Bugs item #2882862, was opened at 2009-10-20 20:50
Message generated for change (Comment added) made by mpg123user
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=733194&aid=2882862&group_id=135704
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: mpg123
Group: 1.0.0
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Alex (mpg123user)
Assigned to: Nobody/Anonymous (nobody)
Summary: Failure on Streaming Site
Initial Comment:
The command,
mpg123 -@ http://www.wfuv.org/audio/wfuvmp3high.pls
works on versions < 1.0.0 (tested with 0.65 and 0.68)
and fails either with "segmentation fault" or after a very short "introductory" audio
on versions >= 1.0.0 (tested with 1.0.0, 1.3.0, 1.5.1, 1.9.0 and 1.9.1)
System: BLFS, kernel 2.6.31.4 (with ALSA 1.20)
Any details, clarifications, etc. available upon request.
-- Alex
----------------------------------------------------------------------
>Comment By: Alex (mpg123user)
Date: 2009-10-23 17:33
Message:
Hi sobukus:
Note: the previous "nobody" (2009-10-23 16:29) was me (not logged in).
Complicated.
A strange thing:
It might depend on the hour - At this point in time
all the versions (see full list below) including 1.9.1
work _OK_ (the "20091023000000" kind below):
160, 161, 162, 164, 170, 171, 172, 173, 180, 181, 190 and 191.
The irony, the BLFS Devel. version, 1.5.1, is the highest with still
"segmentation fault"!
I supposed, to be continued!
----------------------------------------------------------------------
Comment By: Nobody/Anonymous (nobody)
Date: 2009-10-23 16:29
Message:
Hi sobukus:
1. 1.6.3 and 1.8.1 FAIL the "1.9.1" kind (see my 2009-10-21 01:20 post)
2. "20091023000000" works OK (as you predicted):
Several
ICY-META: StreamTitle='';
[with "Welcome WFUV ..." introduction]
Note: Illegal Audio-MPEG-Header 0x27c53050 at offset 0x24f04.
Note: Trying to resync...
Note: Skipped 72 bytes in input.
followed by one line
ICY-META: StreamTitle='';StreamUrl='';
[with Station Sound OK]
3. Crash analysis.
I haven't done anything here yet.
As a note, I assume you want me to use a version with
failure of the "1.0.0" kind.
I'll pick a version as high (i.e., close to 1.9.1) as possible.
-- Alex
----------------------------------------------------------------------
Comment By: Alex (mpg123user)
Date: 2009-10-22 21:39
Message:
Hi sobukus:
Thanks for your helpful comments.
> For the crashes with mpg123 < 1.9:
> Could you tell me what exact version of mpg123 you tried?
As I noted in the submission:
"tested [and failed] with 1.0.0, 1.3.0, 1.5.1, 1.9.0 and 1.9.1"
> I cannot reproduce with 1.8.1 or 1.6.3i
As fate (and probability) would have it, these two are not on the list!
That goes to show that if you're not 100% thorough, simple spot checks
would never protect you from law suits for defamation :)
BTW, I picked 1.5.1 as it was the version included in the latest BLFS
Development book.
I'll be at the subject machine only tomorrow at the earliest.
When I get there, I'll do the following and report:
1. Test 1.6.3i and 1.8.1 (I got your hint: they will work)
2. Test the subversion (mpg123.org/snapshot)
3.
> compile mpg123 with -g in your CFLAGS, run 'ulimit -c 0' in the shell
> before starting mpg123 and when it crashed, verify that you have a file
> named "core" in the current directory and run 'gdb /path/to/mpg123
core',
> enter 'bt' and show me the output of this ...
> Can you collect a backtrace of the crash
> after building mpg123 with debugging symbols?
Try the above. It so happens,I'm sporting a brand new, shiny gdb
version (7.0).
Thanks again for your interest and help,
-- Alex
----------------------------------------------------------------------
Comment By: Thomas Orgis (sobukus)
Date: 2009-10-22 03:43
Message:
Ah, there is it again... the dreaded premature ending because the stream
starts with a piece of mp3 that carries the information that the stream
ends after some seconds.
I removed the code that triggered the controlled ending now, letting
mpg123 stumble over the end when it cannot read anything anymore... the
transition from the intro to the real music is still a bit bumpy (with
on-screen messages at least), because mpg123 complains a bit that the file
is not clean, which it really is not: It's different files cat'ed
together.
We need to ensure that the current state is really safe... so it's just in
subversion for now, please check out the latest http://mpg123.org/snapshot
.
For the crashes with mpg123 < 1.9: Could you tell me what exact version of
mpg123 you tried? I cannot reproduce with 1.8.1 or 1.6.3 . Also, it would
be helpful (in case it really happens only on your setup...) if you
compiled mpg123 with -g in your CFLAGS, run 'ulimit -c 0' in the shell
before starting mpg123 and when it crashed, verify that you have a file
named "core" in the current directory and run 'gdb /path/to/mpg123 core',
enter 'bt' and show me the output of this... In other words, since you use
BLFS, that might be more clear: Can you collect a backtrace of the crash
after building mpg123 with debugging symbols?
Even when current mpg123 does not feature the crash, I want to know what
the problem was... to make sure it's gone for real˙
----------------------------------------------------------------------
Comment By: Alex (mpg123user)
Date: 2009-10-20 21:20
Message:
- Correction:
ALSA 1.0.20
- Additions:
Failures are of two kinds:
1. The "1.0.0" kind:
mpg123[1518]: segfault at ffffffff ip bf80df50 sp b7469831 error 6
Segmentation fault
2. The "1.9.1" kind:
On invokation you get several lines like
ICY-META: StreamTitle='';
accompanied by an "introductory" speech
followed by the "exit" (last) line,
[0:09] Decoding of onair-hi finished
-- Alex
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=733194&aid=2882862&group_id=135704
|