https://jira.opendaylight.org This file is an XML representation of an issue en-us 8.20.10 820010 22-06-2022 https://jira.opendaylight.org/browse/GENIUS-66 genius <p>Currently, getEgressActionInfosForInterface() has a number of fall-through cases. Each of these needs to be verified, and either:</p> <ul> <li>replaced with a break;</li> <li>or documented.</li> </ul> <p>Documentation involves also removing any associated TODOs (e.g. from <a href="https://git.opendaylight.org/gerrit/#/c/53896" class="external-link" target="_blank" rel="nofollow noopener">https://git.opendaylight.org/gerrit/#/c/53896</a>) and @SuppressWarnings.</p> <p>Operating System: All<br/> Platform: All</p> GENIUS-66

IfmUtil.getEgressActionInfosForInterface()'s behaviour needs to be verified and documented

Bug Resolved Done Michael Vorburger Stephen Kitt Mon, 27 Mar 2017 16:25:41 +0000 Thu, 6 Apr 2017 15:29:30 +0000 Thu, 6 Apr 2017 15:29:30 +0000 (unspecified) General 0 3 <p>On looking at this code more closely a 2nd time after a good night's sleep, I can actually now see that the if/s in the case sort of cancel each other out.. </p> <p>It's very confusing / non-obvious / hard to read code - but probably not really a bug then, as it looks like a switch case fall through bug, but because of the nested if actually is not (which I didn't notice yesterday). </p> <p>This really should be refactored more clearly, under this bug.</p> <p>As <a href="https://git.opendaylight.org/gerrit/#/c/53631/" class="external-link" target="_blank" rel="nofollow noopener">https://git.opendaylight.org/gerrit/#/c/53631/</a> and <a href="https://git.opendaylight.org/gerrit/#/c/53875/" class="external-link" target="_blank" rel="nofollow noopener">https://git.opendaylight.org/gerrit/#/c/53875/</a> are just about to add more tests for this code (according to private email from Faseela), perhaps someone could have a go at cleaning up the switch in IfmUtil.getEgressActionInfosForInterface() after those two changes for extended test coverage have been merged.</p> <p>&gt; * replaced with a break;</p> <p>meanwhile had a chat with Faseela on IRC to confirm it's not a bug, it's supposed to be like this, it just confused us.</p> <p>&gt; * or documented.</p> <p>just noticed on <a href="http://checkstyle.sourceforge.net/config_coding.html#FallThrough" class="external-link" target="_blank" rel="nofollow noopener">http://checkstyle.sourceforge.net/config_coding.html#FallThrough</a> that CS has a something that's a bit nicer for this case then the SuppressWarnings, apparently if you put "// falls through" just before the case: without a "break;" that indicates to it that it's intentional and OK.</p> <p>&gt; nicer for this case then the @SuppressWarnings("checkstyle:FallThrough")</p> <p>and, apparently @SuppressWarnings("fallthrough") helps javac understand that we really mean to do that, so we probably should put that as hit to javac.</p> <p>(In reply to Michael Vorburger from comment #3)<br/> &gt; &gt; nicer for this case then the @SuppressWarnings("checkstyle:FallThrough")<br/> &gt; <br/> &gt; and, apparently @SuppressWarnings("fallthrough") helps javac understand that<br/> &gt; we really mean to do that, so we probably should put that as hit to javac.</p> <p>I followed your recommendation to comment the "fall-through":</p> <p><a href="https://git.opendaylight.org/gerrit/#/c/53896/5" class="external-link" target="_blank" rel="nofollow noopener">https://git.opendaylight.org/gerrit/#/c/53896/5</a></p> <p>Considering this resolved, as we've confirmed verbally with Faseela that the switch case "fall through" in IfmUtil.getEgressActionInfosForInterface() is intentional and not a bug, and "documented" this (not the "why" just the "fact that") via inline "// fall through", which Checkstyle detects and then shuts up about.</p> <p>I did actually spend a moment to try to refactor it differently, but figured that the code becomes less not more readable - so let's just leave that as it is.</p> <p>PS: We <b>ARE</b> hoping that the nascent but extremely minimalistic IfmUtilTest can be expanded upon to cover the IfmUtil.getEgressActionInfosForInterface() &amp; more code better!</p> <p>&gt; nascent but extremely minimalistic IfmUtilTest </p> <p>sorry, that was the wrong class of course; the InterfaceManagerConfigurationTest is actually one of our shining examples of Component Tests ... I've just quickly locally verified with <a href="http://www.eclemma.org" class="external-link" target="_blank" rel="nofollow noopener">http://www.eclemma.org</a>, and it does seem to cover the case of VLAN_INTERFACE in IfmUtil.getEgressActionInfosForInterface() .. perhaps future work could cover the other cases (MPLS_OVER_GRE, GRE_TRUNK_INTERFACE, VXLAN_TRUNK_INTERFACE).</p> Development External issue ID 8089 External issue URL Rank 0|i01on3: