en Copyright 2015 Java http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/Analysis/CFLAliasAnalysis/Steensgaard/const-exprs.ll#57d57b1a [AAEval] Make compatible with opaque pointersWith opaque pointers, we cannot use the pointer element type todetermine the LocationSize for the AA query. Instead, -aa-evaltests are now required to have an explicit load or store for anypointer they want to compute alias results for, and the load/storetypes are used to determine the location size.This may affect ordering of results, and sorting within one result,as the type is not considered part of the sorted string anymore.To somewhat minimize the churn, printing still uses faux typedpointer notation. List of files: /llvm-project-15.0.7/llvm/test/Analysis/CFLAliasAnalysis/Steensgaard/const-exprs.ll Tue, 15 Mar 2022 09:18:19 +0000 Nikita Popov <npopov@redhat.com> http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/Analysis/CFLAliasAnalysis/Steensgaard/const-exprs.ll#b81d5baa [test] Use new PM for -aa-eval tests List of files: /llvm-project-15.0.7/llvm/test/Analysis/CFLAliasAnalysis/Steensgaard/const-exprs.ll Tue, 08 Mar 2022 22:00:08 +0000 Arthur Eubanks <aeubanks@google.com> http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/Analysis/CFLAliasAnalysis/Steensgaard/const-exprs.ll#4cd19a6e [BasicAA] Rename -disable-basicaa to -disable-basic-aa to be consistent with the canonical name "basic-aa" List of files: /llvm-project-15.0.7/llvm/test/Analysis/CFLAliasAnalysis/Steensgaard/const-exprs.ll Sat, 27 Jun 2020 03:55:44 +0000 Fangrui Song <maskray@google.com> http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/Analysis/CFLAliasAnalysis/Steensgaard/const-exprs.ll#175400a8 Add regression test for r331976In general, it's difficult to poke the ConstantExpr code in CFLAA, sinceLLVM is so great at eagerly reducing ConstantExprs. :)Sadly, this only shows a functional difference from before the patchbecause CFLAA has some special logic around taking loads of non-pointersinto account. Namely, with the broken select behavior, CFLAA willcompletely fail to take note of @g3. Since CFLAA doesn't have any recordabout @g3 when we do an alias query for @g3 and %a, it conservativelyanswers MayAlias. When we properly take @g3 into account with the newselect logic, we get NoAlias for this query.I suspect that the aforementioned "special logic" isn't completelycorrect, but this test-case should prevent future wonky aliasing resultsfrom appearing for these flavors of ConstantExprs, so I think it's stillworth having.llvm-svn: 332017 List of files: /llvm-project-15.0.7/llvm/test/Analysis/CFLAliasAnalysis/Steensgaard/const-exprs.ll Thu, 10 May 2018 18:37:54 +0000 George Burgess IV <george.burgess.iv@gmail.com>