Hi Igor, thanks for your comments, I’ve used all of them in the code.

Пятница, 28 февраля 2020, 21:42 +03:00 от Igor Munkin <imun@tarantool.org>:

Sasha,

Thanks for the patch. I left some comments below, please consider them.

On 27.02.20, Alexander V. Tikhonov wrote:
> Cleaned up the tests code:
> - added local definitions for the variables
> - added os.exit() routine at the tests finish
> - changed added information messages to test:ok() calls

I propose to reword the commit message the following way:
| test: cleanup tests code
|
| Cleaned up the tests code according to the Lua style guide:
| - made scoped variables local
| - added os.exit call at the end of the test chunk
| - adjusted the messages in test:ok calls

General comment: please adjust the test names regarding the changes I
proposed in the second patch.

>
> Part of #4655
> ---
> test/fix_string_find_recording.test.lua | 6 +++---
> test/fold_bug_LuaJIT_505.test.lua | 7 ++++---
> test/fold_bug_LuaJIT_524.test.lua | 4 ++--
> test/gh.test.lua | 8 ++++----
> test/pairsmm_tarantool_4560.test.lua | 8 +++++---
> test/table_chain_bug_LuaJIT_494.test.lua | 23 ++++++++++++-----------
> test/unsink_64_kptr.test.lua | 7 ++++---
> 7 files changed, 34 insertions(+), 29 deletions(-)
>

<snipped>

> diff --git a/test/fold_bug_LuaJIT_505.test.lua b/test/fold_bug_LuaJIT_505.test.lua
> index 2fee069..3b7d7df 100755
> --- a/test/fold_bug_LuaJIT_505.test.lua
> +++ b/test/fold_bug_LuaJIT_505.test.lua
> @@ -1,8 +1,8 @@
> #!/usr/bin/env tarantool
>
> -tap = require('tap')
> +local tap = require('tap')
>
> -test = tap.test("505")
> +local test = tap.test("lj-505-fold-icorrect-behavior")
> test:plan(1)
>
> -- Test file to demonstrate Lua fold machinery icorrect behavior, details:
> @@ -16,5 +16,6 @@ for _ = 1, 20 do
> local pos_b = string.find(value2, "b", 2, true)
> assert(pos_b == 2, "FAIL: position of 'b' is " .. pos_b)
> end
> +test:ok(true, "LuaJIT/LuaJIT gh-505 assert not occured")

Message is absolutely non-informative. I propose to reword it:
| "string.find offset aritmetics wasn't broken while recording"

>
> -test:ok("PASS")
> +os.exit(test:check() and 0 or 1)

<snipped>

> diff --git a/test/table_chain_bug_LuaJIT_494.test.lua b/test/table_chain_bug_LuaJIT_494.test.lua
> index 06c0f0d..3007360 100755
> --- a/test/table_chain_bug_LuaJIT_494.test.lua
> +++ b/test/table_chain_bug_LuaJIT_494.test.lua
> @@ -1,8 +1,8 @@
> #!/usr/bin/env tarantool
>
> -tap = require('tap')
> +local tap = require('tap')
>
> -test = tap.test("494")
> +local test = tap.test("lj-494-table-chain-infinite-loop")
> test:plan(1)
>
> -- Test file to demonstrate Lua table hash chain bugs discussed in
> @@ -11,19 +11,19 @@ test:plan(1)
> -- https://gist.github.com/corsix/1fc9b13a2dd5f3659417b62dd54d4500
>
> --- Plumbing
> -ffi = require"ffi"
> -ffi.cdef"char* strstr(const char*, const char*)"
> -strstr = ffi.C.strstr
> -cast = ffi.cast
> -str_hash_offset = cast("uint32_t*", strstr("*", ""))[-2] == 1 and 3 or 2
> +local ffi = require("ffi")
> +ffi.cdef("char* strstr(const char*, const char*)")
> +local strstr = ffi.C.strstr
> +local cast = ffi.cast
> +local str_hash_offset = cast("uint32_t*", strstr("*", ""))[-2] == 1 and 3 or 2
> function str_hash(s)
> return cast("uint32_t*", strstr(s, "")) - str_hash_offset
> end
> -table_new = require"table.new"
> +local table_new = require("table.new")
>
> --- Prepare some objects
> -victims = {}
> -orig_hash = {}
> +local victims = {}
> +local orig_hash = {}
> for c in ("abcdef"):gmatch"." do
> v = c .. "{09add58a-13a4-44e0-a52c-d44d0f9b2b95}"
> victims[c] = v
> @@ -174,5 +174,6 @@ collectgarbage()
> for c, v in pairs(victims) do
> str_hash(v)[0] = orig_hash[c]
> end
> +test:ok(true, "LuaJIT/LuaJIT gh-494 successfully checked commit a7dc9e8d23315217c9fe0029bc8ae12c03306b33")

Message is absolutely non-informative. I propose to reword it:
| "table keys collisions are resolved properly (no assertions failed)"

>
> -test:ok("PASS")
> +os.exit(test:check() and 0 or 1)
> diff --git a/test/unsink_64_kptr.test.lua b/test/unsink_64_kptr.test.lua
> index 8995763..a618fd2 100755
> --- a/test/unsink_64_kptr.test.lua
> +++ b/test/unsink_64_kptr.test.lua
> @@ -1,8 +1,8 @@
> #!/usr/bin/env tarantool
>
> -tap = require('tap')
> +local tap = require('tap')
>
> -test = tap.test("232")
> +local test = tap.test("or-232-unsink-64-kptr")
> test:plan(1)
>
> --- From: Thibault Charbonnier <thibaultcha@me.com>
> @@ -40,5 +40,6 @@ end
> for i = 1, 1000 do
> fn(i)
> end
> +test:ok(true, "openresty-lua/resty-core gh-232 allowed its address to be inlined")

Here is a mess with a message, it totally doesn't describe the check
being made. I propose the following comment:
| "allocation is unsunk at the trace exit (no platform failures)"