Move the two static functions to the top, so that forward prototyping
is not required.
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4092
Despite the command line, retain the "--" forms for switches, as
well as the documented "-" forms. Provide "-skew" as a shorter
synonym for the very-long-switch - hope that makes sense.
Make command line handling more robust.
Correct the error handling if an input file is not given.
Give the -dest switch a default port (0 seems most sensible).
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4091
In the process, tighten up command line argument handling a bit, and
improve some error messages.
Also, make the "magic number" 0x2000 a bit more obvious.
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4082
Update the documentation adding information about:
-x switch in esreport and
-cnt <pid> in tsreport
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4080
The ``-cnt`` switch is to be followed by a PID in decimal units.
It checks the values of the field continuity_counter in the specified PID,
warns in there are discontinuities and writes the values of continuity_counter in a file called
'continuity_counter.txt'. Using this automatically turns buffering on (-b).
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4079
If one wants to know about the presence of broken_link_flag, the thing to do is to use esdots -x.
All the recovery point SEI message are reported there (included the value of broken_link_flag)
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4078
lseek does indeed suggest inclusion of <sys/types.h>, as well as
<unistd.h>. Experimentation showed that ading <sys/types.h> didn't
hurt on Mac OS/X, so add both into compat.h -- hopefull this will
solve the bug.
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4077
h.264 RAP as I or I + recovery_point SEI
Two changes done:
1) esdots computes the duration of GOPs, even in the h.264 case. The change made is to consider an I frame without recovery_point SEI as beginning of a new GOP. This can be changed just setting rec_point_required = TRUE in esdots.c.
In this case a new GOP is started when we found I + recovery point SEI
2) nalunit has been modified so it's able to output the information about error recovery point SEI message. Just use 'esreport -x' and SEI recovery point messages will be reported with the same syntax as all the others.
e.g.
83644584/0036: NAL unit 0/6 (SEI) 6: 06 06 01 a4 80 00
@@ Recovery Point SEI
recovery_frame_cnt 0
exact_match_flag 0
broken_link_flag 1
changing_slice_group_idc 0
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4074
Want to use it? Find 'broken_link_flag is 1' and uncomment that part of code.
Than compile and use esdots.
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4073
build, and hopefully also in Linux. Ideally, do the same sort of thing for
the other "macro returning a constant string" cases.
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4070
startup, inspired by work done by Gareth. Still to be fully tested.
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4064
in (for instance) __repr__. The only snag is that I've not yet figured out
how to suppress the warning because the second argument to PyObject_AsReadBuffer
is meant to be "const void **", and I can't specify "const" in Pyrex. And
warnings are not nice...
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4063
Fix "if (err = EOF)", which should be a comparison.
Various pedantries (petty stuff):
* Remove compiler warning on use of build_TS_reader_with_fns - this was because
"char *" is signed on my Mac, whilst "byte *" is not. Fixed by making it
exepct "byte *", and whilst I'm at it make pcapreport.c use the datatypes from
compat.h -- not because they're better than the stdint.h type, because they
clearly aren't, but purely for consistancy (long term, I still want to move
to stdint types - it should be easier now as we don't need to support ancient
compilers))
* Reformat pcapreport.c (this is going to make for a big difference, oh well)
to standard tstools indentation patterns.
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4062
Utility to strip the timestamps off M2TS files and reorder the packets to present a normal TS file.
Known limitations: fails to detect discontinuities, and can produce bad (misordered) TS files as a result
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4060
way!). Next step is to analyse PCR against the pcap timestamps.
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4051
useful sources of data. This *may* help solve the issue of what test data to
use.
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4047
Use of ESOffset sorted out a bit more -- I think it now makes sense.
And since I can't find any way for doctest to "see" C stdout/stderr,
just make rundoctest.py output an explanatory message (and the RED/GREEN
thing just makes it a bit quicker to ignore worrying texts).
--HG--
extra : convert_revision : svn%3Aeff31bef-be4a-0410-a8fe-e47997df2690/trunk%4046
tibs