Jordan, Thank you very much for the feedback, I have a few comments.<div><br></div><div><div class="gmail_quote"><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><i>(1) Despite being on our list, this is probably better suited to a compiler warning.</i></div><div><br></div></div></blockquote><div><br></div><div>I agree, this warning might be better as a compiler warning. I chose to implement this checker as a mainly to learn a bit about the analyzer. This one was on the list and seemed like a good place to get started.</div>
<div> </div><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><div><i>(2) Despite being on our list, "unsigned" isn't actually the interesting thing to check.</i></div>
</div><div><br></div></div></blockquote><div><br></div><div>When I was reading the checker suggestion, I interpreted the purpose to be a more conservative version of a check for unary '+', which, arguably, is often dead code. For example, I have seen structures like this fairly commonly:</div>
<div><br></div><div>int array[] = {</div><div> -3,</div><div> -2,</div><div> -1,</div><div> +1</div><div>};</div><div><br></div><div>Where the '+' is used for alignment, which we wouldn't want to warn about. However, if that array was changed to unsigned, it might be a legitimate warning. I thought the assumption was there's at least a decent chance a unary '+' on unsigned is dead code. The place where I most commonly it pop up was legitimate:</div>
<div><br></div><div>char a = 'A';</div><div>cout << a << " "; // print A</div><div>cout << +a; // prints numerical value of 'A'</div><div><br></div><div>But I hadn't considered the checker was intended to target idempotent or erroneous promotions. If that is the intent, then it seems challenging to decide whether an expression is dead code, or to "force a load", as you put it.</div>
<div><br></div><div><br></div><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><div><i>(3) Macros and templates make this tricky.</i></div>
</div></div></blockquote><div><br></div><div>I thought the that this might have been the reason why the checker was listed as a potential checker, rather than a compiler warning. It does seem like a fairly "noisy" warning. I have run it through some student code. Unfortunately all warnings it produced were false positives, with the exception of one situation similar to the one above.</div>
<div><br></div><div><br></div><div>At this point, I'd be fine with throwing this checker out, as its utility does seem quite limited. If anyone has any ideas on how this checker can be improved to be more useful, I would be interested to hear.</div>
<div><br></div><div>On an unrelated note, do you have any recommendations for what might be a approachable second checker?</div><div><br></div><div>Adam</div><div><br></div><div><br></div><div> </div><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><div> </div></div></div></blockquote><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></div>
<br><div><div><div><div>On Apr 12, 2013, at 23:53 , Adam Schnitzer <<a href="mailto:adamschn@umich.edu" target="_blank">adamschn@umich.edu</a>> wrote:</div><br></div></div><blockquote type="cite"><div><div>
This patch is an implementation of the proposed "different.UnaryPlusWithUnsigned", which I implemented<div>as "alpha.deadcode.UnaryPlusWithUnsigned".<div><br></div><div>It is implemented as a simple AST checker. However, it seems that unary '+' is often removed from the AST</div>
</div><div>as dead code. So some of the basic test cases don't work.</div><div><br></div><div>This is my first (real) patch, so any feedback or criticism is appreciated.</div><div><br></div><div>Adam Schnitzer</div>
</div></div><span><UnaryPlusChecker.patch></span></blockquote></div><br></div></blockquote></div><br></div>