Tarantool development patches archive
 help / color / mirror / Atom feed
From: Kirill Yukhin <kyukhin@tarantool.org>
To: korablev@tarantool.org
Cc: tarantool-patches@freelists.org
Subject: [tarantool-patches] Re: [PATCH] sql: refactor SQL cursor to remove write ones
Date: Fri, 5 Oct 2018 07:39:44 +0300	[thread overview]
Message-ID: <20181005043944.ts6gcu5t2ts7sa4n@tarantool.org> (raw)
In-Reply-To: <20181004120034.pbdqm7thu7kiqa25@tarantool.org>

Hello,
On 04 окт 15:00, Kirill Yukhin wrote:
> Hello,
> On 01 окт 13:31, Kirill Yukhin wrote:
> > In Tarantool opening and positioning cursor for writing
> > have no sense. So refactor SQL code, to perform:
> >   - Creation of ephemeral table w/o any cursors machinery.
> >     No op-code returns register which contains plain pointer
> >     to the ephemeral space.
> >   - OpenRead/OpenWrite opcodes were replaced with single
> >     CursorOpen op-code, which establishes new cursor w/
> >     intention to sebsequent read from the space. This opcode
> >     accepts both plain pointer (in P4 operand) and register
> >     which contains pointer (inn P3) to the ephemeral space.
> > This query scheduler and DML routines thoroughly.
> > 
> > Closes #3182
> > Part of #2362
> > ---
> > Branch: https://github.com/tarantool/tarantool/commits/kyukhin/gh-3182-repair-cursors-2
> > Issue: https://github.com/tarantool/tarantool/issues/3182
> Patch was checked into 2.0 branch.

Seems like the patch broken testing in release mode. This is
due to non-initialized local variable.

Patch in the bottom fixes that. Comitted as obvious into 2.0
branch.

--
Regard, Kirill Yukhin

commit cebcf21114f50141e8aa0a72f30ea47dfbfef16c
Author: Kirill Yukhin <kyukhin@tarantool.org>
Date:   Fri Oct 5 07:19:56 2018 +0300

    sql: fix local var initialization in select planner
    
    Not initialized var led to crashes in non-debug
    modes.
---
 src/box/sql/where.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/src/box/sql/where.c b/src/box/sql/where.c
index 6a1bea0..520a847 100644
--- a/src/box/sql/where.c
+++ b/src/box/sql/where.c
@@ -4591,7 +4591,7 @@ sqlite3WhereBegin(Parse * pParse,	/* The parser context */
 		if (pLoop->wsFlags & WHERE_INDEXED) {
 			struct index_def *idx_def = pLoop->index_def;
 			int iIndexCur;
-			int op;
+			int op = OP_IteratorOpen;
 			/* Check if index is primary. Either of
 			 * points should be true:
 			 * 1. struct Index is non-NULL and is
@@ -4636,7 +4636,6 @@ sqlite3WhereBegin(Parse * pParse,	/* The parser context */
 					}
 				}
 				assert(wctrlFlags & WHERE_ONEPASS_DESIRED);
-				op = OP_IteratorOpen;
 				pWInfo->aiCurOnePass[1] = iIndexCur;
 			} else if (iAuxArg
 				   && (wctrlFlags & WHERE_OR_SUBCLAUSE) != 0) {

      reply	other threads:[~2018-10-05  4:39 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-10-01 10:31 [tarantool-patches] " Kirill Yukhin
2018-10-02 11:56 ` [tarantool-patches] " n.pettik
2018-10-03  9:57   ` Kirill Yukhin
2018-10-04  2:21     ` n.pettik
2018-10-04  7:16       ` Kirill Yukhin
2018-10-04 10:45         ` n.pettik
2018-10-04 11:48           ` Kirill Yukhin
2018-10-04 12:00 ` Kirill Yukhin
2018-10-05  4:39   ` Kirill Yukhin [this message]

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=20181005043944.ts6gcu5t2ts7sa4n@tarantool.org \
    --to=kyukhin@tarantool.org \
    --cc=korablev@tarantool.org \
    --cc=tarantool-patches@freelists.org \
    --subject='[tarantool-patches] Re: [PATCH] sql: refactor SQL cursor to remove write ones' \
    /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