From: Vladislav Shpilevoy via Tarantool-patches <tarantool-patches@dev.tarantool.org>
To: imeevma@tarantool.org
Cc: tarantool-patches@dev.tarantool.org
Subject: Re: [Tarantool-patches] [PATCH v1 20/21] sql: remove MEM_Term flag
Date: Fri, 15 Oct 2021 00:47:33 +0200 [thread overview]
Message-ID: <51eb0bb4-b620-257e-5af3-10c8b23e7b2b@tarantool.org> (raw)
In-Reply-To: <278cf14adddc65006253100ce12604626e748062.1633713433.git.imeevma@gmail.com>
Thanks for the patch!
> diff --git a/src/box/sql/printf.c b/src/box/sql/printf.c
> index b4ab0d0f9..f261a731c 100644
> --- a/src/box/sql/printf.c
> +++ b/src/box/sql/printf.c
> @@ -160,8 +160,22 @@ getTextArg(PrintfArguments * p)
> {
> if (p->nArg <= p->nUsed)
> return 0;
> - struct Mem *mem = p->apArg[p->nUsed++];
> - return (char *)mem_as_str0(mem);
> + struct Mem mem;
> + mem_create(&mem);
> + mem_copy(&mem, p->apArg[p->nUsed++]);
1. mem_copy_as_ephemeral() would be enough. But would be even
better to have mem_snprintf() which would work like snprintf -
save string representation of mem into a provided buffer and
return how many bytes written / need to be written. This though
could be postponed indefinitely, I don't mind.
> + if (mem_to_str(&mem) != 0) {
> + mem_destroy(&mem);
> + return NULL;
> + }
> + char *str = region_alloc(&fiber()->gc, mem.n + 1);
2. printf is called for each row. Which means it will leak the
region. You can either do the snprintf-thing described above to
write directly into the needed buffer, or at least cleanup the
region using a save point in the end of sqlVXPrintf().
> diff --git a/src/box/sql/whereexpr.c b/src/box/sql/whereexpr.c
> index 6849f13ec..ad064d500 100644
> --- a/src/box/sql/whereexpr.c
> +++ b/src/box/sql/whereexpr.c
> @@ -307,12 +307,18 @@ like_optimization_is_valid(Parse *pParse, Expr *pExpr, Expr **ppPrefix,
>
> op = pRight->op;
> if (op == TK_VARIABLE) {
> - Vdbe *pReprepare = pParse->pReprepare;
> - int iCol = pRight->iColumn;
> - pVal = sqlVdbeGetBoundValue(pReprepare, iCol);
> - if (pVal != NULL && mem_is_str(pVal)) {
> - if (mem_as_str0(pVal) == NULL)
> + var = vdbe_get_bound_value(pParse->pReprepare,
> + pRight->iColumn - 1);
> + if (var != NULL && mem_is_str(var)) {
> + uint32_t size = var->n + 1;
> + char *str = region_alloc(&fiber()->gc, size);
3. You leak the fiber's region. AFAIR, it is not cleaned up in the
end of parsing. Try to use pParse->region.
> + if (str == NULL) {
> + diag_set(OutOfMemory, size, "region", "str");
> return -1;
> + }
> + memcpy(str, var->z, var->n);
> + str[var->n] = '\0';
> + z = str;
> }
next prev parent reply other threads:[~2021-10-14 22:47 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-08 17:31 [Tarantool-patches] [PATCH v1 00/21] Refactor non-standard and non-aggragate functions Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 01/21] sql: refactor CHAR() function Mergen Imeev via Tarantool-patches
2021-10-14 22:42 ` Vladislav Shpilevoy via Tarantool-patches
2021-10-25 8:02 ` Mergen Imeev via Tarantool-patches
2021-10-29 23:42 ` Vladislav Shpilevoy via Tarantool-patches
2021-11-02 11:35 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 02/21] sql: refactor GREATEST() and LEAST() functions Mergen Imeev via Tarantool-patches
2021-10-14 22:42 ` Vladislav Shpilevoy via Tarantool-patches
2021-10-25 8:17 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 03/21] sql: refactor HEX() function Mergen Imeev via Tarantool-patches
2021-10-14 22:43 ` Vladislav Shpilevoy via Tarantool-patches
2021-10-25 8:19 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 04/21] sql: refactor LENGTH() function Mergen Imeev via Tarantool-patches
2021-10-14 22:43 ` Vladislav Shpilevoy via Tarantool-patches
2021-10-25 8:30 ` Mergen Imeev via Tarantool-patches
2021-10-29 23:42 ` Vladislav Shpilevoy via Tarantool-patches
2021-11-02 11:39 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 05/21] sql: refactor PRINTF() function Mergen Imeev via Tarantool-patches
2021-10-14 22:44 ` Vladislav Shpilevoy via Tarantool-patches
2021-10-25 8:33 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 06/21] sql: refactor RANDOM() function Mergen Imeev via Tarantool-patches
2021-10-25 8:35 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 07/21] sql: rework RANDOMBLOB() function Mergen Imeev via Tarantool-patches
2021-10-25 8:36 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 08/21] sql: refactor ZEROBLOB() function Mergen Imeev via Tarantool-patches
2021-10-25 8:37 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 09/21] sql: refactor TYPEOF() function Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 10/21] sql: refactor ROUND() function Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 11/21] sql: refactor ROW_COUNT() function Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 12/21] sql: rework UUID() function Mergen Imeev via Tarantool-patches
2021-10-25 8:38 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 13/21] sql: refactor VERSION() function Mergen Imeev via Tarantool-patches
2021-10-08 17:31 ` [Tarantool-patches] [PATCH v1 14/21] sql: refactor UNICODE() function Mergen Imeev via Tarantool-patches
2021-10-14 22:44 ` Vladislav Shpilevoy via Tarantool-patches
2021-10-25 8:40 ` Mergen Imeev via Tarantool-patches
2021-11-02 11:42 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:32 ` [Tarantool-patches] [PATCH v1 15/21] sql: refactor of SOUNDEX() function Mergen Imeev via Tarantool-patches
2021-10-08 17:32 ` [Tarantool-patches] [PATCH v1 16/21] sql: refactor REPLACE() function Mergen Imeev via Tarantool-patches
2021-10-14 22:45 ` Vladislav Shpilevoy via Tarantool-patches
2021-10-25 8:45 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:32 ` [Tarantool-patches] [PATCH v1 17/21] sql: refactor QUOTE() function Mergen Imeev via Tarantool-patches
2021-10-08 17:32 ` [Tarantool-patches] [PATCH v1 18/21] sql: remove unused code Mergen Imeev via Tarantool-patches
2021-10-25 8:51 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:32 ` [Tarantool-patches] [PATCH v1 19/21] sql: remove MEM_Dyn flag Mergen Imeev via Tarantool-patches
2021-10-14 22:46 ` Vladislav Shpilevoy via Tarantool-patches
2021-10-25 8:54 ` Mergen Imeev via Tarantool-patches
2021-10-29 23:43 ` Vladislav Shpilevoy via Tarantool-patches
2021-11-02 11:43 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:32 ` [Tarantool-patches] [PATCH v1 20/21] sql: remove MEM_Term flag Mergen Imeev via Tarantool-patches
2021-10-14 22:47 ` Vladislav Shpilevoy via Tarantool-patches [this message]
2021-10-25 9:57 ` Mergen Imeev via Tarantool-patches
2021-10-08 17:32 ` [Tarantool-patches] [PATCH v1 21/21] sql: make arguments to be const Mergen Imeev via Tarantool-patches
2021-11-02 22:15 ` [Tarantool-patches] [PATCH v1 00/21] Refactor non-standard and non-aggragate functions Vladislav Shpilevoy via Tarantool-patches
2021-11-11 10:48 Mergen Imeev via Tarantool-patches
2021-11-11 10:49 ` [Tarantool-patches] [PATCH v1 20/21] sql: remove MEM_Term flag Mergen Imeev via Tarantool-patches
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=51eb0bb4-b620-257e-5af3-10c8b23e7b2b@tarantool.org \
--to=tarantool-patches@dev.tarantool.org \
--cc=imeevma@tarantool.org \
--cc=v.shpilevoy@tarantool.org \
--subject='Re: [Tarantool-patches] [PATCH v1 20/21] sql: remove MEM_Term flag' \
/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