Issue 6 in wy60: wy60 hangs on start

[wy...@googlecode.com]

Status: New
Owner: ----
Labels: Type-Defect Priority-Medium

New issue 6 by marcus.bointon: wy60 hangs on start
http://code.google.com/p/wy60/issues/detail?id=6

What steps will reproduce the problem?
1. in a terminal window type 'wy60'

What is the expected output?
launches an interactive shell via the wy60 wrapper

What do you see instead?
Nothing, terminal session becomes completely unresponsive to ctrl-c, z, y
etc. If I close the
terminal, I see there are two wy60 processes left running which don't
respond to TERM, so I have to
kill -9.

What version of the product are you using? On what operating system?
2.0.9 compiled from source on Mac OS X 10.6.2

Please provide any additional information below.
Compilation seemed to go fine. wy60 -h and -v work fine, it's defintely a
runnable app. I tried
running it form inside Terminal.app and iTerm with same results. $TERM is
set to 'linux', shell is
zsh.

--
You received this message because you are listed in the owner
or CC fields of this issue, or because you starred this issue.
You may adjust your issue notification preferences at:
http://code.google.com/hosting/settings

[wy...@googlecode.com]

Comment #1 on issue 6 by zod...@gmail.com: wy60 hangs on start
http://code.google.com/p/wy60/issues/detail?id=6

Hmm, I don't have access to a Mac. So, this is a little hard for me to
test. The code
runs fine on other UNIX operating systems. And at some point a user
contributed the
code changes to make it run on OS X. So, at least at one point it was
working fine on
OS X.

I don't think wy60 has changed significantly since then. But maybe, OS X
has changed
in the intervening years?

Can you do any more debugging? Is there such a thing as "strace" or "truss"
which
shows all the system calls that a program makes? That's often a good
starting point
in getting a sense of where the application might be hanging.

[wy...@googlecode.com]

Comment #2 on issue 6 by marcus.bointon: wy60 hangs on start
http://code.google.com/p/wy60/issues/detail?id=6

I found dtruss (apparently like truss, but built using dtrace) which
produced the attached output.

Attachments:
wydump.txt 0 bytes

[wy...@googlecode.com]

Comment #3 on issue 6 by zod...@gmail.com: wy60 hangs on start
http://code.google.com/p/wy60/issues/detail?id=6

Something went wrong. The attachment is zero bytes long.

[wy...@googlecode.com]

Comment #4 on issue 6 by marcus.bointon: wy60 hangs on start
http://code.google.com/p/wy60/issues/detail?id=6

Take 2

Attachments:
wydump.txt 11.4 KB

[wy...@googlecode.com]

Comment #5 on issue 6 by zod...@gmail.com: wy60 hangs on start
http://code.google.com/p/wy60/issues/detail?id=6

The log file looks a little strange and ends rather surprisingly. So, I am
somewhat
puzzled what is going on.

But I did see that "wy60" tries to print "Cannot determine terminal size".
It should
then immediately have terminated. I am not sure why the call to "exit()"
didn't seem
to work.

In any case, the problem with determining the terminal size appears to be
the root
cause for your difficulties. You need to figure out why the stdout file
descriptor
for wy60 does not appear to point to a valid terminal. Make sure you are
running
within a text terminal. Also, make sure you do not redirect stdout.

Of course, it is also possible that dtruss redirects our stdout. That would
be
unfortunate, because that means we would be looking at a completely
different code
path. Make sure that you can rule this out.

If that still doesn't help, see if you can figure out what errno code 25
means.
That's the error that we get when calling "ioctl(1, TIOCGWINSZ, &win)".
Maybe, that
error code will help figure out what's wrong.

[wy...@googlecode.com]

Comment #6 on issue 6 by marcus.bointon: wy60 hangs on start
http://code.google.com/p/wy60/issues/detail?id=6

That's my fault - it generated a fair amount of output so I redirected
output to a file which probably caused the
problem. Without the redirect, the output is simlar, but much less
readable. It ends with an Err#9 and the Err#25
does not occur. There are a couple of Err#2s when it fails to find config
files, but I guessed they didn't matter.
Anyway, new dump attached, hope it makes more sense.
I've had a look around but I can't find a list of error codes for dtruss or
dtrace. I'd guess that the error codes are
related to the OS call rather than the tracer, but I don't know where to
look for docs on that.

Attachments:
wydump.txt 16.9 KB

[wy...@googlecode.com]

Comment #7 on issue 6 by zod...@gmail.com: wy60 hangs on start
http://code.google.com/p/wy60/issues/detail?id=6

I had a chance to look over the log file, and as far as I can tell,
everything looks normal as far as the
logged data is concerned.

The log file only shows the log for the parent process. And it
unfortunately does not decode all function
arguments.

So, I cannot tell with absolute certainty that we are doing everything
correctly. We might give the wrong
function arguments and I just cannot see that in this log file.

But more importantly, as mentioned earlier, the log file does not include
any information about child
processes. And that's likely where the problem is.

Does your tool have an option to descend into child processes? If so, that
would be a very useful option to
enable.

[wy...@googlecode.com]

Comment #8 on issue 6 by mesutt: wy60 hangs on start
http://code.google.com/p/wy60/issues/detail?id=6

same problem on 10.6.4

[wy...@googlecode.com]

Comment #9 on issue 6 by cathcart: wy60 hangs on start
http://code.google.com/p/wy60/issues/detail?id=6

I'm having the same issue on 10.5.8. I set the flag in dtruss for "follow
children as they are forked" and attached the results here. Please let me
know if I can be of any assistance.

Attachments:
dtruss_results.txt 11.3 KB