Bug 728539
Summary: | [abrt] livecd-tools-16.3-1.fc16: kickstart.py:63:read_kickstart:KickstartError: Failed to parse kickstart file 'fedora-livecd-desktop.ks' : The following problem occurred on line 0 of the kickstart file: | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Mads Kiilerich <mads> |
Component: | livecd-tools | Assignee: | Brian Lane <bcl> |
Status: | CLOSED WONTFIX | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | unspecified | Docs Contact: | |
Priority: | unspecified | ||
Version: | 16 | CC: | adam.stokes, anton, bcl, bruno, dhuff, dvlasenk, iprikryl, Jasper.Hartline, jmoskovc, katzj, kklic, lmacken, mtoman, npajkovs |
Target Milestone: | --- | Keywords: | Reopened |
Target Release: | --- | ||
Hardware: | x86_64 | ||
OS: | Unspecified | ||
Whiteboard: | abrt_hash:6bcc1eca4bff647eb97384422da9fe98944ff8df | ||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2011-08-08 15:44:48 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
Mads Kiilerich
2011-08-05 13:05:20 UTC
abrt really shouldn't be catching errors like this. This is perfectly reasonable behavior for a python application that raises a variety of errors. If notabug in livecd-creator then it must be a bug in abrt? Re-opening and re-assigning to abrt. It might however be more of a packaging guidelines issue - it should perhaps be specified more clearly which application behaviour that is acceptable and what is bugs that should be reported and fixed? (FWIW, IMHO it is no more reasonable behaviour for a Python application to exit with an arbitrary exception than it is for a C application to exit with a segmentation fault.) Well, it's a matter of opinion, for me any unhandled exception is a bug and there is no way for ABRT to determine which exception is a real bug and which exception is just deliberately uncaught. On a second thought - it's never ok to consider exception as a proper exit from an application unless it's a situation which should never occur. Reassigning back to livecd-creator. It seems like there are different opinions on this, and unfortunately the opinions are incompatible and conflicting. IMO that is why have the packaging guidelines to clarify how we do in Fedora - and FESCo and the packaging committee that could and should be asked for clarification when the guidelines isn't sufficiently clear on how this issue shold be solved. (Re-reassigning to livecd-creator instead of liveusb-creator.) I went thru the code and this is what I found: except IOError, e: raise errors.KickstartError("Failed to read kickstart file " "'%s' : %s" % (path, e.strerror or e.args[0])) except kserrors.KickstartError, e: raise errors.KickstartError("Failed to parse kickstart file " "'%s' : %s" % (path, e)) can you tell me why it is better to raise an exception instead of just print ("Failed to parse kickstart file '%s' : %s" % (path, e)) if you don't want to catch it anyway? This just leads to printing traceback where error msg would be good enough and unnecessary scares users who thinks this is a bug. The traceback is being generated by the python-imgcreate library, so that isn't the right place to print errors. Maybe this should be caught higher up in livecd-creator, but then you run into a situation where you are having to try to catch all conceivable errors and convert them to prints. It seems more reasonable to me to let it raise the error, which includes both the easy to read output and a traceback to where the original error happened in the library. I think it is a mistake for abrt to be catching tracebacks in python code at all. If the python interpreter crashes, fine. But abrt doesn't (and very likely can't) provide any useful additional information to a traceback. (In reply to comment #8) > The traceback is being generated by the python-imgcreate library, so that isn't > the right place to print errors. > > Maybe this should be caught higher up in livecd-creator, but then you run into > a situation where you are having to try to catch all conceivable errors and > convert them to prints. It seems more reasonable to me to let it raise the > error, which includes both the easy to read output and a traceback to where the > original error happened in the library. > - error msg or error dialog would be imho the right solution, but it's your code and no one can force you to fix it.. > I think it is a mistake for abrt to be catching tracebacks in python code at > all. If the python interpreter crashes, fine. But abrt doesn't (and very likely > can't) provide any useful additional information to a traceback. - it's you one problem with it against many bugs in python code fixed thanks to ABRT reports... |