Historically, we join a new replica off the last checkpoint. As a
result, we must always keep the last memtx snapshot and all vinyl data
files corresponding to it. Actually, there's no need to use the last
checkpoint for joining a replica. Instead we can use the current read
view as both memtx and vinyl support it. This should speed up the
process of joining a new replica, because we don't need to replay all
xlogs written after the last checkpoint, only those that are accumulated
while we are relaying the current read view. This should also allow us
to avoid creating a snapshot file on bootstrap, because the only reason
why we need it is allowing joining replicas. Besides, this is a step
towards decoupling the vinyl metadata log from checkpointing in
particular and from xlogs in general.

Closes #1271

This commit does not change a user visible behaviour. It refactors pwd
module to explicitly divide code that fetches data from passwd / group
databases from code that performs deserialization of the data to Lua
tables.

The idea of splitting of those actions appears when it was observed that
a call of getpw() / getgr() leads to problems on some systems when it is
invoked during passwd / group database traveral.

Now it is more obvious that we don't call getpw() during passwd
traversal and getgr() during group traveral.

Follows up #4428 and #4447.

LTO build fails on warning message:
In file included from /tarantool/src/lib/core/diag.h:33:0,
                 from /tarantool/src/box/engine.h:36,
                 from /tarantool/src/box/memtx_engine.h:40,
                 from /tarantool/src/box/memtx_engine.c:31:
/tarantool/src/box/memtx_engine.c: In function
'metmx_tuple_chunk_delete':
/tarantool/src/trivia/util.h:201:49: error: initialization from
incompatible pointer type [-Werror]
  const typeof( ((type *)0)->member  ) *__mptr = (ptr); \
                                                 ^
/tarantool/src/box/memtx_engine.c:1115:3: note: in expansion of macro
'container_of'
   container_of((typeof(tuple_chunk->data) *)data,
   ^
/tarantool/src/trivia/util.h:201:49: error: (near initialization for
'tuple_chunk') [-Werror]
  const typeof( ((type *)0)->member  ) *__mptr = (ptr); \
                                                 ^
/tarantool/src/box/memtx_engine.c:1115:3: note: in expansion of macro
'container_of'
   container_of((typeof(tuple_chunk->data) *)data,

Closes #4438

In a configuration with several read-only and read-write instances, if
replication_connect_quorum is not greater than the amount of read-only
instances and replication_connect_timeout happens to be small enough
for some read-only instances to form a quorum and exceed the timeout
before any of the read-write instaces start, all these read-only
instances will choose themselves a read-only bootstrap leader.
This 'leader' will successfully bootstrap itself, but will fail to
register any of the other instances in _cluster table, since it isn't
writeable. As a result, some of the read-only instances will just die
unable to bootstrap from a read-only bootstrap leader, and when the
read-write instances are finally up, they'll see a single read-only
instance which managed to bootstrap itself and now gets a
REPLICASET_UUID_MISMATCH error, since no read-write instance will
choose it as bootstrap leader, and will rather bootstrap from one of
its read-write mates.

The described situation is clearly not what user has hoped for, so
throw an error, when a read-only instance tries to initiate the
bootstrap. The error will give the user a cue that he should increase
replication_connect_timeout.

Closes #4321

@TarantoolBot document
Title: replication: forbid to bootstrap read-only masters.

It is no longer possible to bootstrap a read-only instance in an emply
data directory as a master. You will see the following error trying to
do so:
```
ER_BOOTSTRAP_READONLY: Trying to bootstrap a local read-only instance as master
```
Now if you have a fresh instance, which has
`read_only=true` in an initial `box.cfg` call, you need to set up
replication from an instance which is either read-write, or has your
local instance's uuid in its `_cluster` table.

In case you have multiple read-only and read-write instances with
replication set up, and you still see the aforementioned error message,
this means that none of your read-write instances managed to start
listening on their port before read_only instances have exceeded the
`replication_connect_timeout`. In this case you should raise
`replication_connect_timeout` to a greater value.

rows_per_wal does not allow to properly limit size of WAL logs.
A user would need to estimate size of each WAL row to do that.
Now WAL supports option wal_max_size allowing to limit WAL size
much more precise.

Part of #3762

Added support for the following curl options:
* CURLOPT_PROXY
* CURLOPT_PROXYPORT
* CURLOPT_PROXYUSERPWD
* CURLOPT_NOPROXY

@TarantoolBot document
Title: httpc: proxy server options

Use 'proxy' option to specify the proxy server host or IP-address
(optionally may be prefixed with a scheme - e.g. http:// or https://).
'proxy_port' and 'proxy_user_pwd' options may be used to specify the
proxy port (443 for https proxy and 1080 for others by default) and
user credentials (format: [user name]:[password]) respectively.

If 'proxy' option is not set a value from the corresponding
environment variable will be used. Environment variable names are:
'http_proxy', 'https_proxy', 'ftp_proxy' etc. 'all_proxy' variable
is used if no protocol specific proxy was set.

Setting 'proxy' option to an empty string will explicitly disable the
use of a proxy, even if there is an environment variable set for it.

Set 'no_proxy' option to specify a comma separated list of hosts that
do not require a proxy to get reached, even if one is specified by
'proxy' option (or the corresponding environment variable). The only
wildcard available is a single * character, which matches all hosts,
and effectively disables the proxy. 'no_proxy' environment variable
will be used if this option is not set.

Setting 'no_proxy' option to an empty string will explicitly enable
the proxy for all host names, even if there is an environment variable
set for it.

See:
https://curl.haxx.se/libcurl/c/CURLOPT_PROXY.html
https://curl.haxx.se/libcurl/c/CURLOPT_PROXYPORT.html
https://curl.haxx.se/libcurl/c/CURLOPT_USERPWD.html
https://curl.haxx.se/libcurl/c/CURLOPT_NOPROXY.html

We know that madvise(MADV_DONTDUMP) is present
on linux based platforms only (macos doesn't
support it at all, freebsd requires MADV_NOCORE
or similar which is unsupported by now) thus
we should not print a warning on other systems
to not spam users.

Fixes #4464

Closes #2200
Closes #4113
Closes #2233

@TarantoolBot document
Title: The box.internal.sql_function_create is forbidden
Legacy mechanism box.internal.sql_function_create to make some
Lua function available in SQL is forbidden now.

To make some function available in SQL you need to use
box.schema.func.create() mechanism: you need to specify
1) function language and language-specific options(e.g. you
   are able to define a persistent Lua function)
2) whether this function is_deterministic or not: deterministic
   functions allows to generate more efficient SQL VDBE bytecode
   so you better specify it when it is true
3) the function returns type: a Tarantool type string describes
   a type of value returned by function
4) param_list - a table of Tarantool's types strings desccribe
   function argument types
5) exports - a table of Tarantool's frontends where this function
   should be available ('LUA' by default). You need to specify
   {'LUA', 'SQL'} to make function available both in SQL requests
   and visible in box.func folder

Example:
-- Case1: C function
-- function1.so has int divide() symbol
box.schema.func.create("function1.divide", {language = 'C',
			returns = 'number',
			param_list = {'number', 'number'},
			is_deterministic = true,
			exports = {'LUA', 'SQL'}})
box.execute('SELECT "function1.divide"(6, 3)')
- metadata:
  - name: '"function1.divide"(6, 3)'
    type: number
  rows:
  - [2]
box.schema.func.drop("function1.divide")

-- Case2: Persistent Lua function
box.schema.func.create("SUMMARIZE", {language = 'LUA',
			returns = 'number',
			body = 'function (a, b) return a + b end',
			param_list = {'number', 'number'},
			is_deterministic = true,
			exports = {'LUA', 'SQL'}})
box.execute('SELECT summarize(1, 2)')
- metadata:
  - name: summarize(1, 2)
    type: number
  rows:
  - [3]
box.schema.func.drop("summarize")

Moreover there is a special predefined Lua function LUA that
allows to evaluate a custom Lua expressions in SQL.
You need to pass a string in form "return ...." to LUA function
that returns more than one value of any type.

Example:
box.execute('SELECT lua(\'return 1 + 1\')')
- metadata:
  - name: lua('return 1 + 1')
    type: any
  rows:
  - [2]
box.execute('SELECT lua(\'return box.cfg.memtx_memory\')')
- metadata:
  - name: lua('return box.cfg.memtx_memory')
    type: any
  rows:
  - [268435456]

Now it is possible to move all SQL builtin functions to
Tarantool's function hash. An existent FuncDef function
representation was replaced with func_sql_builtin class.
It has a sql-specific method :call and :finalize, while
port API call is not supported and protected with stubs.

This patch removes FuncDef hash and sql_function_create endpoint,
but doesn't introduce something instead. Therefore few affected
tests are disabled. A required functionality would be fixed in
the next patch.

Following tests using sql_create_function are broken now.
They are going to be fixed in the next commit:
 sql-tap/alias.test.lua sql-tap/check.test.lua
 sql-tap/func5.test.lua sql-tap/lua_sql.test.lua
 sql-tap/subquery.test.lua sql-tap/trigger9.test.lua
 sql/errinj.result sql/errinj.test.lua
 sql/func-recreate.test.lua

Part of #2200, #4113, #2233

@TarantoolBot document
Title: SQL builtins priveleges

All SQL built-ins are executed on SQL privilege level that is
undefined yet.

Now all SQL builtins are unique, i.e. each function has the only
hash table entry. This technique requires to remove static argc
checks for some builtins(substr, round, like, trim, count,
group_concat). Now they raise a runtime error in case of invalid
usage.

Updated error messages correspondingly, to provide more
informative messages in such cases. Taking into account upcoming
changes, all built-in functions names are uppercased.

Needed for #2200, #4113, #2233

Replaced sql function flag SQL_FUNC_MINMAX with couple new flags
SQL_FUNC_MIN and SQL_FUNC_MAX. This allows to distinguish MIN and
MAX function by flag instead of using user_data context.

This allows to delete user_data field in FundDef object in
further refactoring.

Needed for #2200, #4113, #2233

Changed sql_vdbe_mem_alloc_region routine name to the
vdbe_mem_alloc_blob_region because we are going to introduce
a new function with a really similar (but more appropriate) name
vdbemem_alloc_on_region in following patch. Due to the fact that
the function used only in vdbe.c module it has been moved there
and has been marked as "static".

Needed for #2200, #4113, #2233

Before a46b5200 SQL implementation featured min()/max() functions
overloading: if one argument was passed, then aggregate version would be
invoked; otherwise - scalar one. We decided to get rid of it and rename
scalar version to LEAST()/GREATEST() correspondingly. However, assertion
inside their implementations has been remained: it verifies that number
of passed arguments is greater than 1. On the other hand, now one can
pass literally any number of arguments to this function, including one
(which results in fired mentioned assertion) and zero (which leads to
NULL dereference in expr.c: these functions are marked with
SQL_FUNC_NEEDCOLL flag, and as a consequence they are assumed to have at
least one argument). Firstly, let's place check that number of passed
arguments more than one. Secondly, let's not assume that functions with
SQL_FUNC_NEEDCOLL must have any arguments.

Closes #4453

This is a followup for 7691154a

In case of throwing client error because of inactive function
we did not destroy used port. It could possibly cause huge
memory leaks as could be seen with top or its analogues when
performing net.box test run in a loop.

Closes #4388

Currently we only enter orphan mode when instance fails to connect to
replication_connect_quorum remote instances during local recovery.
On bootstrap and manual replication configuration change an error is
thrown. We better enter orphan mode on manual config change, and leave
it only in case we managed to sync with replication_connect_quorum
instances.

Closes #4424

@TarantoolBot document
Title: document reaction on error in replication configuration change.

Now when issuing `box.cfg{replication={uri1, uri2, ...}}` and failing to
sync with replication_connect_quorum remote instances, the server will
throw an error, if it is bootstrapping, and just set its state to orphan
in all other cases (recovering from existing xlog/snap files or manually
changing box.cfg.replication on the fly). To leave orphan mode, you may
wait until the server manages to sync with replication_connect_quorum
instances.
In order to leave orphan mode you need to make the server sync with
enough instances. To do so, you may either:
1) set replication_connect_quorum to a lower value
2) reset box.cfg.replication to exclude instances that cannot
   be reached or synced with
3) just set box.cfg.replication to "" (empty string)

In 3-node replica-set, registering with the leader node
does not guarantee the registration record arrives to the second
peer immediately. The joining node may bootstrap faster than
the registration record arrives to the second peer, in which
case replication will fail to create a full mesh.

A follow up on gh-4305, fix failing args.test.py

To make it possible to develop and test related features on
systems without systemd.

WITH_SYSTEMD cmake flag is used to generate systemd related files:
unit, generator script, etc. To keep this behavior and make it possible
to use NOTIFY_SOCKET without other systemd-related stuff,
I added WITH_NOTIFY_SOCKET cmake flag.

It also required some changes to support other OS:

SOCK_CLOEXEC (not available on macOS) flag for socket()
is replaced with `fcntl(fd, F_SETFD, FD_CLOEXEC)` which has the same effect.

MSG_NOSIGNAL flag for sendmsg is also not available on macOS.
However it has SO_NOSIGPIPE flag for setsockopt which disables SIGPIPE.
So it requires different solution for different OS. Inspired by
https://nwat.xyz/blog/2014/01/16/porting-msg_more-and-msg_nosigpipe-to-osx/

Have to reduce send-buffer size to 4MB because larger values
are not supported on macOS by default. This value should be enough
for all systems because notification messages are usually less than 1KB.

Fixes #4436

Tuple fields can be named, accessed by name, indexed by name, but
till this commit field names couldn't be used in update
operations. Now it is possible.

This patch is a teaser of updates by JSON path.

Part of #1261

Rope will not be used by tuple_update directly in
next patches.

Rope library stores alloc, split and free functions by pointer,
that is a huge slog at performance, as any other virtual
function call. There is no reason, why the rope library may not
become a template, except historical ones.

The patch not only removes virtual functions, but also makes rope
deletion almost no-op in case if free() function is not defined.

A second motivation point of that patch is that original rope
structure was too big. Again, because it stored several pointers
to functions. In forthcoming patches on #1261 multiple ropes can
be created per each update, so it makes sense to reduce size of
this structure.

They are always region_aligned_alloc and region pointer. Lets use
them always inside tuple_update.c, with no necessity to pass
explicitly.

The patch is motivated by forthcoming updates by JSON path, which
will strongly complicate and perhaps slow down the tuple_update.c
code. The present patch as well as some next ones should smooth
these problems.

This problem is similar to the one fixed in commit
3c6c1cc9 (lua:fix decimal comparison
with nil)
We should handle box.NULL the same way.

Closes #4454

CentOS 6 and FreeBSD 12 implementations of getpwuid() rewind
setpwent()-{getpwent()}-endpwent() loop to a start that leads to a hang
during pwd.getpwall() invoke. The same is true for a getgrgid() call
during setgrent()-{getgrent()}-endgrent() loop.

The commit modifies pwd module to avoid getpwuid() calls during passwd
database traversal and to avoid getgrgid() calls when traversing groups.

The commit also fixes the important regression on CentOS 6 after
f5d8331e ('lua: workaround
pwd.getpwall() issue on Fedora 29'): tarantool hungs during startup due
to added getpwall() call. This made tarantool unusable on CentOS 6 at
all.

Aside of that the commit fixes another pwd.getgrall() problem: the
function gaves password entries instead of group entries.

Fixes #4428.
Fixes #4447.
Part of #4271.

This change is related to #4391. The objective was to collect additional
information about modules, but it's hard to do without changing API.

This patch will allow to monkey-patch report generation and achieve the
same results without interfering the daemon behavior.

There is a problem with calculating .msg_namelen field
of msghdr struct. Instead of

    .msg_name   = &sa,
    .msg_namelen = sizeof(sa.sun_family) + strlen(sd_unix_path),

it must set as

    .msg_namelen = sizeof(sa) // larger value than current invalid one

It works on linux but when I tried to enable this feature for macOS
it didn't (maybe because of different order of fields in the struct).

Instead of fixing calculation, I've replaced original sendmsg call
with sendto, because it's a convenient shortcut which
simplifies code and can prevent such mistakes.

Required for #4436

This function implements common way of precise comparison between
unsigned integer and floating point values (doubles). Currently, it is
used in tuple comparators, but we need the same thing in SQL. Hence,
let's move it to header containing set of utilities.

To compare floating point values and integers in SQL functions
compare_uint_float() and compare_int_float() are used. Unfortunately,
they contain bug connected with checking border case: that's not correct
to cast UINT64_MAX (2^64 - 1) to double. Proper way is to use exp2(2^64)
or predefined floating point constant. To not bother fixing function
which in turn may contain other tricky places, let's use instead already
verified double_compare_uint64(). So that we have unified way of
integer<->float comparison.

Closes #4422

@TarantoolBot document
Title: Introduce <WITH ENGINE> clause for CREATE TABLE statement

To allow user to specify engine as per table option, CREATE TABLE
statement has been extended with optional <WITH ENGINE = engine_name>
clause. This clause comes at the end of CREATE TABLE statement.
For instance:

CREATE TABLE t_vinyl (id INT PRIMARY KEY) WITH ENGINE = 'vinyl';

Name of engine is considered to be string literal ergo should be
enclosed in single quotation marks and be lower-cased. Note that engine
specified in WITH ENGINE clause overwrites default engine, which is set
via 'pragma sql_default_engine'.

Error logging in engine_find() seems to be redundant: error message is
displayed twice (since its callers always push error on the top).  Let's
remove this duplication.

Name duplicates are allowed for savepoints (both in our SQL
implementation and in ANSI specification). ANSI SQL states that previous
savepoint should be deleted. What is more, our doc confirms this fact
and says that "...it is released before the new savepoint is set."
Unfortunately, it's not true - currently old savepoint remains in the
list. For instance:

SAVEPOINT t;
SAVEPOINT t;
RELEASE SAVEPOINT t;
RELEASE SAVEPOINT t; -- no error is raised

Let's fix this and remove old savepoint from the list.

This allows us to completely remove SQL specific struct Savepoint and
use instead original struct txn_savepoint.

This procedure is processed in several steps. Firstly, we add name
to struct txn_savepoint since we should be capable of operating on named
savepoints (which in turn is SQL feature). Still, anonymous (in the sense
of name absence) savepoints are also valid. Then, we add list (as
implementation of stailq) of savepoints to struct txn: it allows us to
find savepoint by its name. Finally, we patch rollback to/release
savepoint routines: for rollback tail of the list containing savepoints
is cut (but subject of rollback routine remains in the list); for
release routine we cut tail including node being released.

We are going to merge struct psql_txn with struct txn as a part of SQL
integration into NoSQL, so let's move counter of deferred foreign key
violations directly to struct txn.

Closes #4413

@TarantoolBot document
Title: update operations on decimal fields.

tuple:update and space:update now support deicmal operands for
arithmetic operations ('+' and '-'). The syntax is as usual:
```
d = box.tuple.new(decimal.new('1'))
---
...
d:update{{'+', 1, decimal.new('0.5')}}
---
- [1.5]
...
```

Insertion ('!') and assignment ('=') are also supported:
```
a = decimal.new('1')
---
...
b = decimal.new('1e10')
---
...
c = decimal.new('1e-10')
---
...
d = box.tuple.new{5, a, 6, b, 7, c, "string"}
---
...
d
---
- [5, 1, 6, 10000000000, 7, 0.0000000001, 'string']
...

d:update{{'!', 3, dec.new('1234.5678')}}
---
- [5, 1, 1234.5678, 6, 10000000000, 7, 0.0000000001, 'string']
...
d:update{{'=', -1, dec.new('0.12345678910111213')}}
---
- [5, 1, 6, 10000000000, 7, 0.0000000001, 0.12345678910111213]

When performing an arithmetic operation ('+', '-'), where either the
updated field or the operand is decimal, the result will be decimal.

When both the updated field and the operand are decimal, the result
will, of course, be decimal.

...
```

Closes #4333

@TarantoolBot document
Title: Document decimal field type.

Decimals may now be stored in spaces. A corresponding field type is
introduced: 'decimal'. Decimal values are also allowed in 'scalar',
'any' and 'number' fields.

'decimal' field type is appropriate for both memtx HASH and TREE
indices, as well as for vinyl TREE index.

```
To create an index 'pk' over a decimal field, say
```
tarantool> box.space.test:create_index('pk', {parts={1, 'decimal'}})
---
- unique: true
  parts:
  - type: decimal
    is_nullable: false
    fieldno: 1
  id: 0
  space_id: 512
  type: TREE
  name: pk
...
```
Now you can insert some decimal values:
```
tarantool> for i = 1,10 do
         > box.space.test:insert{decimal.new((i-5)/10)}
         > end
---
...
```
tarantool> box.space.test:select{}
---
- - [-0.4]
  - [-0.3]
  - [-0.2]
  - [-0.1]
  - [0]
  - [0.1]
  - [0.2]
  - [0.3]
  - [0.4]
  - [0.5]
...
```
Decimals may alse be inserted into `scalar` and `number` fields. In this
case all the number values are sorted correctly:
```
tarantool> box.schema.space.create('test')
tarantool> box.space.test:create_index('pk', {parts={1, 'number'}})
tarantool> box.space.test:insert{-1.0001, 'number'}
---
- [-1.0001, 'number']
...

tarantool> box.space.test:insert{decimal.new(-1.00001), 'decimal'}
---
- [-1.00001, 'decimal']
...

tarantool> box.space.test:insert{-1, 'number'}
---
- [-1, 'number']
...

tarantool> box.space.test:insert{decimal.new(-0.999), 'decimal'}
---
- [-0.999, 'decimal']
...

tarantool> box.space.test:insert{-0.998, 'number'}
---
- [-0.998, 'number']
...

tarantool> box.space.test:insert{-0.9, 'number'}
---
- [-0.9, 'number']
...

tarantool> box.space.test:insert{-0.95, 'number'}
---
- [-0.95, 'number']
...

tarantool> box.space.test:insert{decimal.new(-0.92), 'decimal'}
---
- [-0.92, 'decimal']
...

tarantool> box.space.test:insert{decimal.new(-0.971), 'decimal'}
---
- [-0.971, 'decimal']
...

tarantool> box.space.test:select{}
---
- - [-1.0001, 'number']
  - [-1.00001, 'decimal']
  - [-1, 'number']
  - [-0.999, 'decimal']
  - [-0.998, 'number']
  - [-0.971, 'decimal']
  - [-0.95, 'number']
  - [-0.92, 'decimal']
  - [-0.9, 'number']
...

```
Uniqueness is also preserved between decimals and other number types:
```
tarantool> box.space.test:insert{-0.92}
---
- error: Duplicate key exists in unique index 'pk' in space 'test'
...

tarantool> box.space.test:insert{decimal.new(-0.9)}
---
- error: Duplicate key exists in unique index 'pk' in space 'test'
...

```

You can also set decimal fields in space format:
```
tarantool> _ = box.schema.space.create('test')
---
...

tarantool> _ = box.space.test:create_index('pk')
---
...

tarantool> box.space.test:format{{name='id', type='unsigned'}, {name='balance', type='decimal'}}
---
...

tarantool> box.space.test:insert{1}
---
- error: Tuple field 2 required by space format is missing
...

tarantool> box.space.test:insert{1, 'string'}
---
- error: 'Tuple field 2 type does not match one required by operation: expected decimal'
...

tarantool> box.space.test:insert{1, 1.2345}
---
- error: 'Tuple field 2 type does not match one required by operation: expected decimal'
...

tarantool> box.space.test:insert{1, decimal.new('1337.420')}
---
- [1, 1337.420]
...

```

Update decNumber library, add methods to convert decimals to uint64_t
and int64_t, add unit tests.
Also replace decimal_round() function with decimal_round_with_mode() to
allow setting rounding mode.
We need to round with mode DEC_ROUND_DOWN in to_int64 conversions in
order to be consistent with double to int conversions.
It will be needed to compute hints for decimal fields.

Prerequisite #4333