[Java] False negative with squid:S128 - Switch cases should end with an unconditional "break" statement
116 views
Skip to first unread message
Francis Galiegue
Oct 13, 2015, 8:57:47 AM10/13/15
to SonarQube
Hello,
On Tuesday, October 13, 2015 at 11:49:12 AM UTC+2, David Racodon wrote:
Hi,Any chance you can deal with the following false negative?
While I disagree with this rule as a whole (there are perfectly legitimate uses of fall through in switch cases), here it is not a false positive; if your value is 1 in the switch case, the return of the value is not unconditional and depends on whether j == 3.
The rule itself, in fact, should probably be rephrased, with the message being something like "this switch case falls through the next one; ensure that this is intended" or something.
Regards
David Racodon
Oct 13, 2015, 9:53:40 AM10/13/15
to Francis Galiegue, SonarQube
Hi Francis,
here it is not a false positive; if your value is 1 in the switch case, the return of the value is not unconditional and depends on whether j == 3.
I'm talking about false negative here, not false positive. I don't get any issue but I do expect one to be raised on "case 1".
Regards,
--
You received this message because you are subscribed to the Google Groups "SonarQube" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sonarqube+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sonarqube/8ba1c95f-9f33-4415-8e12-5e4c5445e629%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Francis Galiegue
Oct 13, 2015, 11:28:25 AM10/13/15
to SonarQube, fgal...@gmail.com
Hello,
On Tuesday, October 13, 2015 at 3:53:40 PM UTC+2, David Racodon wrote:
Hi Francis,here it is not a false positive; if your value is 1 in the switch case, the return of the value is not unconditional and depends on whether j == 3.I'm talking about false negative here, not false positive. I don't get any issue but I do expect one to be raised on "case 1".
Uhm, yes, sorry for the noise. Although I have some difficulty about the description of it being a "false negative"; rather an "undetected case".
Regardless of whether this case is actually detected though, I believe the rule's message should be changed to something along the lines of my suggestion.
Regards
David Racodon
Oct 13, 2015, 11:32:17 AM10/13/15
to Francis Galiegue, SonarQube
Uhm, yes, sorry for the noise. Although I have some difficulty about the description of it being a "false negative"; rather an "undetected case".
This is the same for me.
Regardless of whether this case is actually detected though, I believe the rule's message should be changed to something along the lines of my suggestion.
I wouldn't mind your proposal about the issue message to be implemented.
--
You received this message because you are subscribed to the Google Groups "SonarQube" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sonarqube+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sonarqube/02aa1512-174b-43c5-b65c-6762718cfd7f%40googlegroups.com.
Michael Gumowski
Oct 14, 2015, 3:00:17 AM10/14/15
to David Racodon, Francis Galiegue, SonarQube
Hey guys,
Thank you for the feedback.
I created the following JIRA ticket to handle the issue: https://jira.sonarsource.com/browse/SONARJAVA-1336
Regards,
To view this discussion on the web visit https://groups.google.com/d/msgid/sonarqube/CACTFSAhrX8pAVrcjBygunutF_me5msZviUDKnpB%2BgRBJfM_hGA%40mail.gmail.com.
0 new messages