From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 67294 invoked by alias); 31 Jan 2018 18:02:11 -0000 Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org Received: (qmail 67280 invoked by uid 89); 31 Jan 2018 18:02:10 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-16.4 required=5.0 tests=AWL,BAYES_00,GIT_PATCH_1,GIT_PATCH_2,GIT_PATCH_3,RCVD_IN_DNSWL_NONE autolearn=ham version=3.3.2 spammy= X-HELO: mail-io0-f194.google.com Received: from mail-io0-f194.google.com (HELO mail-io0-f194.google.com) (209.85.223.194) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Wed, 31 Jan 2018 18:02:04 +0000 Received: by mail-io0-f194.google.com with SMTP id f4so16149992ioh.8 for ; Wed, 31 Jan 2018 10:02:03 -0800 (PST) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:in-reply-to:references:from:date :message-id:subject:to:cc; bh=ds4RiD0ppIeipB+rqNuUPL5JXFuaatwXvq0TIJzoYjk=; b=eh3V8BROE3JMri6drZ0HNAd9uXIhlRb8of0ChZ2EBe5jtfa1mNrpa3CEpRBtZ95zqw eBKvvwNVX4uGJeGHK0rdQzt/zxLzWXt4K4w5Fds8bV7svIDd2GUuuSsZ2NG0YcDRRJJA HyPyB1By8cAoIbkpk4pR/f0r2JT3swY3rP4TXEH/TWZd5y14P+Nu7jlE+Jjx22TdRZ1+ 2f5MsiX7RUjtzlk2JYqhVsP+1Q9yRLlZqPeT38dV1SoNg1dZsnIHnGdRYqgx4kOmJHqm 7N4fyMkWLo/gMB9IbFY9/UZLCD30hpH9VyLKUPxI+8TDVQ/Y3Om9lTp7xQtB7g0Qbb73 V6xA== X-Gm-Message-State: AKwxytcsBiJOdoFlZ1zRjdJARIAoB3fKhD7GbrT1Erm6HPM9cc5MuvbN GkFkxOpMA/BipwqoNZycLQ9X9p6w/O5DaeXiILgbrA== X-Google-Smtp-Source: AH8x227sd0JA1Y6y38s9S/xlmHONMFK0JpQ4w5xefOr5BKHX9VuMK7F1WCMRhSf5N4udQbX5CUSiEhCV/kzi7/68Mxo= X-Received: by 10.107.175.8 with SMTP id y8mr35443636ioe.37.1517421722048; Wed, 31 Jan 2018 10:02:02 -0800 (PST) MIME-Version: 1.0 Received: by 10.107.17.206 with HTTP; Wed, 31 Jan 2018 10:01:41 -0800 (PST) In-Reply-To: <20180131174529.GP2063@tucnak> References: <20180124231943.GJ2063@tucnak> <8ce8bce4-07bf-705e-cf4b-fc6d7858250a@gmail.com> <20180125225354.GU2063@tucnak> <20180131174529.GP2063@tucnak> From: Jason Merrill Date: Wed, 31 Jan 2018 18:36:00 -0000 Message-ID: Subject: Re: [C++ PATCH] Fix constexpr handling of arrays with unknown bound (PR c++/83993) To: Jakub Jelinek Cc: Martin Sebor , Nathan Sidwell , gcc-patches List Content-Type: text/plain; charset="UTF-8" X-IsSubscribed: yes X-SW-Source: 2018-01/txt/msg02472.txt.bz2 OK. On Wed, Jan 31, 2018 at 12:45 PM, Jakub Jelinek wrote: > On Wed, Jan 31, 2018 at 11:02:36AM -0500, Jason Merrill wrote: >> > In this case I was just extending existing warning and wanted >> > consistency with that. Both can be changed in a GCC9 follow-up, >> > or if Jason/Nathan want it now, even for GCC8, sure. > > So, do you want %qE as in the patch or %E? Note, sidx might > actually not be a INTEGER_CST (as we only verify_constant it which > allows all kinds of expressions), though usually it will be. > E.g. the > if (!tree_fits_shwi_p (index) > || (i = tree_to_shwi (index)) < 0) > { > diag_array_subscript (ctx, ary, index); > case will call it if index is some reduced_constant_expression_p > which is not INTEGER_CST, initializer_constant_valid_p* has lots > of cases, sure not all apply to integral types. > >> > Well, even zero subscript is wrong if lval is false, i.e. a[0] + 1 >> > rather than e.g. &a[0]. So we'd need to have another wordings >> > for the case when the index is zero. >> >> But if lval is false we want the value of the array, and if we know >> the value we know the bounds, so that shouldn't be needed. > > You're right, when dereferencing it we'll error already earlier, > because the array of unknown bounds can't be constexpr. > > I think the " with unknown bounds" part still might be useful to users, > it can be seen in the type as [], but it might take some time to the users > to figure out that is the reason why we disallow it. > > So like this? > > 2018-01-31 Jakub Jelinek > > PR c++/83993 > * constexpr.c (diag_array_subscript): Emit different diagnostics > if TYPE_DOMAIN (arraytype) is NULL. > (cxx_eval_array_reference, cxx_eval_store_expression): For arrays > with NULL TYPE_DOMAIN use size_zero_node as nelts. > > * g++.dg/init/pr83993-1.C: New test. > * g++.dg/cpp0x/pr83993.C: New test. > > --- gcc/cp/constexpr.c.jj 2018-01-19 23:34:04.897278768 +0100 > +++ gcc/cp/constexpr.c 2018-01-24 13:38:40.572913190 +0100 > @@ -2270,13 +2270,20 @@ diag_array_subscript (const constexpr_ct > tree sidx = fold_convert (ssizetype, index); > if (DECL_P (array)) > { > - error ("array subscript value %qE is outside the bounds " > - "of array %qD of type %qT", sidx, array, arraytype); > + if (TYPE_DOMAIN (arraytype)) > + error ("array subscript value %qE is outside the bounds " > + "of array %qD of type %qT", sidx, array, arraytype); > + else > + error ("non-zero array subscript %qE is used with array %qD of " > + "type %qT with unknown bounds", sidx, array, arraytype); > inform (DECL_SOURCE_LOCATION (array), "declared here"); > } > - else > + else if (TYPE_DOMAIN (arraytype)) > error ("array subscript value %qE is outside the bounds " > "of array type %qT", sidx, arraytype); > + else > + error ("non-zero array subscript %qE is used with array of type %qT " > + "with unknown bounds", sidx, arraytype); > } > } > > @@ -2361,7 +2368,12 @@ cxx_eval_array_reference (const constexp > > tree nelts; > if (TREE_CODE (TREE_TYPE (ary)) == ARRAY_TYPE) > - nelts = array_type_nelts_top (TREE_TYPE (ary)); > + { > + if (TYPE_DOMAIN (TREE_TYPE (ary))) > + nelts = array_type_nelts_top (TREE_TYPE (ary)); > + else > + nelts = size_zero_node; > + } > else if (VECTOR_TYPE_P (TREE_TYPE (ary))) > nelts = size_int (TYPE_VECTOR_SUBPARTS (TREE_TYPE (ary))); > else > @@ -3439,7 +3451,12 @@ cxx_eval_store_expression (const constex > tree nelts, ary; > ary = TREE_OPERAND (probe, 0); > if (TREE_CODE (TREE_TYPE (ary)) == ARRAY_TYPE) > - nelts = array_type_nelts_top (TREE_TYPE (ary)); > + { > + if (TYPE_DOMAIN (TREE_TYPE (ary))) > + nelts = array_type_nelts_top (TREE_TYPE (ary)); > + else > + nelts = size_zero_node; > + } > else if (VECTOR_TYPE_P (TREE_TYPE (ary))) > nelts = size_int (TYPE_VECTOR_SUBPARTS (TREE_TYPE (ary))); > else > --- gcc/testsuite/g++.dg/init/pr83993-1.C.jj 2018-01-24 13:45:43.430864528 +0100 > +++ gcc/testsuite/g++.dg/init/pr83993-1.C 2018-01-24 13:44:59.352869530 +0100 > @@ -0,0 +1,11 @@ > +// PR c++/83993 > +// { dg-do compile } > + > +extern const int a[]; > +const int *const b = &a[0]; > + > +int > +foo () > +{ > + return b[0]; > +} > --- gcc/testsuite/g++.dg/cpp0x/pr83993.C.jj 2018-01-24 14:09:01.846716177 +0100 > +++ gcc/testsuite/g++.dg/cpp0x/pr83993.C 2018-01-24 14:08:41.246718212 +0100 > @@ -0,0 +1,49 @@ > +// PR c++/83993 > +// { dg-do compile { target c++11 } } > +// { dg-options "" } > + > +extern const int a[]; > +const int b[5] = { 1, 2, 3, 4, 5 }; > +extern const int c[4]; > +constexpr const int *d = &a[0]; > +constexpr const int *d2 = a; > +constexpr const int *e = &a[1]; // { dg-error "non-zero array subscript '1' is used with array 'a' of type 'const int \\\[\\\]' with unknown bounds" } > +constexpr const int *f = &b[0]; > +constexpr const int *f2 = b; > +constexpr const int *g = &b[5]; > +constexpr const int *h = &b[6]; // { dg-error "array subscript value '6' is outside the bounds of array 'b' of type 'const int \\\[5\\\]'" } > +constexpr const int *i = &c[0]; > +constexpr const int *i2 = c; > +constexpr const int *j = &c[4]; > +constexpr const int *k = &c[5]; // { dg-error "array subscript value '5' is outside the bounds of array 'c' of type 'const int \\\[4\\\]'" } > +extern const int l[]; > + > +void > +foo () > +{ > + extern const int l[3]; > + constexpr const int *m = &l[0]; > + constexpr const int *m2 = l; > + constexpr const int *n = &l[1]; > + static_assert (m == m2, ""); > +} > + > +constexpr const int *m = &l[0]; > +constexpr const int *m2 = l; > +constexpr const int *n = &l[1]; // { dg-error "non-zero array subscript '1' is used with array 'l' of type 'const int \\\[\\\]' with unknown bounds" } > +static_assert (d == d2 && f == f2 && i == i2 && m == m2, ""); > +const int o[] = { 1, 2 }; > +constexpr const int *p = &o[0]; > +constexpr const int *p2 = o; > +constexpr const int *q = &o[2]; > +constexpr const int *r = &o[3]; // { dg-error "array subscript value '3' is outside the bounds of array 'o' of type 'const int \\\[2\\\]'" } > +struct S { char a; char b[]; } s; > +constexpr const char *t = &s.b[0]; > +constexpr const char *t2 = s.b; > +constexpr const char *u = &s.b[1]; // { dg-error "non-zero array subscript '1' is used with array of type 'char \\\[\\\]' with unknown bounds" } > +struct V { int a; }; > +extern V v[]; > +constexpr V *w = &v[0]; > +constexpr V *w2 = v; > +constexpr int *x = &v[0].a; > +constexpr int y = a[0]; // { dg-error "the value of 'a' is not usable in a constant expression" } > > > Jakub