From: "n.pettik" <korablev@tarantool.org> To: tarantool-patches@freelists.org Cc: Vladislav Shpilevoy <v.shpilevoy@tarantool.org> Subject: [tarantool-patches] Re: [PATCH 6/6] sql: introduce ALTER TABLE ADD CONSTRAINT UNIQUE/PRIMARY KEY Date: Wed, 16 Jan 2019 23:06:45 +0300 [thread overview] Message-ID: <A57AFCB3-F2AF-4FD5-BFD7-2DE0DBE1719A@tarantool.org> (raw) In-Reply-To: <3804c32a-4e68-d876-b811-119ff669852b@tarantool.org> >> diff --git a/src/box/sql/build.c b/src/box/sql/build.c >> index 1d7b6c827..0314be957 100644 >> --- a/src/box/sql/build.c >> +++ b/src/box/sql/build.c >> @@ -2083,6 +2083,19 @@ generate_index_id(struct Parse *parse, uint32_t space_id, int cursor) >> return iid_reg; >> } >> +static void >> +pk_check_existence(struct Parse *parse, uint32_t space_id, int cursor) > > 1. It is worth mentioning in a comment that the cursor is opened on _index > space. Or rename it to _index_cursor. It do not mind if in such a case you > omit a comment. I’d prefer second option: diff --git a/src/box/sql/build.c b/src/box/sql/build.c index 080914576..cdeb85bd7 100644 --- a/src/box/sql/build.c +++ b/src/box/sql/build.c @@ -2089,12 +2089,12 @@ generate_index_id(struct Parse *parse, uint32_t space_id, int cursor) } static void -pk_check_existence(struct Parse *parse, uint32_t space_id, int cursor) +pk_check_existence(struct Parse *parse, uint32_t space_id, int _index_cursor) { struct Vdbe *v = sqlite3GetVdbe(parse); int tmp_reg = ++parse->nMem; sqlite3VdbeAddOp2(v, OP_Integer, space_id, tmp_reg); - int found_addr = sqlite3VdbeAddOp4Int(v, OP_NotFound, cursor, 0, + int found_addr = sqlite3VdbeAddOp4Int(v, OP_NotFound, _index_cursor, 0, > >> +{ >> + struct Vdbe *v = sqlite3GetVdbe(parse); >> + int tmp_reg = ++parse->nMem; > > 2. Why not to use a truly temp register? As I know, we have facilities for > it. Parse.nMem are not temporary. Parse.nTempReg - is. I am afraid that they don’t work. I greped through code a bit and found no real usages of tmp regs. What is more: int sqlite3GetTempReg(Parse * pParse) { if (pParse->nTempReg == 0) { return ++pParse->nMem; } return pParse->aTempReg[--pParse->nTempReg]; } It always returns real register since nTempReg == 0 and is never incremented. I’ve opened issue: https://github.com/tarantool/tarantool/issues/3942 https://github.com/tarantool/tarantool/issues/3943 >> + sqlite3VdbeAddOp2(v, OP_Integer, space_id, tmp_reg); >> + int found_addr = sqlite3VdbeAddOp4Int(v, OP_NotFound, cursor, 0, >> + tmp_reg, 1); >> + sqlite3VdbeAddOp4(v, OP_Halt, SQLITE_ERROR, ON_CONFLICT_ACTION_FAIL, 0, >> + "multiple primary keys are not allowed", P4_STATIC); >> + sqlite3VdbeJumpHere(v, found_addr); >> +} >> + >> /** >> * Add new index to table's indexes list. >> * We follow convention that PK comes first in list. >> diff --git a/test/sql-tap/alter.test.lua b/test/sql-tap/alter.test.lua >> index 1aad555c0..925753749 100755 >> --- a/test/sql-tap/alter.test.lua >> +++ b/test/sql-tap/alter.test.lua >> @@ -517,6 +517,62 @@ test:do_catchsql_test( >> -- </alter-7.11> >> }) >> +test:do_test( >> + "alter-8.1.0", >> + function() >> + format = {} >> + format[1] = { name = 'id', type = 'scalar'} >> + format[2] = { name = 'f2', type = 'scalar'} >> + s = box.schema.create_space('T', {format = format}) >> + end, >> + {}) >> + >> +test:do_catchsql_test( >> + "alter-8.1.1", >> + [[ >> + ALTER TABLE t ADD CONSTRAINT pk PRIMARY KEY("id"); >> + ]], { >> + 0 >> + }) >> + >> +test:do_test( >> + "alter-8.1.2", >> + function() >> + i = box.space.T.index[0] >> + return i.id > > 3. Why not return box.space.T.index[0].id? … diff --git a/test/sql-tap/alter.test.lua b/test/sql-tap/alter.test.lua index 925753749..318b0f68d 100755 --- a/test/sql-tap/alter.test.lua +++ b/test/sql-tap/alter.test.lua @@ -538,8 +538,7 @@ test:do_catchsql_test( test:do_test( "alter-8.1.2", function() - i = box.space.T.index[0] - return i.id + return box.space.T.index[0].id end, 0)
next prev parent reply other threads:[~2019-01-16 20:06 UTC|newest] Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-01-09 12:13 [tarantool-patches] [PATCH 0/6] Introduce ALTER TABLE ADD CONSTRAINT UNIQUE/PK Nikita Pettik 2019-01-09 12:13 ` [tarantool-patches] [PATCH 1/6] sql: move constraint name to struct contraint_parse Nikita Pettik 2019-01-14 14:04 ` [tarantool-patches] " Vladislav Shpilevoy 2019-01-16 20:06 ` n.pettik 2019-01-16 20:54 ` Vladislav Shpilevoy 2019-01-17 10:56 ` Konstantin Osipov 2019-01-17 17:14 ` n.pettik 2019-01-09 12:13 ` [tarantool-patches] [PATCH 2/6] sql: rework ALTER TABLE grammar Nikita Pettik 2019-01-14 14:05 ` [tarantool-patches] " Vladislav Shpilevoy 2019-01-16 20:06 ` n.pettik 2019-01-16 20:54 ` Vladislav Shpilevoy 2019-01-17 11:51 ` Konstantin Osipov 2019-01-17 17:14 ` n.pettik 2019-01-18 1:42 ` Konstantin Osipov 2019-01-09 12:13 ` [tarantool-patches] [PATCH 3/6] sql: remove start token from sql_create_index args Nikita Pettik 2019-01-09 12:13 ` [tarantool-patches] [PATCH 4/6] sql: refactor getNewIid() function Nikita Pettik 2019-01-14 14:05 ` [tarantool-patches] " Vladislav Shpilevoy 2019-01-09 12:13 ` [tarantool-patches] [PATCH 5/6] sql: fix error message for improperly created index Nikita Pettik 2019-01-14 14:06 ` [tarantool-patches] " Vladislav Shpilevoy 2019-01-16 20:06 ` n.pettik 2019-01-09 12:13 ` [tarantool-patches] [PATCH 6/6] sql: introduce ALTER TABLE ADD CONSTRAINT UNIQUE/PRIMARY KEY Nikita Pettik 2019-01-14 14:06 ` [tarantool-patches] " Vladislav Shpilevoy 2019-01-16 20:06 ` n.pettik [this message] 2019-01-16 20:54 ` Vladislav Shpilevoy
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=A57AFCB3-F2AF-4FD5-BFD7-2DE0DBE1719A@tarantool.org \ --to=korablev@tarantool.org \ --cc=tarantool-patches@freelists.org \ --cc=v.shpilevoy@tarantool.org \ --subject='[tarantool-patches] Re: [PATCH 6/6] sql: introduce ALTER TABLE ADD CONSTRAINT UNIQUE/PRIMARY KEY' \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: link
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox