<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><br class=""><div><blockquote type="cite" class=""><div class="">On Oct 21, 2014, at 2:09 PM, David Blaikie <<a href="mailto:dblaikie@gmail.com" class="">dblaikie@gmail.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><br class="Apple-interchange-newline"><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><div class="gmail_quote" style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;">On Tue, Oct 21, 2014 at 2:01 PM, Frédéric Riss<span class="Apple-converted-space"> </span><span dir="ltr" class=""><<a href="mailto:friss@apple.com" target="_blank" class="">friss@apple.com</a>></span><span class="Apple-converted-space"> </span>wrote:<br class=""><blockquote class="gmail_quote" style="margin: 0px 0px 0px 0.8ex; border-left-width: 1px; border-left-color: rgb(204, 204, 204); border-left-style: solid; padding-left: 1ex;"><div style="word-wrap: break-word;" class=""><br class=""><div class=""><span class=""><blockquote type="cite" class=""><div class="">On Oct 20, 2014, at 2:06 PM, David Blaikie <<a href="mailto:dblaikie@gmail.com" target="_blank" class="">dblaikie@gmail.com</a>> wrote:</div><br class=""><div class=""><div dir="ltr" class="">I've checked this out on the GDB test suite & it seems fine.<br class=""><br class="">Looking at the code itself, I think it could be made a bit more obvious/uniform to have VariableDIE always refer to the definition and have just a single conditional to decide whether to attach the DW_AT_specification or not.<br class=""><br class="">Actually this works pretty well & even adds another source fidelity benefit, allowing the definition, while out-of-line of the class, to be attributed to the scope in which the definition occurs (eg: namespace x { struct foo { static int i; }; int foo::i; } - putting the definition in namespace 'x' instead of just forcing it out to the global namespace) and exposes a bug in Clang where the definition is scoped to the decl context instead of the lexical context. (patch attached to show a slightly short-sighted fix there)<br class=""></div></div></blockquote><div class=""><br class=""></div></span><div class="">Yeah, this looks an even nicer cleanup. I re-tested it just to be sure. Can you just commit what’s in your branch, or do you want me to reroll your reworked patch with a proper commit message?</div></div></div></blockquote><div class=""><br class=""></div><div class="">It'll need the clang change first, or we'll end up putting the static variable's definition inside the class (DW_AT_class_type with both a DW_AT_member and a DW_AT_variable as children, both describing the static member variable, with the latter pointing to the former via DW_AT_specification) & I'll need to fix/work up test coverage for it (so if you want to do those things instead, I won't complain - but I can do them).<br class=""></div></div></div></blockquote><div><br class=""></div><div>Oh that’s what it does, sorry, I thought you were just talking about file/line attribution of the definition… Interestingly it doesn’t seem to bother lldb much, but it’s obviously too ugly to live! The file/line issue is of course related. I thought it was the backend grabbing the wrong information from the Decls, but it’s in fact the frontend passing the wrong information down.</div><div><br class=""></div><blockquote type="cite" class=""><div class=""><div class="gmail_quote" style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;"><div class="">What was your thinking on the scoping issue in the frontend? It's certainly not the full solution - the full solution, imho would be to emit a declaration (potentially the DW_AT_member in the class) whenever the lexical context and the decl context differ. This would be the precise source fidelity. (& in the backend, do the right thing for fields that differ between the two - so we can describe the location of the definition in the .cpp file as different from the location of the declaration in the .h file, etc)<br class=""></div></div></div></blockquote><div><br class=""></div><div>Well, you have sent out the wrong patch for the frontend part, thus I misunderstood the exact issue you were trying to solve. If you send the right patch I can try to make sense of it and get an opinion on the matter.</div><div><br class=""></div><div>Fred</div><br class=""><blockquote type="cite" class=""><div class=""><div class="gmail_quote" style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;"><div class=""> - David</div><div class=""> </div><blockquote class="gmail_quote" style="margin: 0px 0px 0px 0.8ex; border-left-width: 1px; border-left-color: rgb(204, 204, 204); border-left-style: solid; padding-left: 1ex;"><div style="word-wrap: break-word;" class=""><div class=""><div class="">Thanks!<br class=""></div></div></div></blockquote><div class=""><br class="">Thanks for your patience,<br class=""><br class="">- David<br class=""> </div><blockquote class="gmail_quote" style="margin: 0px 0px 0px 0.8ex; border-left-width: 1px; border-left-color: rgb(204, 204, 204); border-left-style: solid; padding-left: 1ex;"><div style="word-wrap: break-word;" class=""><div class=""><div class=""></div><div class="">Fred</div><br class=""><blockquote type="cite" class=""><div class=""><span class=""><div dir="ltr" class="">If we did a few more smart things in the frontend, we could get better fidelity for declarations/definitions of namespace/global variables - removing the weird special case introduced with my patch. Probably the right thing here is to optimize in the frontend, if the definition appears in the decl context (eg: if the global variable's decl context is the same as the lexical context) then just produce the definition. If the decl and lexical context differ, produce both a declaration (in the decl context) and a definition (in the lexical context) - this would be the best source fidelity, but probably isn't too worthwhile.<br class=""><br class="">Oh, and if we wanted to be really good here, we could make sure we emit any attributes onto the definition that differ from the declaration - note GCC does the right thing here, putting DW_AT_decl_line/file on the definitions if the line/file happens to be different from the declaration's line/file. We don't do that - we just put nothing on the definition (& I don't think we do this for function declarations/definitions either (again, here, GCC gets this right)... could, but probably not important)<br class=""><br class=""><br class=""></div><div class="gmail_extra"><br class=""><div class="gmail_quote">On Thu, Oct 16, 2014 at 8:46 AM, Frederic Riss<span class="Apple-converted-space"> </span><span dir="ltr" class=""><<a href="mailto:friss@apple.com" target="_blank" class="">friss@apple.com</a>></span><span class="Apple-converted-space"> </span>wrote:<br class=""><blockquote class="gmail_quote" style="margin: 0px 0px 0px 0.8ex; border-left-width: 1px; border-left-color: rgb(204, 204, 204); border-left-style: solid; padding-left: 1ex;">ping<br class=""><br class=""><a href="http://reviews.llvm.org/D5457" target="_blank" class="">http://reviews.llvm.org/D5457</a><br class=""><br class=""><br class=""><br class="">_______________________________________________<br class="">llvm-commits mailing list<br class=""><a href="mailto:llvm-commits@cs.uiuc.edu" target="_blank" class="">llvm-commits@cs.uiuc.edu</a><br class=""><a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits" target="_blank" class="">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</a><br class=""></blockquote></div><br class=""></div></span><span class=""><static_var_def.diff></span><span class=""><global_var.diff></span></div></blockquote></div></div></blockquote></div></div></blockquote></div><br class=""></body></html>