Bug 175036
| Summary: | shell escape reads from stdin even while in pdb | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Patrick Mansfield <patmans> |
| Component: | anaconda | Assignee: | David Cantrell <dcantrell> |
| Status: | CLOSED RAWHIDE | QA Contact: | Mike McLean <mikem> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | rawhide | ||
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2006-01-31 20:44:21 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
|
Description
Patrick Mansfield
2005-12-05 21:05:23 UTC
I'm trying to reproduce this error here, but I need some more information. When/where are you pressing enter to spawn a shell? What python error is happening where you click debug to go to pdb? Log file(s) and/or screen shots could help me a bit. I'm working on additions to anaconda for iscsi, and hit a bug in my development
python code.
Any exception hit in python that causes us to enter pdb will have a problem. I
don't know if there other programs spawned by anaconda that read stdin, I doubt it.
I am hitting "enter" to input data to pdb, not to the shell, but the shell is
grabbing parts of the input.
Hopefully, this output explains any missing details. I finally typed something
like "bbtt<enter>enter" to get the backtrace at the end.
File "/usr/lib/anaconda/iw/iscsi_gui.py", line 60, in getNext
self.iscsi.startup()
File "/usr/lib/anaconda/iscsi.py", line 92, in startup
os.unlink(INITIATOR_FILE)
OSError: [Errno 2] No such file or directory: '/etc/initiatorname.iscsi'
> /usr/lib/anaconda/iscsi.py(92)startup()
-> os.unlink(INITIATOR_FILE)
(Pdb) bt
sh: bt: command not found
sh-3.00# *** NameError: name 't' is not defined
(Pdb) bt
sh: bt: command not found
sh-3.00# *** NameError: name 't' is not defined
(Pdb) btt
sh: btt: command not found
sh-3.00# (Pdb)
sh-3.00# (Pdb)
sh-3.00# (Pdb) bt
sh: bt: command not found
sh-3.00# /usr/bin/anaconda(1198)?()
-> intf.run(id, dispatch)
/usr/lib/anaconda/gui.py(822)run()
-> self.icw.run (self.runres)
/usr/lib/anaconda/gui.py(1386)run()
-> gtk.main()
/usr/lib/anaconda/gui.py(923)nextClicked()
-> return
/usr/lib/anaconda/iw/iscsi_gui.py(60)getNext()
-> self.iscsi.startup()
> /usr/lib/anaconda/iscsi.py(92)startup()
-> os.unlink(INITIATOR_FILE)
OK, I definitely see what you're talking about now. I've been trying to recreate this problem here using the latest rawhide anaconda and it's not happening. Can I see your iscsi_gui.py and associated changes to anaconda and see if I can trigger the problem that way? I have a whole set of changes, I can send them all, but I already fixed my problem(s). It would be easier to add a syntax error in something like packages.py:betaNagScreen() or such, that will allow you to immediately hit an exception and reproduce the problem. I assume you were planning on sending in the anaconda changes. Jeremy is probably a better contact for that, but feel free to send them to me via email. I did exactly what you suggested: threw in a syntax error, got the exception window, clicked debug, and was in pdb. But it worked fine. I was not getting the behavior you described. Which is why I was interested in seeing your changes. I don't want to close this until I see your changes because right now the only thing I have to go on is closing it as WORKSFORME. I have fixed this problem in rawhide, but introduced a new problem. Once you drop in to Pdb, keyboard echo is gone. I'm working on that and once it's finished, I'll close this bug. This comment is mostly for record keeping anyway. Fixed now. |