<div dir="ltr"><br><div class="gmail_extra"><br><br><div class="gmail_quote">On Tue, Aug 26, 2014 at 1:45 PM, David Blaikie <span dir="ltr"><<a href="mailto:dblaikie@gmail.com" target="_blank">dblaikie@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div class=""><br><br>On Mon, Aug 18, 2014 at 2:33 PM, Adrian Prantl <<a href="mailto:aprantl@apple.com" target="_blank">aprantl@apple.com</a>> wrote:<br>
> IMO, the best we can do is to use the end of the CompoundStmt as the cleanup location if there is one, and otherwise give up and use line 0.<br>
<br></div>OK. The "let's do better when there's braces" isn't something I'm worrying about for now - it's a possible incremental improvement after we figure out the "worst case" behavior.<div class="">
<br>
<br>><br>> ```<br>> if (c)<br>> foo(); // cleanups = line 0<br>><br>> if (c) {<br>> } // cleanups<br>><br>> if (c)<br>> foo();<br>> else {<br>> bar();<br>> } // cleanups<br>
<br></div>In this case, I'm not sure the close } to the else is any better than the last line of the else, it still feels like part of the else, not the end of the if scope.<br><br>Anyway, using line zero (did this by hand and used no-integrated-as because Clang doesn't cope well with emitting line table entries for line 0) here's GDB's behavior in some examples:<br>
<br>GCC is GCC 4.8<br>ToT is Clang ToT<br>End is this patch applied to Clang ToT to use the end of the if as the location<br>Zer is the end patch applied, and then manually editing the assembly (and assembling with the system assembler, not the integrated assembler) to have the non-exceptional dtor call be on line 0.<br>
<br>So we get the totally right behavior from GDB the same as we would by assigning it to the end. In the cases where we get different behavior, it's better - we don't step to lines that don't execute, but we still get a bad location walking up from the dtor call (end up wherever the preceeding code was - which means sometimes we end up on the last line of the function where the EH cleanup is ascribed to). Seems better, but not perfect. (but we have no way to get perfect)<br>
</div></blockquote><div><br></div><div>& one other thing I didn't highlight was where the dtor was ascribed to in GCC, ToT, and End. I guess ToT and End are clear - that's the whole contention of this patch, ToT ascribes it to the 'if' (line 2 in all the examples below) and End ascribes it to the end of the if (line N-1 in all the examples below). GCC ascribes it to N-1 as well. As discussed previously, the way to get Clang to do what GCC is doing, is to do 'End' as well as moving the EH cleanup block after the normal cleanup. Then you get the advantages (1 (even better than "Zero", since it gets the dtor location 'right') and 2 (same as "Zero")) and disadvantages (weird nexting behavior in 3 and 4).<br>
</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">
<br><font face="courier new, monospace">1: for<br>2: if<br>3: func();<br>4: }<br><br>GCC: 1 2 3 1 2 1 2 3 1 4<br>ToT: 1 2 3 2 1 2 2 1 2 3 2 1 4<br>End: 1 2 3 3 1 2 3 1 2 3 3 1 4<br>Zer: 1 2 3 1 2 1 2 3 1 4, dtor skipped and ascribed to 4<br>
<br>1: for<br>2: if<br>3: func();<br>4: else<br>5: func();<br>6: }<br><br>GCC: 1 2 3 1 2 5 1 2 3 1 6<br>ToT: 1 2 3 2 1 2 5 2 2 3 2 1 6<br>End: 1 2 3 1 2 5 1 2 3 1 6<br>Zer: 1 2 3 1 2 5 1 2 3 1 6, dtor skipped and ascribed to 5<br>
<br>1: for<br>2: if<br>3: ;<br>4: else<br>5: ;<br>6: }<br><br>GCC: 1 2 5 1 2 5 1 2 5 1 6<br>ToT: 1 2 2 1 2 2 1 2 2 1 6<br>End: 1 2 5 1 2 5 1 2 5 1 6<br>Zer: 1 2 1 2 1 2 1 6, dtor skipped and ascribed to 6<br>
<br>1: for<br>2: if<br>3: ;<br>4: }<br><br>GCC: 1 2 3 1 2 3 1 2 3 1 4<br>ToT: 1 2 2 1 2 2 1 2 2 1 4<br>End: 1 2 3 1 2 3 1 2 3 1 4<br>Zer: 1 2 1 2 1 2 1 4, dtor skipped and ascribed to 4</font><br><br>> ```<br>
><br>> <a href="http://reviews.llvm.org/D4956" target="_blank">http://reviews.llvm.org/D4956</a><br>><br>><br></div>
</blockquote></div><br></div></div>