Git log: commit be10be8f9e888f6ebe16f04759321b1dd730ce90
Author: Igor Munkin <imun@tarantool.org>
Date: Mon Apr 19 17:54:18 2021 +0300
luajit: bump new version
LuaJIT submodule is bumped to introduce the following change:
* test: fix directory detection in lua-Harness suite
This changeset fixes the hidden bug in lua-Harness test suite that was
revealed by applying 9610c7410ce210cb29de0a616f357c1a17392047 ('hotfix:
update libcurl submodule to 7.76.0'). This was 314-th patch in 2.8.0
release, so RPM-related tools confuse the testing machinery using the
patch revision in the building tree names[1].
[1]: https://github.com/tarantool/tarantool/runs/2361010932#step:5:10032
Follows up #5844
commit d2bd451095890c1caaedb2d1c2fc3a49c554915e
Author: Mergen Imeev <imeevma@gmail.com>
Date: Fri Apr 16 13:10:31 2021 +0300
Add test and changelog for issue gh-5890
commit 3a6371265147512d9e072f53e32f67bf98940d20
Author: Mergen Imeev <imeevma@gmail.com>
Date: Fri Apr 16 11:36:44 2021 +0300
Add changelog for select uuid in SQL
commit 9610c7410ce210cb29de0a616f357c1a17392047
Author: Nikita Pettik <korablev@tarantool.org>
Date: Fri Apr 16 12:25:22 2021 +0300
hotfix: update libcurl submodule to 7.76.0
It was accidentally reverted by 660686e3f.
commit d15c633e464096d76cc949f7ab291be1e0d0be2f
Author: Artem Starshov <artemreyt@tarantool.org>
Date: Thu Apr 15 12:40:34 2021 +0300
coverity: add errinj macroses to fix dead code
Our coverity (scan.coverity.com) reports about dead code (CWE-561)
in cases where variable of type `struct errinj *` is compared with
NULL at if condition. Dead code appears due to function-like macro
`errinj()` always returns NULL at non-debug version, so execution
never goes into if block.
Added macroses ERROR_INJECT_INT and ERROR_INJECT_DOUBLE which solves
this problem like it does ERROR_INJECT macro with boolean errinj type.
Usage you can see on example of ERRINJ_STDIN_ISATTY fix in this commit.
Follows up #5040
commit 7c9ce0b4bee2aef893b5edb6379dd96ea3c2aaab
Author: Nikita Pettik <korablev@tarantool.org>
Date: Thu Apr 15 16:12:59 2021 +0300
Add changelos for memtx mvcc issues
commit 660686e3f95ee210f421157c4534107d8eb75c6c
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Thu Apr 15 13:54:47 2021 +0300
txm: fix a small mistake that led to an assertion
Follow-ip of #5628
commit 1faa45f9340666d85580b70b31b2392d7498c2e2
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Wed Apr 14 15:56:35 2021 +0300
txm: implement read gap trackers
After implementation of point hole trackers only one problem has
left. When a transaction reads a value from a TREE index by key
that key may be partial and/or search type may be GE, GT etc. In
this case point hole hash table doesn't provide track of writes
into some key interval.
Implement gap trackers for those purposes. The main idea is to
store those trackers in a story of a closest tuple to the right of
the lookup (aka successor) or directly in index if there no
successor (for example if the space in empty).
Closes #5628
commit 85b6b46a11aba474efae23c6da5739be0a590aac
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Thu Apr 15 08:46:01 2021 +0300
memtx: refactor tree iterator start function
Reorder some code, rename a bit and and unify branches.
Part of #5628
commit ef0f2b78fb65e19088e07d9aa2639cf34f375e88
Author: Iskander Sagitov <i.sagitov@tarantool.org>
Date: Wed Apr 14 17:09:38 2021 +0300
test: fix nondeterminism of ER_TUPLE_FOUND error
ER_TUPLE_FOUND error message depends on the previous tests, so we can
just check error code in these tests to avoid it.
commit e17a53843fa19603213d5a0f4d08750335223436
Author: mechanik20051988 <mechanik20051988@tarantool.org>
Date: Wed Apr 14 11:15:39 2021 +0300
tuple: do not allow autofill of nested arrays with nulls
In previous version there was inconsistent/ behaviour for update
operations to nested arrays. When user try to update non existent
field of nested array along with previous update existent field,
it's works, otherwise no. For example we create tuple and two
update operations:
`t = box.tuple.new({1, 2, {11, 22}})`
`op1 = {'=', '[3][1]', 11}`
`op2 = {'=', '[3][4]', 44}`
Then in case we apply operations separately `t:update({op1})` and
then `t:update({op2})`, second operation fails but is case we apply
them together t:update({op1, op2}) they both complete successfully.
Follow up #3378
commit 1b330121ec7dafcc0fc6ab63279e0dc52f9f947c
Author: Roman Khabibov <roman.habibov@tarantool.org>
Date: Tue Mar 9 16:47:11 2021 +0300
box: set box.cfg options via environment variables
Add ability to set box.cfg options via environment variables. These
variables should have name `TT_<OPTION>`. When Tarantool instance is
started under tarantoolctl utility, environment variables have higher
priority than tarantoolctl configuration file.
Closes #5602
Co-authored-by: Leonid Vasiliev <lvasiliev@tarantool.org>
Co-authored-by: Alexander Turenko <alexander.turenko@tarantool.org>
@TarantoolBot document
Title: Set box.cfg options via environment variables
Now, it is possible to set box.cfg options via environment variables.
The name of variable should correspond the following pattern:
`TT_<NAME>`, where `<NAME>` is uppercase box.cfg option name. For
example: `TT_LISTEN`, `TT_READAHEAD`.
Array values are separated by comma. Example:
```sh
export TT_REPLICATION=localhost:3301,localhost:3302
```
An empty variable is the same as unset one.
commit 8d1a4bdbf5f819d353a61d7e62ae48cd3f04c5ab
Author: Cyrill Gorcunov <gorcunov@gmail.com>
Date: Wed Apr 14 20:10:43 2021 +0300
cfg: provide types for logger options
This needed for fast type check when fetching
options from environment variable.
Part-of #5602
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
commit 8cb3b77d8bc50eefbfd301cd2e5d9fa3a83c51e3
Author: Mergen Imeev <imeevma@gmail.com>
Date: Thu Apr 8 11:51:08 2021 +0300
sql: introduce mem_get_agg()
This patch introduces mem_get_agg(). This function is used to receive
address of memory allocated for accumulation structure of the aggregate
function.
Closes #5818
commit 718d43eec34f91b575c3d8a780be6725ce39c470
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 22 12:27:33 2021 +0300
sql: introduce mem_len()
This patch introduces mem_len(). This function is used to receive
length of string or binary value of MEM. If MEM is not of STRING or
VARBINARy type this function returns -1.
Part of #5818
commit deabc2ce5b6af2487bcf2e16d9572cfd5049613c
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 22 11:54:01 2021 +0300
sql: introduce mem_get_bin()
This patch introduces mem_get_bin(). This function is used to receive
binary value from MEM. If value of MEM is not binary value, it is
converted to binary value if possible. MEM is not changed.
Part of #5818
commit 716a33d1fdfc36a8311be0bb60d39d446d37d24a
Author: Mergen Imeev <imeevma@gmail.com>
Date: Thu Mar 18 12:41:41 2021 +0300
sql: introduce mem_get_str0() and mem_as_str0()
This patch introduces mem_get_str0() and mem_as_str0(). Function
mem_get_str0() is used to receive NULL-terminated string from MEM. If
value of MEM is not NULL-terminated string, it is converted to
NULL-terminated string if possible. MEM is not changed. Function
mem_as_str0() is also used to receive NULL-terminated string from MEM,
however if MEM does not contain NULL-terminated string it converts MEM
to MEM that contains NULL-terminated string and returns its value.
Part of #5818
commit 953a847cfd3aec8d5648bc8c3bbfb65fcefff626
Author: Mergen Imeev <imeevma@gmail.com>
Date: Wed Mar 17 14:20:37 2021 +0300
sql: introduce mem_get_bool()
This patch introduces mem_get_bool(). This function is used to receive
boolean value from MEM. If value of MEM is not boolean, it is
converted to boolean if possible. MEM is not changed.
Part of #5818
commit d4e55ae8ab7dd2e80ffa5f89a4a319d8892631eb
Author: Mergen Imeev <imeevma@gmail.com>
Date: Wed Mar 17 14:13:30 2021 +0300
sql: introduce mem_get_double()
This patch introduces mem_get_double(). This function is used to receive
double value from MEM. If value of MEM is not double, it is converted to
double if possible. MEM is not changed.
Part of #5818
commit bde6d26e968143010e20448dcb8fb0e53f9adfb0
Author: Mergen Imeev <imeevma@gmail.com>
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
commit dbdc4eec848e16c83eb4a7150935cde39bddcb2c
Author: Mergen Imeev <imeevma@gmail.com>
Date: Wed Mar 17 13:20:37 2021 +0300
sql: introduce mem_get_int()
This patch introduces mem_get_int() function. This function is used to
receive integer value from MEM. If value of MEM is not integer, it is
converted to integer if possible. MEM is not changed.
Part of #5818
commit 655c4df556e23d98e03a82b0d050df209ba79449
Author: Mergen Imeev <imeevma@gmail.com>
Date: Wed Mar 17 12:43:08 2021 +0300
sql: introduce mem_cast_implicit()
This patch introduces mem_cast_implicit(). This function is used to
convert a MEM to given type according to implicit cast rules.
Part of #5818
commit c8b8fd666ae5e7cdc1d70f382f81233b82abf33b
Author: Mergen Imeev <imeevma@gmail.com>
Date: Wed Mar 17 12:27:09 2021 +0300
sql: introduce mem_cast_explicit()
This patch introduces mem_cast_explicit(). This function is used to
convert a MEM to a given field type according to explicit cast rules.
Part of #5818
commit b47715c1cc251b9c8d0d1692687eacdbe048e8b8
Author: Mergen Imeev <imeevma@gmail.com>
Date: Wed Mar 17 11:55:48 2021 +0300
sql: introduce mem_to_str() and mem_to_str0()
This patch introduces mem_to_str() and mem_to_str0() functions. These
functions are used to convert a MEM to a MEM that contains string value.
These functions defines the rules that are used during convertion from
values of all other types to STRING.
Part of #5818
commit 985eca7cfc1bb2de444c3295f2fc5c6b8496cfa6
Author: Mergen Imeev <imeevma@gmail.com>
Date: Wed Mar 17 11:39:13 2021 +0300
sql: introduce mem_to_number()
This patch introduces mem_to_number(). This function is used to convert
MEM to MEM contains number.
Part of #5818
commit 5b438e244e19d4c1f6ca4399d1cf89f8ae1ab9eb
Author: Mergen Imeev <imeevma@gmail.com>
Date: Wed Mar 17 11:03:22 2021 +0300
sql: introduce mem_to_double()
This patch intruduces mem_to_double(). This function is used to convert
a MEM to a MEM that contains double value. This function defines the
rules that are used during conversion from values of all other types to
double.
Part of #5818
commit 829daeed1b67326c11a8457156f5ba3ba883403f
Author: Mergen Imeev <imeevma@gmail.com>
Date: Wed Mar 17 10:46:29 2021 +0300
sql: introduce mem_to_int*() functions
This patch introduces mem_to_int() and mem_to_int_precise() functions.
These functions are used to convert a MEM to a MEM that contains
integer value. These functions defines the rules that are used during
convertion from values of all other types to INTEGER.
Part of #5818
commit 1581f8831b7929a40e0faf587501e08bacccbe35
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 16 15:14:32 2021 +0300
sql: introduce mem_set_null_clear()
This patch introduces mem_set_null_clear() function. This function sets
"cleared" NULL to MEM. This NULL is used only in internal VDBE
operations.
Part of #5818
commit bdb64c1d6423c4f43ae870bccaf4cda6c374823d
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 16 15:02:08 2021 +0300
sql: introduce mem_set_agg()
This patch introduces mem_set_agg() function. This function stores
aggregation function to MEM and allocates enough memory to hold
accumulation structure for aggregate function.
Part of #5818
commit d15f2a72950d0997a8509434a9c7fb3e1b1d72eb
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 16 14:40:37 2021 +0300
sql: introduce mem_set_frame()
This patch introduces mem_set_frame() function. This function clears the
MEM and sets a frame to MEM. Frames used for internal VDBE operations.
Part of #5818
commit f22686dcb08b60648fe064bb7e8115cde066f678
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 16 14:36:24 2021 +0300
sql: refactor mem_set_ptr()
This patch refactors mem_set_ptr() function. Also, it moves the function
to all others mem_set_*() functions.
Part of #5818
commit 8bdc9da82030cba9fdf33187097c7881ee13a76e
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 16 14:26:55 2021 +0300
sql: introduce mem_set_invalid()
This patch introduces mem_set_invalid() function. This function clears
MEM and invalidates it.
Part of #5818
commit f149b5ade6131c9011bae67f0039da016e2916c9
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 16 13:50:54 2021 +0300
sql: introduce mem_set_*() for map and array
This patch introduces set of mem_set_*() functions for MAP and ARRAY.
These functions clears MEM and sets it to given binary value. Binary
value must be MAP or ARRAY. Degree of clearing and type of allocation of
the binary value is determined by the function used.
Part of #5818
commit fe92a240d3aa8cf31d202c9671ba639c76412b89
Author: Mergen Imeev <imeevma@gmail.com>
Date: Wed Mar 17 11:11:42 2021 +0300
sql: introduce mem_set_zerobin()
This patch introduces mem_set_zerobin() function. This function clears
MEM and sets it to binary value that consists of given number of zero
bytes.
Part of #5818
commit 95d49373e965749540dbf132bece8e924893c91b
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 16 11:55:05 2021 +0300
sql: introduce mem_copy_bin()
This patch introduces mem_copy_bin() function. This function copies
given binary value to a newly allocated by MEM memory.
Part of #5818
commit 2a4efdd5a203bee5c08f1b89313a8ec87dcc1b10
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 16 10:17:33 2021 +0300
sql: introduce mem_set_bin_*() functions
This patch introduces set of mem_set_bin_*() functions. These functions
clears MEM and sets it to given binary value. Degree of clearing and
type of allocation of the binary value is determined by the function
used.
Part of #5818
commit d089a6cf06c7119c15cb7cd82bace7b25cf64ac6
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 15 16:59:27 2021 +0300
sql: introduce mem_copy_str() and mem_copy_str0()
This patch introduces mem_copy_str() and mem_copy_str0() functions.
These functions copy given string to newly allocated by MEM memory.
String given to mem_copy_str0() should be NULL-terminated. This is not
necessary for mem_copy_str().
Part of #5818
commit 79273e38dc2a4bb977314017f37efd6a0895dccc
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 15 14:28:29 2021 +0300
sql: introduce mem_set_str_*() functions
This patch introduces set of mem_set_str_*() functions. These functions
clears MEM and sets it to given string value. Degree of clearing and
type of allocation of the string is determined by the function used.
Part of #5818
commit d689a3afdcffa9efd31a1e3ad4f2c54d45d26cb9
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 15 12:42:59 2021 +0300
sql: move mem_set_bool() and mem_set_double()
This patch performs a small refactoring of mem_set_double() and moves
this function and mem_set_bool() to another place so that they are part
of the code block containing all mem_set_*() functions.
Part of #5818
commit c4a86696bcae71e63f911c637324910a04ee2202
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 15 12:38:26 2021 +0300
sql: introduce mem_set_uint()
This patch introduces mem_set_uint(). This function clears MEM and sets
it to given unsigned value.
Part of #5818
commit c9b6a278ef25a318ede630d32ae9daacd0df9980
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 15 11:29:40 2021 +0300
sql: introduce mem_set_int()
This patch introduces mem_set_int(). This function clears MEM and sets
it to given integer value.
Part of #5818
commit 1aad8fdb283296c9561b9d5c873f03f7c15ee920
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 15 10:43:29 2021 +0300
sql: introduce mem_set_null()
This patch introduces mem_set_null(). This function clears MEM and sets
it to NULL.
Part of #5818
commit 4d801264cd86924fdad3ad6e256c9cd3e65449a9
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 15 14:29:18 2021 +0300
sql: Initialize MEM in sqlVdbeAllocUnpackedRecord()
This patch adds initialization for newly created MEM objects in
sqlVdbeAllocUnpackedRecord(). Changing a MEM without being
initialized may give us unexpected result.
Part of #5818
commit 8bc1b41123a5bc1ae4d6d6e6446c3e10876f6ed8
Author: Mergen Imeev <imeevma@gmail.com>
Date: Sun Mar 14 22:20:11 2021 +0300
sql: introduce bitwise operations for MEM
This patch introduces mem_bit_and(), mem_bit_or(), mem_shift_left(),
mem_shift_right(), and mem_bit_not(), which perform bitwise operations
on MEM.
Part of #5818
commit 80cbac4dc16a90309d1db2cebf43c90028379dad
Author: Mergen Imeev <imeevma@gmail.com>
Date: Sun Mar 14 13:13:34 2021 +0300
sql: introduce mem_cmp_*() functions
This patch introduces set of mem_cmp_*() functions. These functions are
used to compare MEMs.
Part of #5818
commit de0e77f9719b20cb4aae4485bbd59a2307831202
Author: Mergen Imeev <imeevma@gmail.com>
Date: Sun Mar 14 11:51:52 2021 +0300
sql: introduce arithmetic operations for MEM
This patch introduces mem_add(), mem_sub(), mem_mul(), mem_div() and
mem_rem(), which perform arithmetic operations on two MEMs. Operands
must contain values of numeric types or values that can be converted
to a number according to implicit casting rules.
Part of #5818
commit 0c3d0af3738c67f4db89b3183d019cdb00bb4494
Author: Mergen Imeev <imeevma@gmail.com>
Date: Sat Mar 13 21:30:54 2021 +0300
sql: introduce mem_concat()
This patch introduces mem_concat(). Function mem_concat() concatenates
values from two MEMs in case these values are strings or binaries and
writes the result to the third MEM.
Part of #5818
commit f9b7049535cbb9a16eaa783e727be44105182d2e
Author: Mergen Imeev <imeevma@gmail.com>
Date: Sat Mar 13 18:03:25 2021 +0300
sql: remove sql_column_to_messagepack()
Function sql_column_to_messagepack() has almost the same functionality
as sql_vdbe_mem_encode_tuple(). Due to this it is not necessary to
have sql_column_to_messagepack(), so it is removed in this commit.
Part of #5818
commit f7a3eff7ff6bb03d29c0dc1fadd8c86fab7df494
Author: Mergen Imeev <imeevma@gmail.com>
Date: Sat Mar 13 15:43:38 2021 +0300
sql: rework vdbe_decode_msgpack_into_mem()
The original vdbe_decode_msgpack_into_mem() returns a MEM that contains
string and binary values as ephemeral. This patch renames this function
to mem_from_mp_ephemeral() and introduces new function mem_from_mp(),
which returns a MEM that contains string and binary values in newly
allocated memory.
This patch changes behavior for this query:
CREATE TABLE t1(m VARBINARY primary key);
INSERT INTO t1 VALUES(x'6178'), (x'6278'), (x'6379');
SELECT count(*), substr(m,2,1) AS mx FROM t1 GROUP BY mx;
Before this patch:
tarantool> SELECT count(*), substr(m,2,1) AS mx FROM t1 GROUP BY mx;
---
- metadata:
- name: COLUMN_1
type: integer
- name: MX
type: string
rows:
- [2, 'y']
- [1, 'y']
...
After this patch.
tarantool> SELECT count(*), substr(m,2,1) AS mx FROM t1 GROUP BY mx;
---
- metadata:
- name: COLUMN_1
type: integer
- name: MX
type: string
rows:
- [2, 'x']
- [1, 'y']
...
Part of #5818
Closes #5890
commit 1d24bef87cb9171194a58a2326287310f0f857fe
Author: Mergen Imeev <imeevma@gmail.com>
Date: Fri Mar 5 15:30:23 2021 +0300
sql: rework mem_move()
This patch reworks mem_move(). This function moves all content of source
MEM to destination MEM. Source mem is set to NULL.
Part of #5818
commit aae2c0a3cf619200b0a876128468dfcd8c619cfb
Author: Mergen Imeev <imeevma@gmail.com>
Date: Thu Mar 4 18:19:06 2021 +0300
sql: introduce mem_copy_as_ephemeral()
This patch intoduces mem_copy_as_ephemeral(). This function copies value
from source MEM to destination MEM. In case value is of string or binary
type and its allocation type is not static, it copied as value with
ephemeral allocation type.
Part of #5818
commit 29c93d020fa059d1c4ec21197a7bab535f9c86b2
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 23 00:53:03 2021 +0300
sql: introduce mem_copy()
This patch introduces mem_copy(). This function copies value from source
MEM to destination MEM. In case value is string or binary and have not
static allocation type, it is copied to newly allocated memory.
Part of #5818
commit 2d0870fcbce3404faa8908b05dab814e1ae21baf
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 2 11:27:48 2021 +0300
sql: introduce mem_is_*() functions()
This patch introduces mem_is_*() functions that allows to check current
MEM state.
Part of #5818
commit 65959761cbb0f387fe7c89920419f636e7b659da
Author: Mergen Imeev <imeevma@gmail.com>
Date: Thu Mar 4 14:07:35 2021 +0300
sql: introduce mem_destroy()
This patch introduces mem_destroy(). This function should be used to
free and destroy all objects owned by MEM, if necessary.
Part of #5818
commit 5360768a27bf5a98d19756f9ab91b8a1ed3b9cdf
Author: Mergen Imeev <imeevma@gmail.com>
Date: Thu Mar 4 13:32:07 2021 +0300
sql: introduce mem_create()
This patch introduces mem_create(). This function should be used to
initialize a MEM. MEM should be initialized before usage.
Part of #5818
commit c976ea84386c62a43625d215bcf9487bea3f467b
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 2 13:52:11 2021 +0300
sql: introduce mem_str()
This patch introduces mem_str() which allows to receive value of MEM as
a string. Due to the limitations of static_alloc(), this function cannot
be used to safely retrieve a value of MEM converted to string. This
function is suitable for debugging, displaying the value in an error
message, etc.
Part of #5818
commit c14e14f9129b902ba63f90d2beb4b9e233ae2787
Author: Mergen Imeev <imeevma@gmail.com>
Date: Thu Apr 1 15:09:15 2021 +0300
sql: disable unused code in sql/vdbemem.c
This patch disables unused code in sql/vdbemem.c. It will simplify
refactoring.
Part of #5818
commit 91ec46e436507dcec13e1f4d858cbc1b6939a0d1
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Mar 2 10:58:15 2021 +0300
sql: remove unused MEM-related functions
Part of #5818
commit 5bc230c661a89586a144c64084724f04b7e10db3
Author: Mergen Imeev <imeevma@gmail.com>
Date: Tue Apr 6 17:31:47 2021 +0300
sql: refactor port_vdbemem_*() functions
This patch simplifies port_vdbemem_dump_lua() and
port_vdbemem_get_msgpack() functions.
Part of #5818
commit 89db11f40fc88430bbb0995fa502940b72180a61
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 1 18:13:06 2021 +0300
sql: move MEM-related functions to mem.c/mem.h
This patch moves all MEM-related functions in mem.c/mem.h.
Part of #5818
commit c60fff9279e3b146ef38fe365630d5da2d58afef
Author: Mergen Imeev <imeevma@gmail.com>
Date: Thu Mar 4 17:22:45 2021 +0300
sql: remove NULL-termination in OP_ResultRow
There is no need for NULL-termination, since MEMs converted to msgpack
and msgpack do not require NULL-terminated strings.
Needed for #5818
commit bc4b8459c454a4a56e34d4ece6a3f7cc9af9411b
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 22 13:28:36 2021 +0300
sql: disable unused code in sql/legacy.c
This patch disables unused code in sql/legacy.c. It will simplify
refactoring.
Needed for #5818
commit 743d6d40a816eb8d4a07afa69e8682ed72add8ff
Author: Mergen Imeev <imeevma@gmail.com>
Date: Mon Mar 22 13:17:39 2021 +0300
sql: disable unused code in sql/analyze.c
This patch disables unused code in sql/analyze.c. It will simplify
refactoring.
Needed for #5818
commit da9b7ea7ce874dd38112b7519702112952c60e2d
Author: Mergen Imeev <imeevma@gmail.com>
Date: Thu Mar 4 17:17:18 2021 +0300
sql: enhance vdbe_decode_msgpack_into_mem()
Currently, vdbe_decode_msgpack_into_mem() creates a MEM that is not
properly initialized in case msgpack contains MP_EXT, MP_MAP, or
MP_ARRAY fields. This patch makes it so that after execution of this
function, all created MEMs are properly initialized.
Closes #5011
Closes #5704
Closes #5913
Needed for #5818
commit c0e253fe56612030f45dc08540c53f2395c81531
Author: Alexander Turenko <alexander.turenko@tarantool.org>
Date: Tue Apr 13 12:16:10 2021 +0300
security: update libcurl from 7.71.1 to 7.76.0
The reason of the update is to protect us against possible MITM attack
from a malicious HTTPS proxy server with trusted certificate when TLS
1.3 is used (CVE-2021-22890, [1]). libcurl versions prior to 7.76.0 can
skip a TLS handshake with a target host in this circumstances.
Other vulnerabilities fixed in the (7.71.1; 7.76.0] version range do not
look relevant to our built-in http client. See [2] for the full list.
The CMake version requirement is updated from 3.1 to 3.2, because curl's
CMakeLists.txt has the following clause at beginning:
| cmake_minimum_required(VERSION 3.2...3.16 FATAL_ERROR)
(It was there in vanilla curl 7.71.1 too and we had to remove it in
order to support CMake 2. Now we don't support CMake 2, so it is good
time to get rid of the extra patch upward vanilla curl repository.)
According to the CMake versions table in
8a7702b112edf81a33f79cfe0094a551cd008b74 ('github-ci: purge Debian
Jessie from CI'), CMake 3.2+ is available on all supported OSes.
[1]: https://curl.se/docs/CVE-2021-22890.html
[2]: https://curl.se/docs/vulnerabilities.html
@TarantoolBot document
Title: Now we require CMake 3.2 to build tarantool
In https://github.com/tarantool/doc/issues/1780 we requested to update
the CMake minimum version to 3.1. Now it is time for 3.2. See details in
the linked commit.
Please, update the 'Building from source' manual.
commit 38d0b0c1f62886f40db1279da64f2fac75473119
Author: Roman Khabibov <roman.habibov@tarantool.org>
Date: Sun Dec 20 13:08:40 2020 +0500
build: install libcurl headers
Ship libcurl headers to system path "${PREFIX}/include/tarantool"
in the case of libcurl included as bundled library or static
build. It is needed to use SMTP client with tarantool's libcurl
instead of system libcurl.
See related issue: https://github.com/tarantool/smtp/issues/24
Closes #4559
commit 4bde1dbc0fddbff04889f4a483dd97083435dd3a
Author: Roman Khabibov <roman.habibov@tarantool.org>
Date: Tue Jan 19 00:45:32 2021 +0300
build: enable smtp
Enable smtp and smtps protocols in bundled libcurl. It is needed
to use SMTP client with tarantool's libcurl instead of system
libcurl.
See related issue: https://github.com/tarantool/smtp/issues/24
Part of #4559
commit c5049edcdc99a0b137374b4276c6c7065332c501
Author: Alexander V. Tikhonov <avtikhon@tarantool.org>
Date: Wed Apr 14 18:31:53 2021 +0300
test: ASAN leak on feedback demon call using cURL
Found that after patch set of 5 commits:
aa97a18575a5fba29550e131c4764b81841612b9 ("feedback_daemon: count and report some events")
e9c9832a0f3635fa0179a2cc1bfd334dea8cdc64 ("feedback_daemon: generate report right before sending")
bc15e0f070c89e00cd76b0432d906697aae25202 ("feedback_daemon: send feedback on server start")
670acf0d163436a6853dc54d521ff7f037acba8d ("feedback_daemon: rename `send_test` to `send`")
c5d595bcae1a539604895b186cdda86a212909e8 ("feedback_daemon: include server uptime in the report")
began to leak feedback demon on using cURL. To avoid of it decided to
add LSAN suppresions:
1. for /lib/x86_64-linux-gnu/libcrypto.so
leak:libcrypto.so*
2. for /lib/x86_64-linux-gnu/libssl.so
leak:libssl.so*
Needed for tarantool/tarantool-qa#116
commit 8a09e18ce723ffaed85f04d27a388ffdb681308f
Author: Sergey Kaplun <skaplun@tarantool.org>
Date: Wed Apr 7 18:16:50 2021 +0300
luajit: bump new version
LuaJIT submodule is bumped to introduce the following changes:
* test: disable too deep recursive PUC-Rio test
* test: disable PUC-Rio hanging GC test
* test: disable PUC-Rio test checking -h option
* test: disable PUC-Rio test for checking arg layout
* test: disable PUC-Rio tests for several -l options
* test: disable PUC-Rio test for syntax level error
* test: disable PUC-Rio test for non-ascii variable
* test: disable PUC-Rio test for fast function name
* test: disable PUC-Rio test for variables in error
* test: disable PUC-Rio test for getfenv in tailcall
* test: remove string.gfind assert in PUC-Rio test
* test: use math.fmod in PUC-Rio tests
* test: disable locale-dependent PUC-Rio tests
* test: adapt PUC-Rio test for %q in string.format
* test: disable PUC-Rio test for per-coroutine hooks
* test: adapt PUC-Rio test with activeline check
* test: disable PUC-Rio test for tailcall info
* test: adapt PUC-Rio test with count hooks
* test: adapt PUC-Rio test for debug in vararg func
* test: adapt PUC-Rio tests with vararg functions
* test: disable PUC-Rio suite tests for line hook
* test: adapt PUC-Rio tests counting GC steps
* test: disable PUC-Rio tests for bytecode header
* test: disable PUC-Rio tests confused by -v output
* test: adapt PUC-Rio test for arg presence
* test: remove quotes in progname from PUC-Rio
* test: adapt PUC-Rio suite for out-of-source build
* test: build auxiliary C libs from PUC-Rio Lua 5.1
* test: add PUC-Rio Lua 5.1 test suite
Within this changeset PUC-Rio Lua 5.1 suite[1] is added to Tarantool
testing. Considering Tarantool specific changes in runtime the suite
itself is adjusted in LuaJIT submodule.
<test/luajit-test-init.lua> pretest runner is adjusted by setting custom
`_loadstring()` and `_dofile()` global function to run tests correctly
for out-of-source build.
Also, this patch excludes PUC-Rio-Lua-5.1 test suite from ASAN checks.
[1]: https://www.lua.org/tests/lua5.1-tests.tar.gz
Closes #5845
Closes #5686
Closes #5694
Closes #5701
Closes #5708
Closes #5710
Closes #5711
Closes #5712
Part of #4473
Reviewed-by: Igor Munkin <imun@tarantool.org>
Signed-off-by: Igor Munkin <imun@tarantool.org>
commit 07b7062b4891151de09673aca21ba58c64fbf777
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Tue Apr 13 13:44:58 2021 +0300
box: add successor argument in replace function
Add an argument which will receive successor of the inserted tuple.
Part of #5628
commit bf3c3b4bf7a1580aff816be5df6a1fd371c375e8
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Tue Apr 13 11:22:09 2021 +0300
bps: add successor argument to insertion function
Common insertion function in BSP tree can make a replacement
or insertion.
When a replacement happens, a replaced element is returned via
special agrument.
Let's add another argument of the function for the case of
insetion: an element before which then insertion happened will
be returned via that argument.
Part of #5628
commit df7af44301399678e1fef2efe8dcd683aa3e8b63
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Wed Apr 14 09:51:06 2021 +0300
txm: fix space len and count calculation
When MVCC engine is enabled, an index can contain temporary
non-committed tuples that are not visible outside their
transaction.
That's why internal index size can be greater than visible by
standalone observer.
In order to fix space:count and index:count etc we have to use
special adjustment from MVCC engine.
Fixes #5972
commit 27a496b7e222ea4aa234d5187a5d4202c2a7756f
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Tue Apr 13 14:37:26 2021 +0300
txm: implement point hole trackers
TX manager has a problem, kwown as phantom read.
If a transaction reads a value by key and founds nothing, even
no dirty tuples, then the read is not recorded anywhere.
Implement a special storage for point reads that founds nothing,
and use it in conflict resolution.
Part of #5628
commit 26428c0c5a4866a6908c206a4794ccf4fed12cce
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Tue Apr 13 14:45:26 2021 +0300
box: add unique ID to index
Part of #5628
commit 45e828d90f274d9fb5a74977986053b20c96d22f
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Tue Apr 13 05:39:12 2021 +0300
txm: fix a bug that happened with alter
See test for details.
Part of #5628
commit 8f502d1f91afbabdf6288189bfc71efe7358694b
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Fri Apr 9 19:27:18 2021 +0300
txm: simlification: don't link stories with clean tuples
Before this patch there was a way to build history of a key when
the terminal element in list is a clean tuple, not its
story. That was a kind of optimization desigend to avoid excess
story creation.
Unfortunately that kind of otimization is possible to use only
for one-index spaces, and even in that case it doesn't work.
Let's remove optimization that doesn't work.
Part of #5628
commit 7d511a1a94fd3e241d6c050554fc132387fc16ab
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Sun Apr 11 19:16:33 2021 +0300
box: remove unused check
Since DUP_REPLACE mode is always used with given old_tuple,
there's no reason to check in runtime
Part of #5628
commit 1210648e46b9a3682e7ba279643e175987d83a38
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Fri Dec 25 08:34:05 2020 +0300
txm: code reorder
Part of #5628
commit c6a8b384080b10a7e0513a76c656a9b81df0b905
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Tue Dec 22 16:15:10 2020 +0300
txm: minor changes
Part of #5628
commit 8474a96e9b2f66f0ad4343fe8fce9311f43927e2
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Thu Apr 1 10:58:54 2021 +0300
txm: completely switch off MVCC engine for ephemeral spaces
Since some operations with an ephemeral space are done without a
transaction - we cannot make it consistent and should not even try.
Part of #5628
commit 535e2d4288b58d4ee7cd85000d13b39170b23edb
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Wed Apr 7 18:54:55 2021 +0300
box: fix indents
Part of #5958
commit 453b14d5b894a9d5be4360e6c93e1e37a06215e7
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Wed Apr 7 18:46:47 2021 +0300
box: store request type implicitly instead of faking xrow
Part of #5958
commit 6e6554cb0fe29ee04a7141c5e43ed428a8bebd09
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Wed Apr 7 18:04:56 2021 +0300
box: store request type in 16-bit interger
It seems that it'll never need it greater that UINT16_MAX.
Maybe it will allow us to save some memory.
Part of #5958
commit 4dbd46442a77e2f65d2ae2c0e413f6493fa11374
Author: Aleksandr Lyapunov <alyapunov@tarantool.org>
Date: Wed Apr 14 08:03:47 2021 +0300
txm: switch to normal mode earlier durion recovery with MVCC
Usually during final recovery, when .xlog files are loaded and
applied, a special memtx_space_replace_primary_key function is
used for faster recovery. The function only updates primary key
while all other indexes are scheduled to build at the end of
recovery.
This approach doesn't work with MVCC engine and synchro spaces:
When transactions appear in xlog and they may or may not be
committed, we should process the in normal mode throught all
indexes in order to leave transaction history in each index.
Fixes #5610
Fixes #5973
commit 57ab8c2e7dbc68b031fad652cb5d917f9a2760fe
Author: Alexander V. Tikhonov <avtikhon@tarantool.org>
Date: Mon Apr 12 13:09:00 2021 +0300
github-ci: add more cleanup for running WS
Found that 'actions/checkout' does not remove all temporary files from
previous runs in submodules [1], it runs only 'git clean --xffd' [2]:
libgcov profiling error:/home/ubuntu/actions-runner/_work/tarantool/tarantool/src/lib/small/CMakeFiles/small.dir/small/small_class.c.gcda:overwriting an existing profile data with a different timestamp
To avoid of it added:
git submodules foreach --recursive 'git clean -xffd'
to 'actions/environment' which is run after each 'actions/checkout'.
Part of #5986
[1]: https://github.com/tarantool/tarantool/runs/2318244478?check_suite_focus=true#step:5:4012
[2]: https://github.com/actions/checkout/issues/358
commit ef55e4881c0b38f7520ebb5817d47e785ff55c45
Author: Sergey Kaplun <skaplun@tarantool.org>
Date: Wed Apr 14 16:25:49 2021 +0300
luajit: bump new version
LuaJIT submodule is bumped to introduce the following changes:
* tools: introduce --leak-only memprof parser option
Within this changeset the new Lua module providing post-processing
routines for parsed memory events is introduced:
* memprof/process.lua: post-process the collected events
The changes provide an option showing only heap difference. One can
launch memory profile parser with the introduced option via the
following command:
$ tarantool -e 'require("memprof")(arg)' - --leak-only filename.bin
Closes #5812
Reviewed-by: Igor Munkin <imun@tarantool.org>
Signed-off-by: Igor Munkin <imun@tarantool.org>
commit afcef769baea699c80bf25593aee8fd178f4cb50
Author: Cyrill Gorcunov <gorcunov@gmail.com>
Date: Wed Sep 30 13:28:51 2020 +0300
test: add box.lib test
Part-of #4642
Co-developed-by: Vladislav Shpilevoy <v.shpilevoy@tarantool.org>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
commit f463b5fa1bfc19049d145e495ac4a5f2a3673590
Author: Cyrill Gorcunov <gorcunov@gmail.com>
Date: Tue Feb 9 00:55:35 2021 +0300
box: implement box.lib module
Currently to run "C" function from some external module one
have to register it first in "_func" system space. This is
a problem if node is in read-only mode (replica).
Still people would like to have a way to run such functions
even in ro mode. For this sake we implement "box.lib" lua module.
Unlike `box.schema.func` interface the `box.lib` does not defer module
loading procedure until first call of a function. Instead a module
is loaded immediately and if some error happens (say shared
library is corrupted or not found) it pops up early.
The need of use stored C procedures implies that application is
running under serious loads most likely there is modular structure
present on Lua level (ie same shared library is loaded in different
sources) thus we cache the loaded library and reuse it on next
load attempts. To verify that cached library is up to day the
module_cache engine test for file attributes (device, inode, size,
modification time) on every load attempt.
Since both `box.schema.func` and `box.lib` are using caching to minimize
module loading procedure the pass-through caching scheme is
implemented:
- box.lib relies on module_cache engine for caching;
- box.schema.func does snoop into box.lib hash table when attempt
to load a new module, if module is present in box.lib hash then
it simply referenced from there and added into own hash table;
in case if module is not present then it loaded from the scratch
and put into both hashes;
- the module_reload action in box.schema.func invalidates module_cache
or fill it if entry is not present.
Closes #4642
Co-developed-by: Vladislav Shpilevoy <v.shpilevoy@tarantool.org>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
@TarantoolBot document
Title: box.lib module
Overview
========
`box.lib` module provides a way to create, delete and execute
`C` procedures from shared libraries. Unlike `box.schema.func`
methods the functions created with `box.lib` help are not persistent
and live purely in memory. Once a node get turned off they are
vanished. An initial purpose for them is to execute them on
nodes which are running in read-only mode.
Module functions
================
`box.lib.load(path) -> obj | error`
-----------------------------------
Loads a module from `path` and return an object instance
associate with the module, otherwise an error is thrown.
The `path` should not end up with shared library extension
(such as `.so`), only a file name shall be there.
Possible errors:
- IllegalParams: module path is either not supplied
or not a string.
- SystemError: unable to open a module due to a system error.
- ClientError: a module does not exist.
- OutOfMemory: unable to allocate a module.
Example:
``` Lua
-- Without error handling
m = box.lib.load('path/to/library)
-- With error handling
m, err = pcall(box.lib.load, 'path/to/library')
if err ~= nil then
print(err)
end
```
`module:unload() -> true | error`
---------------------------------
Unloads a module. Returns `true` on success, otherwise an error
is thrown. Once the module is unloaded one can't load new
functions from this module instance.
Possible errors:
- IllegalParams: a module is not supplied.
- IllegalParams: a module is already unloaded.
Example:
``` Lua
m = box.lib.load('path/to/library')
--
-- do something with module
--
m:unload()
```
If there are functions from this module referenced somewhere
in other places of Lua code they still can be executed because
the module continue sitting in memory until the last reference
to it is closed.
If the module become a target to the Lua's garbage collector
then unload is called implicitly.
`module:load(name) -> obj | error`
----------------------------------
Loads a new function with name `name` from the previously
loaded `module` and return a callable object instance
associated with the function. On failure an error is thrown.
Possible errors:
- IllegalParams: function name is either not supplied
or not a string.
- IllegalParams: attempt to load a function but module
has been unloaded already.
- ClientError: no such function in the module.
- OutOfMemory: unable to allocate a function.
Example:
``` Lua
-- Load a module if not been loaded yet.
m = box.lib.load('path/to/library')
-- Load a function with the `foo` name from the module `m`.
func = m:load('foo')
```
In case if there is no need for further loading of other
functions from the same module then the module might be
unloaded immediately.
``` Lua
m = box.lib.load('path/to/library')
func = m:load('foo')
m:unload()
```
`function:unload() -> true | error`
-----------------------------------
Unloads a function. Returns `true` on success, otherwise
an error is thrown.
Possible errors:
- IllegalParams: function name is either not supplied
or not a string.
- IllegalParams: the function already unloaded.
Example:
``` Lua
m = box.lib.load('path/to/library')
func = m:load('foo')
--
-- do something with function and cleanup then
--
func:unload()
m:unload()
```
If the function become a target to the Lua's garbage collector
then unload is called implicitly.
Executing a loaded function
===========================
Once function is loaded it can be executed as an ordinary Lua call.
Lets consider the following example. We have a `C` function which
takes two numbers and returns their sum.
``` C
int
cfunc_sum(box_function_ctx_t *ctx, const char *args, const char *args_end)
{
uint32_t arg_count = mp_decode_array(&args);
if (arg_count != 2) {
return box_error_set(__FILE__, __LINE__, ER_PROC_C, "%s",
"invalid argument count");
}
uint64_t a = mp_decode_uint(&args);
uint64_t b = mp_decode_uint(&args);
char res[16];
char *end = mp_encode_uint(res, a + b);
box_return_mp(ctx, res, end);
return 0;
}
```
The name of the function is `cfunc_sum` and the function is built into
`cfunc.so` shared library.
First we should load it as
``` Lua
m = box.lib.load('cfunc')
cfunc_sum = m:load('cfunc_sum')
```
Once successfully loaded we can execute it. Lets call the
`cfunc_sum` with wrong number of arguments
``` Lua
cfunc_sum()
| ---
| - error: invalid argument count
```
We will see the `"invalid argument count"` message in output.
The error message has been set by the `box_error_set` in `C`
code above.
On success the sum of arguments will be printed out.
``` Lua
cfunc_sum(1, 2)
| ---
| - 3
```
The functions may return multiple results. For example a trivial
echo function which prints arguments passed in.
``` Lua
cfunc_echo(1,2,3)
| ---
| - 1
| - 2
| - 3
```
Module and function caches
==========================
Loading a module is relatively slow procedure because operating
system needs to read the library, resolve its symbols and etc.
Thus to speedup this procedure if the module is loaded for a first
time we put it into an internal cache. If module is sitting in
the cache already and new request to load comes in -- we simply
reuse a previous copy. In case if module is updated on a storage
device then on new load attempt we detect that file attributes
(such as device number, inode, size, modification time) get changed
and reload module from the scratch. Note that newly loaded module
does not intersect with previously loaded modules, the continue
operating with code previously read from cache.
Thus if there is a need to update a module then all module instances
should be unloaded (together with functions) and loaded again.
Similar caching technique applied to functions -- only first function
allocation cause symbol resolving, next ones are simply obtained from
a function cache.
commit 4a6b3a85be25c8b42dca3f5cd09851bcdfb86383
Author: Cyrill Gorcunov <gorcunov@gmail.com>
Date: Thu Apr 1 18:58:26 2021 +0300
box/schema.func: switch to new module api
Since we have low level module api now we can switch
the box.schema.func code to use it. In particular we
define schema_module structure as a wrapper over the
modules api -- it carries a pointer to general module
structure.
Because of modules reload functionality the schema modules
carry own cache of modules instances. Thus to make overall
code somehow close to modules api we do:
1) cache operations are renamed to cache_find/put/update/del;
2) C functions are switched to use module_func low level api;
3) because low level modules api carry own references we can
take no explicit reference when calling a function.
Part-of #4642
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
commit d3b835de4d1507f18a2b63678e69720e1637b276
Author: Cyrill Gorcunov <gorcunov@gmail.com>
Date: Wed Feb 24 00:35:03 2021 +0300
box/module_cache: introduce modules subsystem
The modules subsystem hides some low-level operations under API.
In particular the modules subsystem is responsible for
- modules lookup in Lua's "package.search" storage
- modules caching to eliminate expensive load procedure
- function symbol resolving
Because naming is intersecting with current module functions
sitting in box/func, lets rename the later to schema_module
prefix. We will use this prefix in next patches to point the
modules in box.schema.func are just a particular user of
the general modules engine.
Part-of #4642
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
commit a5cd73aee7912d5c1b23e5de9b000da6678c361e
Author: Cyrill Gorcunov <gorcunov@gmail.com>
Date: Tue Feb 23 10:46:13 2021 +0300
box/func: module_reload -- drop redundant argument
The only purpose of the module argument is to
notify the caller that the module doesn't exist.
Lets simplify the code and drop this argument.
Part-of #4642
Acked-by: Serge Petrenko <sergepetrenko@tarantool.org>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
commit b9f2bf4e9b36fdc4ab547a6fcf63054e18d8da8c
Author: Cyrill Gorcunov <gorcunov@gmail.com>
Date: Thu Apr 1 18:03:08 2021 +0300
box/func: fix modules functions restore
In commit 96938fafb (Add hot function reload for C procedures)
an ability to hot reload of modules has been introduced.
When module is been reloaded his functions are resolved to
new symbols but if something went wrong it is supposed
to restore old symbols from the old module.
Actually current code restores only one function and may
crash if there a bunch of functions to restore. Lets fix it.
Fixes #5968
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
commit b259e9306f717883026520c975f2610c9fbc7d43
Author: Vladislav Shpilevoy <v.shpilevoy@tarantool.org>
Date: Mon Apr 5 19:05:59 2021 +0200
applier: process synchro rows after WAL write
Applier used to process synchronous rows CONFIRM and ROLLBACK
right after receipt before they are written to WAL.
That led to a bug that the confirmed data became visible, might be
accessed by user requests, then the node restarted before CONFIRM
finished its WAL write, and the data was not visible again. That
is just like if it would be rolled back, which is not acceptable.
Another case - CONFIRM WAL write could simply fail due to any
reason (no disk space, OOM), but the transactions would remain
confirmed anyway.
Also that produced some hacks in the limbo's code to support the
confirmation and rollback of transactions not yet written to WAL.
The patch makes the synchro rows processed only after they are
written to WAL. Although the 'rollback' case above might still
happen if the xlogs were in the kernel caches, and the machine was
powered off before they were flushed to disk. But that is not
related to qsync specifically.
To handle the synchro rows after WAL write the patch makes them go
to WAL in a blocking way (journal_write() instead of
journal_write_try_async()). Otherwise it could happen that a
CONFIRM/ROLLBACK is being written to WAL and would clear the limbo
afterwards, but a new transaction arrives with a different owner,
and it conflicts with the current limbo owner.
Closes #5213
commit 645141133478fcd10b1ac4bd2e993a84da6a7750
Author: mechanik20051988 <mechanik20.05.1988@gmail.com>
Date: Tue Apr 13 15:16:39 2021 +0300
update: allow update(delete) absent nullable fields
In previous patch update(insert) operation for absent nullable fields
was allowed. This patch allows to update(delete) operation for absent
nullable fileds.
Closes #3378
commit 2bb373b9f5589cbda667f3d7465aae23075e49bc
Author: Mary Feofanova <m.feofanova@tarantool.org>
Date: Tue Mar 16 01:50:51 2021 +0300
update: allow update absent nullable fields
Update operations could not insert with gaps. This patch changes
the behavior so that the update operation fills the missing fields
with nulls.
Part of #3378
@TarantoolBot document
Title: Allow update absent nullable fields
Update operations could not insert with gaps. Changed the behavior
so that the update operation fills the missing fields with nulls.
For example we create space `s = box.schema.create_space('s')`,
then create index for this space `pk = s:create_index('pk')`, and
then insert tuple in space `s:insert{1, 2}`. After all of this we
try to update this tuple `s:update({1}, {{'!', 5, 6}})`. In previous
version this operation fails with ER_NO_SUCH_FIELD_NO error, and now
it's finished with success and there is [1, 2, null, null, 6] tuple in
space.
commit 4b0f81ea71b898c742ba4c340d86b22e9ffc768f
Author: Mary Feofanova <m.feofanova@tarantool.org>
Date: Thu Mar 18 11:07:11 2021 +0300
test: change compact.test.lua bad upsert
Prepare this test for upcoming #3378 fix:
bad upserts will become good, so we need
another way to do them.
commit 2ede3be3564f04ba2c80ed454a2ba6b0ca11450d
Author: mechanik20051988 <mechanik20.05.1988@gmail.com>
Date: Tue Apr 13 22:18:55 2021 +0300
iproto: implement ability to run multiple iproto threads
There are users that have specific workloads where iproto thread
is the bottleneck of throughput: iproto thread's code is 100% loaded
while TX thread's core is not. For such cases it would be nice to have
a capability to create several iproto threads.
Closes #5645
@TarantoolBot document
Title: implement ability to run multiple iproto threads
Implement ability to run multiple iproto threads, which is useful
in some specific workloads where iproto thread is the bottleneck
of throughput. To specify count of iproto threads, user should used
iproto_threads option in box.cfg. For example if user want to start
8 iproto threads, he must enter `box.cfg{iproto_threads=8}`. Default
iproto threads count == 1. This option is not dynamic, so user can't
change it after first setting, until server restart. Distribution of
connections per threads is managed by OS kernel.
commit 440a4f3060358e315dbd8da5f28d775ed940baf6
Author: mechanik20051988 <mechanik20.05.1988@gmail.com>
Date: Fri Apr 9 13:42:43 2021 +0300
iproto: fix error with struct rmean
There was two problems with struct rmean:
- For correct access for rmean struct fields, this struct should be
created in tx thread.
- In case when rmean_new return NULL in net_cord_f tarantool hangs
and does not terminate in any way except on SIGKILL.
Also net_slabc cache was not destroyed. Moved allocation and deallocation
of rmean structure to iproto_init/iproto_free respectively. Added
slab_cache_destroy for net_slabc for graceful resource releases.
commit 91d8e7ae2a43a6b7898c1c08b81d1a6fd9dfd907
Author: mechanik20051988 <mechanik20.05.1988@gmail.com>
Date: Fri Mar 26 18:14:33 2021 +0300
Fix a potential error with the rmean structure fields access
The fields of the rmean structure can be accessed from
multiple threads, so we must use atomic operations to get/set
fields in this structure. Also in the comments to the functions
i wrote in which threads they should be called to correctly access
the fields of the rmean structure.
commit 820d2be666f5eb235e66c173142dc07a799decc4
Author: Alexander Turenko <alexander.turenko@tarantool.org>
Date: Tue Apr 6 12:59:55 2021 +0300
build: fix configuring using cmake3 command
`cmake` command was hardcoded for configuring libcurl, however only
`cmake3` may be installed in a system. Now we use the same cmake command
for configuring libcurl as one that is used for configuring tarantool
itself.
The problem exists since 2.6.0-196-g2b0760192 ('build: enable cmake in
curl build').
Fixes #5955
