTS, and only issues the DTS if
it's different. In that case, parsers tend to estimate the next DTS
based on the previous DTS and the duration, which can accumulate
rounding errors.
2019-11-14 12:16:04 +00:00
Jan Schmidt
c94d50090b switchbin: Free path objects on finalize
Clean up path objects nicely when shutting down,
first by dropping pointers to elements during dispose,
and then by making sure to drop the ref to the path object
when finalizing the switch bin.

Fixes valgrind checks in the unit test.
2019-11-13 10:15:32 +00:00
Jan Schmidt
ed63012d70 switchbin: Add current-path property
Returns the index of the currently selected processing
path, or MAX-UINT if none
2019-11-13 10:15:32 +00:00
Jan Schmidt
6d292c86e9 switchbin: Add docs
Add documentation clauses and enrol switchbin to generate
plugin docs
2019-11-13 10:15:32 +00:00
Jan Schmidt
e367258eef switchbin: Initial checkin
Add code from Stream Unlimited implementing a bin
which switches between different internal decoding/processing
chains based on input caps
2019-11-13 10:15:31 +00:00
Jan Alexander Steffens (heftig)
346bca80af
rtmp2: Fix NULL check in gst_rtmp_meta_transform
Coverity rightly complains that checking a pointer for NULL after
dereferencing it is pointless.

Remove the check, and to be safe, assert that gst_buffer_add_meta
returns non-NULL.

CID 1455485
2019-11-12 12:20:34 +01:00
Jan Alexander Steffens (heftig)
f730f4a694
rtmp2: Check for missing GstRtmpMeta
The message buffers are created using `gst_rtmp_message_new` and thus
always contain a GstRtmpMeta. Add checks to appease Coverity's static
analysis.

CID 1455596
CID 1455384
2019-11-12 12:20:30 +01:00
Vivia Nikolaidou
8d7489a734 rtmp2sink: Add a check that meta isn't NULL before accessing
It really can't be NULL, this is just to convince coverity

CID 1455553
2019-11-12 12:36:38 +02:00
Nicolas Dufresne
44322b1dfc vc1parse: Avoid division by zero assertion
A framerate of 0/1 is valid, but we cannot calculate the frame duration
in this context. Simply protect against this case.

Related to #660
2019-11-11 16:23:18 -05:00
Nicolas Dufresne
a5113fe8c8 vc1parser: Relax ASF Binding Byte validation
According to the spec, the least significant bit is reserved and should
always we set to 1. Though, some wrong file has been found. Considering
how low important this reserved bit is, relax the validation.

Related to #660
2019-11-11 16:22:54 -05:00
Edward Hervey
7bceb6c3ff bad: Avoid using deprecated API
GTimeval is deprecated
2019-11-08 10:43:08 +00:00
Edward Hervey
60cec38591 tsdemux: Handle continuity mismatch in more cases
Packets of a given PID are meant to have sequential continuity counters
(modulo 16). If there are not sequential, this is the sign of a broken
stream, which we then consider as a discontinuity.

But if that new packet is a frame start (PUSI is true), then we can resume
from that packet without any damage.
2019-11-07 09:17:25 +00:00
Ederson de Souza
fe8e2a001c debugutils: clockselect, a pipeline that enables clock selection
Sometimes, one wants to force a clock on some pipelines - for instance,
when testing TSN related pipelines, one usually uses GstPtpClock or
CLOCK_REALTIME (assuming system realtime clock is in sync with network
one). Until now, one needs to write an application for that - not
difficult, but quite boring if one just wants to test something. This
patch presents a new element to help that: clockselect.

clockselect is a pipeline with two properties to select a clock. One
property, "clock-id", enables one to choose between "monotonic",
"realtime", "ptp" or "default" clock - where default keeps pipeline
behaviour of choosing a clock based on its elements. The other property,
"ptp-domain" gives one the choice of which PTP domain should be used.

Some very simple tests also added for this new element.
2019-11-06 08:58:53 -08:00
Niels De Graef
d8f61515d8 Don't pass default GLib marshallers for signals
By passing NULL to `g_signal_new` instead of a marshaller, GLib will
actually internally optimize the signal (if the marshaller is available
in GLib itself) by also setting the valist marshaller. This makes the
signal emission a bit more performant than the regular marshalling,
which still needs to box into `GValue` and call libffi in case of a
generic marshaller.

Note that for custom marshallers, one would use
`g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-11-06 14:27:46 +00:00
Michael Olbrich
58479bec37 sdp: don't leak the ghost pad
The peer is already gone when pad_removed_cb() called, so the ghost cannot
be removed. Use g_object_set_data() instead to remember the ghost pad.

Copied from similar code in GstRTPBin.
2019-11-06 02:11:20 +00:00
Aaron Boxer
6892078b00 basetsmux: allow null J2K profile
Since we are not requiring that profile equals GST_JPEG2000_PARSE_PROFILE_BC_SINGLE,
(as the standard requires) we can allow profile to be null. We relax this condition because
OpenJPEG can't create broadcast profiles.
2019-11-05 21:21:51 +00:00
Aaron Boxer
1414e58dfa jpeg2000parse: fail caps negotiation if caps are NOT fixed 2019-11-05 21:21:51 +00:00
Aaron Boxer
7b3491adf7 jpeg2000parse: use pre_push_frame to reset parser 2019-11-05 21:21:51 +00:00
Aaron Boxer
bfee115d66 jpeg2000parse: parse_event: call base class at end
derived class should do it's work first before calling base
2019-11-05 21:21:51 +00:00
Aaron Boxer
67cffd70ad jpeg2000parse: do hard reset if gst_base_parse_finish_frame fails 2019-11-05 21:21:51 +00:00
Aaron Boxer
a35157debf jpeg2000parse: initialize some variables to make valgrind happy 2019-11-05 21:21:51 +00:00
Aaron Boxer
fae0664824 jpeg2000parse: use GST_INT in caps for profile
Negotiation failed with GST_UINT
2019-11-05 21:21:51 +00:00
Aaron Boxer
1344d9f560 jpeg2000parse: make explicit that codec_format is for src caps 2019-11-05 21:21:51 +00:00
Aaron Boxer
969e30c035 jpeg2000parse: refactor
1. only recalculate src codec format if sink caps change
2. use correct value for "jp2c" magic in J2C box ID
3. only parse J2K magic once, and store result
4. more sanity checks comparing caps to parsed codec
2019-11-05 21:21:51 +00:00
Aaron Boxer
2b6b1a2b04 jpeg2000parse: set parsed to TRUE in src caps 2019-11-05 21:21:51 +00:00
Aaron Boxer
453a65b8e9 jpeg2000parse: only cache caps parameters when caps have in fact changed 2019-11-05 21:21:51 +00:00
Aaron Boxer
ead5dba3ac jpeg2000parse: fix typos in media format 2019-11-05 21:21:51 +00:00
Aaron Boxer
88efbd2344 jpeg2000parse: add reset method
Also add three new struct members, currently unused.
2019-11-05 21:21:51 +00:00
Aaron Boxer
6d3429af34 documentation: fixed a heap o' typos 2019-11-05 09:11:25 -05:00
Vivia Nikolaidou
2386858a91 Add files from gst-rtmp
For master, without autotools.
2019-11-05 13:52:55 +00:00
Olivier Crête
3a9c224ac2 ristsrc: Apply BINDTODEVICE to socket created by udpsrc too 2019-11-04 20:47:23 +00:00
Yeongjin Jeong
732735cab0 h265parse: Fix wrong NALU minimum length check
Fixes a problem where an EOS/EOB NALU placed at the end of
an AU is detected as an other AU and create a buffer that
does not have valid pts.
2019-11-04 14:16:49 +09:00
Yeongjin Jeong
0318458b0e h264parse: Fix wrong NALU minimum length check
Fixes a problem where an EOS/EOB NALU placed at the end of
an AU is detected as an other AU and create a buffer that
does not have valid pts.
2019-11-04 14:16:49 +09:00
Edward Hervey
ef16d7558f mpegtsmux: Add SCTE-35 support
This adds two properties:
* scte-35-pid: If not 0, enables the SCTE-35 support for the current
  program. This will write the proper PMT and send SCTE-35 NULL
  commands (i.e. heartbeats) at a regular interval
* scte-35-null-interval: This specifies the interval at which the
  NULL commands should be sent

Sending SCTE-35 commands is done by creating the appropriate SCTE-35
GstMpegtsSection and then sending them on the muxer. See the
associated example
2019-10-31 12:31:27 +00:00
Edward Hervey
6a9108884c mpegts: Add support for SCTE-35 sections
Not all commands are supported, but the most common ones are.
Both parsing and packetizing is supported
2019-10-31 12:31:27 +00:00
Edward Hervey
5464775aef tsmux: Disable bluray-isms from PMT
We were unconditionally adding top-level descriptors in the PMT which
were only related to bluray support for PS3 (from 10 years ago).

These should be re-added conditionally
2019-10-31 12:31:27 +00:00
Edward Hervey
878edacc05 mpegtspacketizer: Fix off-by-one error
This went un-noticed for 6 years :( The issue is that for short
sections (without subtables and CRC), we would always fail when
checking whether we had enough data or not and then default to the
long section checking.

Use the long section checking would then cause interesting side-effects
for short sections (such as believing they were already seen and therefore
would be dropped/ignored).
2019-10-31 12:31:27 +00:00