<html><head><meta http-equiv="Content-Type" content="text/html charset=windows-1252"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;"><div>Let's do it in two patches, but yes, that sounds like a great test case and new feature to add. It might be nice to add it to TraversalDumper as well—it's just a debug checker that rarely comes in handy, but as Prashant noted it is a surprising inconsistency.</div><div><br></div><div>Thanks!</div><div>Jordan</div><div><br></div><br><div><div>On Jan 20, 2014, at 6:22 , Zach Davis <<a href="mailto:zdavkeos@gmail.com">zdavkeos@gmail.com</a>> wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div dir="ltr"><div><div><div><div><div><br></div>I am finally getting a chance to look at this again. For a use-case/test-case I have added runCheckersForPreStmt to processSwitch and modified the UndefBranchChecker.cpp code to also look at SwitchStmt's so I get warning about this kind of thing:<br>
<br></div>int a[2];<br></div>switch (a[1]) { ...<br><br>>>><br><br>undef_test.c:7:5: warning: Branch condition evaluates to a garbage value<br> switch (a[1]) {<br> ^~~~~~ ~~~~<br><br></div>Does this sound like a reasonable test case?<br>
Should I add the UndefBranchChecker.cpp changes and corresponding test case in misc-ps.c to the patch I submit?<br></div><div>Are there other use-cases I should be considering?<br></div><div><br></div>Zach <br></div><div class="gmail_extra">
<br><br><div class="gmail_quote">On Thu, Dec 12, 2013 at 11:26 AM, Jordan Rose <span dir="ltr"><<a href="mailto:jordan_rose@apple.com" target="_blank">jordan_rose@apple.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div style="word-wrap:break-word"><div>That sounds right. Let me know if you have any more questions!</div><span class="HOEnZb"><font color="#888888"><div><br></div><div>Jordan</div></font></span><div><div class="h5"><br>
<div><div>On Dec 12, 2013, at 7:17 , Zach Davis <<a href="mailto:zdavkeos@gmail.com" target="_blank">zdavkeos@gmail.com</a>> wrote:</div><br><blockquote type="cite"><div dir="ltr">Let me see if I understand the plan:<br>
<br>1. Add a runCheckersForPreStmt() call to ExprEngine::processBranch<br>right after the call to runCheckersForBranchStmt(), effectively<br>chaining the two together. The loop can the evaluate the nodes<br>
returned from both calls?<br><br>2. Add a runCheckersForPreStmt() call to the top of<br>ExprEngine::processSwitch. The loop will then need to evaluate the<br>returned nodes as well as explore and evaluate the switch body as it<br>
currently does?<br><br>3. Add a runCheckersForPreStmt() call to<br>ExprEngine::processIndirectGoto and add a loop to evaluate any<br>returned nodes.<br><br>Zach<br><br>> Hi, Zach. That's great to hear. The code lives in ExprEngine.cpp, and<br>
> you can see in functions like ExprEngine::VisitUnaryOperator<br>> (ExprEngineC.cpp) that it basically just consists of a call to the<br>> CheckerManager to run the pre-statement checks, then a loop over the<br>
> results to actually perform the evaluation. In<br>> ExprEngine::processBranch the loop is already in place, so you just<br>> need to add the second callback; ExprEngine::processSwitch also has a<br>> loop but may need a bit more<br>
> restructuring. ExprEngine::processIndirectGoto is the last of these<br>> and probably needs a completely new loop added.<br></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Wed, Dec 11, 2013 at 12:08 PM, Jordan Rose <span dir="ltr"><<a href="mailto:jordan_rose@apple.com" target="_blank">jordan_rose@apple.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><br>
On Dec 6, 2013, at 10:07 , Zach Davis <<a href="mailto:zdavkeos@gmail.com" target="_blank">zdavkeos@gmail.com</a>> wrote:<br>
<br>
> > Right; what I meant was that we'd prefer to make this work rather than adding switch statements to checkBranchCondition, but right now neither one works. Sorry for the inconvenience.<br>
> ><br>
> > Jordan<br>
><br>
> I had this same problem and ended up using an ast-matcher in a clang tool instead to get the information I wanted.<br>
><br>
> I would be interested in helping add this feature though.<br>
><br>
><br>
> I found the bug report (18175), but could use some pointers to get started on a patch.<br>
<br>
</div>Hi, Zach. That's great to hear. The code lives in ExprEngine.cpp, and you can see in functions like ExprEngine::VisitUnaryOperator (ExprEngineC.cpp) that it basically just consists of a call to the CheckerManager to run the pre-statement checks, then a loop over the results to actually perform the evaluation. In ExprEngine::processBranch the loop is already in place, so you just need to add the second callback; ExprEngine::processSwitch also has a loop but may need a bit more restructuring. ExprEngine::processIndirectGoto is the last of these and probably needs a completely new loop added.<br>
<br>
Please feel free to ask more specific questions, and when you're ready send the first iteration of your patch to cfe-commits.<br>
<br>
Thanks for picking this up!<br>
<span><font color="#888888">Jordan<br>
</font></span></blockquote></div><br></div>
</blockquote></div><br></div></div></div>
</blockquote></div><br></div>
</blockquote></div><br></body></html>