From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from [87.239.111.99] (localhost [127.0.0.1]) by dev.tarantool.org (Postfix) with ESMTP id 11D786B962; Wed, 14 Apr 2021 03:39:52 +0300 (MSK) DKIM-Filter: OpenDKIM Filter v2.11.0 dev.tarantool.org 11D786B962 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=tarantool.org; s=dev; t=1618360792; bh=y5CKQ/z0RRmItMLWtgdHLggQ1gaifBokBP18fDxs5Fg=; h=Date:To:References:In-Reply-To:Subject:List-Id:List-Unsubscribe: List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To:Cc: From; b=bkVhvsh5C4nLaNaWXb23AYLPKZx/X6ar6PjLy9Zzleirkwu3uGtZ3QrgutyW4uDN2 cLDOfzDG33AEG1KjTGUTPx6gXROQk7UykwTCp5B6hiWqWxjBXm9AXLKFSNvcpsq5Yx vl+3JOA1UzN9VB6u3nCKYSs9/Y7FqmhD1LDXKQ2w= Received: from smtp30.i.mail.ru (smtp30.i.mail.ru [94.100.177.90]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by dev.tarantool.org (Postfix) with ESMTPS id A34646EC5B for ; Wed, 14 Apr 2021 03:39:50 +0300 (MSK) DKIM-Filter: OpenDKIM Filter v2.11.0 dev.tarantool.org A34646EC5B Received: by smtp30.i.mail.ru with esmtpa (envelope-from ) id 1lWTZN-0003km-P7; Wed, 14 Apr 2021 03:39:50 +0300 Date: Wed, 14 Apr 2021 03:39:48 +0300 To: Vladislav Shpilevoy Message-ID: <20210414003948.GA140789@tarantool.org> References: <29a16259-d26a-4391-4a8d-0439fec146a0@tarantool.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <29a16259-d26a-4391-4a8d-0439fec146a0@tarantool.org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eAau8CL7WIMRKs4sN3D3tLDjz0dLbV79QFUyzQ2Ujvy7cMT6pYYqY16iZVKkSc3dCLJ7zSJH7+u4VD18S7Vl4ZUrpaVfd2+vE6kuoey4m4VkSEu530nj6fImhcD4MUrOEAnl0W826KZ9Q+tr5ycPtXkTV4k65bRjmOUUP8cvGozZ33TWg5HZplvhhXbhDGzqmQDTd6OAevLeAnq3Ra9uf7zvY2zzsIhlcp/Y7m53TZgf2aB4JOg4gkr2biojnA7/qPBUIXEYe84jVlrWGw== X-Mailru-Sender: 5C3750E245F362008BC1685FEC6306ED37247840F31BC2161F71E938D9020A4149EBE2174CCC94795105BD0848736F9966FEC6BF5C9C28D97E07721503EA2E00ED97202A5A4E92BF7402F9BA4338D657ED14614B50AE0675 X-Mras: Ok Subject: Re: [Tarantool-patches] [PATCH v5 46/52] sql: introduce mem_get_uint() X-BeenThere: tarantool-patches@dev.tarantool.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: Tarantool development patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , From: Mergen Imeev via Tarantool-patches Reply-To: Mergen Imeev Cc: tarantool-patches@dev.tarantool.org Errors-To: tarantool-patches-bounces@dev.tarantool.org Sender: "Tarantool-patches" Thank you for the review! My answer, diff and new patch below. On Wed, Apr 14, 2021 at 01:04:05AM +0200, Vladislav Shpilevoy wrote: > Thanks for the fixes! > > On 09.04.2021 23:08, Mergen Imeev via Tarantool-patches wrote: > > Thank you for the review! My answer and new patch below. > > > > > > On 30.03.2021 02:08, Vladislav Shpilevoy wrote: > >> Thanks for the patch! > >> > >> On 23.03.2021 10:36, Mergen Imeev via Tarantool-patches wrote: > >>> This patch introduces mem_get_unsigned() function which is used to > >>> receive unsigned value from MEM. > >>> > >>> Part of #5818 > >>> --- > >>> src/box/sql/func.c | 16 +++++++++++----- > >>> src/box/sql/mem.c | 37 +++++++++++++++++++++++++++---------- > >>> src/box/sql/mem.h | 6 +++--- > >>> src/box/sql/sqlInt.h | 3 --- > >>> src/box/sql/vdbeapi.c | 6 ------ > >>> 5 files changed, 41 insertions(+), 27 deletions(-) > >>> > >>> diff --git a/src/box/sql/func.c b/src/box/sql/func.c > >>> index 0fa0f6ac7..a851d98f2 100644 > >>> --- a/src/box/sql/func.c > >>> +++ b/src/box/sql/func.c > >>> @@ -118,9 +118,12 @@ port_vdbemem_dump_lua(struct port *base, struct lua_State *L, bool is_flat) > >>> luaL_pushint64(L, n); > >>> break; > >>> } > >>> - case MP_UINT: > >>> - luaL_pushuint64(L, sql_value_uint64(param)); > >>> + case MP_UINT: { > >>> + uint64_t u; > >>> + mem_get_unsigned(param, &u); > >>> + luaL_pushuint64(L, u); > >> Maybe we could make 2 functions? One to get the value and ignore > >> the errors, and the other to get as an out parameter + return an > >> error? > >> > >> For instance, mem_to_uint() - returns uint64_t and internally asserts > >> that the value is correct. And mem_get_uint() works like your version. > >> > >> The same for the other get functions whose result is often ignored. > > For some functions I created a "proxy" functions in func.c the way you > > described, but not for this function since it is only used in a few places of > > sql/func.c. Should I do this for all functions? In func.c I mean. I see this as > > temporary measure, since I hope we will rework built-in functions one day. > > Unfortunately, 'hope' is not enough. And it is highly possible the code > will live for long. Therefore I think we need to make it solid where possible > and clearly state it is unsafe or add assertions where it is not possible. > > Here mem_get_uint() result is ignored always. Even if it fails. I think it > must be called something like mem_get_uint_unsafe() and return the uint as > 'return', not via an out argument. Then at least we would see it is broken > when we are around this code again, and it won't raise questions if it is a > known issue, and why it is not fixed (this must be in a comment for the > function). Understood. I created a new function, mem_get_uint_unsafe(). Diff: diff --git a/src/box/sql/func.c b/src/box/sql/func.c index 6e6978bbc..5503a9b16 100644 --- a/src/box/sql/func.c +++ b/src/box/sql/func.c @@ -201,9 +201,7 @@ absFunc(sql_context * context, int argc, sql_value ** argv) UNUSED_PARAMETER(argc); switch (sql_value_type(argv[0])) { case MP_UINT: { - uint64_t u; - mem_get_uint(argv[0], &u); - sql_result_uint(context, u); + sql_result_uint(context, mem_get_uint_unsafe(argv[0])); break; } case MP_INT: { @@ -1171,7 +1169,7 @@ charFunc(sql_context * context, int argc, sql_value ** argv) if (sql_value_type(argv[i]) == MP_INT) x = 0xfffd; else - mem_get_uint(argv[i], &x); + x = mem_get_uint_unsafe(argv[i]); if (x > 0x10ffff) x = 0xfffd; c = (unsigned)(x & 0x1fffff); diff --git a/src/box/sql/mem.h b/src/box/sql/mem.h index 4aadcd3f7..ca8a75b50 100644 --- a/src/box/sql/mem.h +++ b/src/box/sql/mem.h @@ -776,6 +776,18 @@ mem_get_int_unsafe(const struct Mem *mem) int mem_get_uint(const struct Mem *mem, uint64_t *u); +/** + * Return value of MEM converted to uint64_t. This function is not safe, since + * its behaviour is undefined if mem_get_uint() returned an error. + */ +static inline uint64_t +mem_get_uint_unsafe(const struct Mem *mem) +{ + uint64_t u; + mem_get_uint(mem, &u); + return u; +} + /** * Simple type to str convertor. It is used to simplify * error reporting. New patch: commit b9ca33e93110ecf167329a0f58473371de5c7c45 Author: Mergen Imeev Date: Wed Mar 17 13:55:37 2021 +0300 sql: introduce mem_get_uint() This patch introduces mem_get_uint() function. This function is used to receive unsigned value from MEM. If value of MEM is not unsigned, it is converted to unsigned if possible. MEM is not changed. Part of #5818 diff --git a/src/box/sql/func.c b/src/box/sql/func.c index 0db698174..5503a9b16 100644 --- a/src/box/sql/func.c +++ b/src/box/sql/func.c @@ -201,7 +201,7 @@ absFunc(sql_context * context, int argc, sql_value ** argv) UNUSED_PARAMETER(argc); switch (sql_value_type(argv[0])) { case MP_UINT: { - sql_result_uint(context, sql_value_uint64(argv[0])); + sql_result_uint(context, mem_get_uint_unsafe(argv[0])); break; } case MP_INT: { @@ -1169,7 +1169,7 @@ charFunc(sql_context * context, int argc, sql_value ** argv) if (sql_value_type(argv[i]) == MP_INT) x = 0xfffd; else - x = sql_value_uint64(argv[i]); + x = mem_get_uint_unsafe(argv[i]); if (x > 0x10ffff) x = 0xfffd; c = (unsigned)(x & 0x1fffff); diff --git a/src/box/sql/mem.c b/src/box/sql/mem.c index adf5e236b..ab31029df 100644 --- a/src/box/sql/mem.c +++ b/src/box/sql/mem.c @@ -1074,6 +1074,33 @@ mem_get_int(const struct Mem *mem, int64_t *i, bool *is_neg) return -1; } +int +mem_get_uint(const struct Mem *mem, uint64_t *u) +{ + if ((mem->flags & MEM_Int) != 0) + return -1; + if ((mem->flags & MEM_UInt) != 0) { + *u = mem->u.u; + return 0; + } + if ((mem->flags & (MEM_Str | MEM_Blob)) != 0) { + bool is_neg; + if (sql_atoi64(mem->z, (int64_t *)u, &is_neg, mem->n) != 0 || + is_neg) + return -1; + return 0; + } + if ((mem->flags & MEM_Real) != 0) { + double d = mem->u.r; + if (d >= 0 && d < (double)UINT64_MAX) { + *u = (uint64_t)d; + return 0; + } + return -1; + } + return -1; +} + int mem_copy(struct Mem *to, const struct Mem *from) { @@ -2283,16 +2310,6 @@ sql_value_boolean(sql_value *val) return b; } -uint64_t -sql_value_uint64(sql_value *val) -{ - int64_t i = 0; - bool is_neg; - mem_get_int((struct Mem *) val, &i, &is_neg); - assert(!is_neg); - return i; -} - const unsigned char * sql_value_text(sql_value * pVal) { diff --git a/src/box/sql/mem.h b/src/box/sql/mem.h index f3d9043e5..ca8a75b50 100644 --- a/src/box/sql/mem.h +++ b/src/box/sql/mem.h @@ -768,6 +768,26 @@ mem_get_int_unsafe(const struct Mem *mem) return i; } +/** + * Return value for MEM of UNSIGNED type. For MEM of all other types convert + * value of the MEM to UNSIGNED if possible and return converted value. Original + * MEM is not changed. + */ +int +mem_get_uint(const struct Mem *mem, uint64_t *u); + +/** + * Return value of MEM converted to uint64_t. This function is not safe, since + * its behaviour is undefined if mem_get_uint() returned an error. + */ +static inline uint64_t +mem_get_uint_unsafe(const struct Mem *mem) +{ + uint64_t u; + mem_get_uint(mem, &u); + return u; +} + /** * Simple type to str convertor. It is used to simplify * error reporting. @@ -835,9 +855,6 @@ sql_value_double(struct Mem *); bool sql_value_boolean(struct Mem *val); -uint64_t -sql_value_uint64(struct Mem *val); - const unsigned char * sql_value_text(struct Mem *); diff --git a/src/box/sql/sqlInt.h b/src/box/sql/sqlInt.h index 0af247ebf..6ead9b261 100644 --- a/src/box/sql/sqlInt.h +++ b/src/box/sql/sqlInt.h @@ -448,9 +448,6 @@ sql_column_double(sql_stmt *, int iCol); bool sql_column_boolean(struct sql_stmt *stmt, int column); -uint64_t -sql_column_uint64(struct sql_stmt *stmt, int column); - const unsigned char * sql_column_text(sql_stmt *, int iCol); diff --git a/src/box/sql/vdbeapi.c b/src/box/sql/vdbeapi.c index 5e5957496..1126425bc 100644 --- a/src/box/sql/vdbeapi.c +++ b/src/box/sql/vdbeapi.c @@ -476,12 +476,6 @@ sql_column_boolean(struct sql_stmt *stmt, int i) return sql_value_boolean(columnMem(stmt, i)); } -uint64_t -sql_column_uint64(sql_stmt * pStmt, int i) -{ - return sql_value_uint64(columnMem(pStmt, i)); -} - const unsigned char * sql_column_text(sql_stmt * pStmt, int i) {