Tarantool development patches archive
 help / color / mirror / Atom feed
* [Tarantool-patches] [PATCH luajit] test: disabled flaky test for OSX/ARM64
@ 2026-09-28  6:42 Evgeniy Temirgaleev via Tarantool-patches
  2026-09-28 12:45 ` Sergey Bronnikov via Tarantool-patches
  0 siblings, 1 reply; 5+ messages in thread
From: Evgeniy Temirgaleev via Tarantool-patches @ 2026-09-28  6:42 UTC (permalink / raw)
  To: Sergey Kaplun, Sergey Bronnikov; +Cc: tarantool-patches

From: Temir Galeev <temir.galeev@bk.ru>

The test `fix-mips64-spare-side-exit-patching.test.lua`
is flaky on OSX and ARM64.

The problem is in the limited range for mcode chunk's
addresses on ARM64 and in the OSX strategy for `mmap`
to ignore user's hint [1].

1. https://github.com/tarantool/tarantool/issues/13261

Branch: https://github.com/tarantool/luajit/tree/tmr_g/gh-13621-flaky-fix-mips64-se

---
 .../fix-mips64-spare-side-exit-patching.test.lua                | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/test/tarantool-tests/fix-mips64-spare-side-exit-patching.test.lua b/test/tarantool-tests/fix-mips64-spare-side-exit-patching.test.lua
index dd9f2073..0a9e230f 100644
--- a/test/tarantool-tests/fix-mips64-spare-side-exit-patching.test.lua
+++ b/test/tarantool-tests/fix-mips64-spare-side-exit-patching.test.lua
@@ -4,6 +4,8 @@ local test = tap.test('fix-mips64-spare-side-exit-patching'):skipcond({
   ['Disabled on *BSD due to #4819'] = jit.os == 'BSD',
   -- We need to fix the MIPS behaviour first.
   ['Disabled for MIPS architectures'] = jit.arch:match('mips'),
+  ['Disabled on *OSX/ARM64 due to #13261'] = jit.os == 'OSX' and
+                                             jit.arch:match('arm64'),
 })
 
 local generators = require('utils').jit.generators
-- 
2.49.0


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [Tarantool-patches] [PATCH luajit] test: disabled flaky test for OSX/ARM64
  2026-09-28  6:42 [Tarantool-patches] [PATCH luajit] test: disabled flaky test for OSX/ARM64 Evgeniy Temirgaleev via Tarantool-patches
@ 2026-09-28 12:45 ` Sergey Bronnikov via Tarantool-patches
  2026-09-29 10:31   ` Evgeniy Temirgaleev via Tarantool-patches
  0 siblings, 1 reply; 5+ messages in thread
From: Sergey Bronnikov via Tarantool-patches @ 2026-09-28 12:45 UTC (permalink / raw)
  To: Evgeniy Temirgaleev, Sergey Kaplun; +Cc: tarantool-patches

Hello, Evgeniy!

Thanks for the patch! LGTM with minor comments below.

Sergey

On 9/28/26 09:42, Evgeniy Temirgaleev wrote:
> From: Temir Galeev <temir.galeev@bk.ru>
>
> The test `fix-mips64-spare-side-exit-patching.test.lua`
> is flaky on OSX and ARM64.
>
> The problem is in the limited range for mcode chunk's
> addresses on ARM64 and in the OSX strategy for `mmap`
> to ignore user's hint [1].
The commit message should contain a trailer "Closes 
tarantool/tarantool#13261", right?
>
> 1. https://github.com/tarantool/tarantool/issues/13261
>
> Branch: https://github.com/tarantool/luajit/tree/tmr_g/gh-13621-flaky-fix-mips64-se
>
> ---
>   .../fix-mips64-spare-side-exit-patching.test.lua                | 2 ++
>   1 file changed, 2 insertions(+)
>
> diff --git a/test/tarantool-tests/fix-mips64-spare-side-exit-patching.test.lua b/test/tarantool-tests/fix-mips64-spare-side-exit-patching.test.lua
> index dd9f2073..0a9e230f 100644
> --- a/test/tarantool-tests/fix-mips64-spare-side-exit-patching.test.lua
> +++ b/test/tarantool-tests/fix-mips64-spare-side-exit-patching.test.lua
> @@ -4,6 +4,8 @@ local test = tap.test('fix-mips64-spare-side-exit-patching'):skipcond({
>     ['Disabled on *BSD due to #4819'] = jit.os == 'BSD',
>     -- We need to fix the MIPS behaviour first.
>     ['Disabled for MIPS architectures'] = jit.arch:match('mips'),
> +  ['Disabled on *OSX/ARM64 due to #13261'] = jit.os == 'OSX' and

s/*OSX/macOS/

Feel free to ignore.

OS X is the old official name (used from 2012 to 2016). It is best 
avoided now,

except when referring to specific historical versions (e.g., OS X El 
Capitan).

macOS is the current official name. It is generally not abbreviated further.

osx / OSX is an unofficial, colloquial abbreviation. It appears in tags, 
filenames,

and casual conversation. While technically incorrect, it is widely 
recognized.


> +                                             jit.arch:match('arm64'),
>   })
>   
>   local generators = require('utils').jit.generators

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [Tarantool-patches] [PATCH luajit] test: disabled flaky test for OSX/ARM64
  2026-09-28 12:45 ` Sergey Bronnikov via Tarantool-patches
@ 2026-09-29 10:31   ` Evgeniy Temirgaleev via Tarantool-patches
  2026-09-29 11:22     ` Sergey Bronnikov via Tarantool-patches
  0 siblings, 1 reply; 5+ messages in thread
From: Evgeniy Temirgaleev via Tarantool-patches @ 2026-09-29 10:31 UTC (permalink / raw)
  To: Sergey Bronnikov; +Cc: tarantool-patches

Hi, Sergey! Thanks for review. Please, see my answers below.

The branch is force pushed with new commit message.

28.09.2026 15:45, Sergey Bronnikov пишет:
> Hello, Evgeniy!
> 
> Thanks for the patch! LGTM with minor comments below.
> 
> Sergey
> 
> On 9/28/26 09:42, Evgeniy Temirgaleev wrote:
>> From: Temir Galeev <temir.galeev@bk.ru>
>>
>> The test `fix-mips64-spare-side-exit-patching.test.lua`
>> is flaky on OSX and ARM64.
>>
>> The problem is in the limited range for mcode chunk's
>> addresses on ARM64 and in the OSX strategy for `mmap`
>> to ignore user's hint [1].
> The commit message should contain a trailer "Closes tarantool/ 
> tarantool#13261", right?

None. The ticket is more general. It's about mcode allocation problem
due to address range limitation.

Updated the commit description.

>>
>> 1. https://github.com/tarantool/tarantool/issues/13261
>>
>> Branch: https://github.com/tarantool/luajit/tree/tmr_g/gh-13621-flaky- 
>> fix-mips64-se
>>
>> ---
>>   .../fix-mips64-spare-side-exit-patching.test.lua                | 2 ++
>>   1 file changed, 2 insertions(+)
>>
>> diff --git a/test/tarantool-tests/fix-mips64-spare-side-exit- 
>> patching.test.lua b/test/tarantool-tests/fix-mips64-spare-side-exit- 
>> patching.test.lua
>> index dd9f2073..0a9e230f 100644
>> --- a/test/tarantool-tests/fix-mips64-spare-side-exit-patching.test.lua
>> +++ b/test/tarantool-tests/fix-mips64-spare-side-exit-patching.test.lua
>> @@ -4,6 +4,8 @@ local test = tap.test('fix-mips64-spare-side-exit- 
>> patching'):skipcond({
>>     ['Disabled on *BSD due to #4819'] = jit.os == 'BSD',
>>     -- We need to fix the MIPS behaviour first.
>>     ['Disabled for MIPS architectures'] = jit.arch:match('mips'),
>> +  ['Disabled on *OSX/ARM64 due to #13261'] = jit.os == 'OSX' and
> 
> s/*OSX/macOS/

Fixed, thanks!

> 
> Feel free to ignore.
> 
> OS X is the old official name (used from 2012 to 2016). It is best 
> avoided now,
> 
> except when referring to specific historical versions (e.g., OS X El 
> Capitan).
> 
> macOS is the current official name. It is generally not abbreviated 
> further.
> 
> osx / OSX is an unofficial, colloquial abbreviation. It appears in tags, 
> filenames,
> 
> and casual conversation. While technically incorrect, it is widely 
> recognized.
> 
> 
>> +                                             jit.arch:match('arm64'),
>>   })
>>   local generators = require('utils').jit.generators

-- 
Updated commit message:

test: disabled flaky test for macOS/ARM64

The test `fix-mips64-spare-side-exit-patching.test.lua`
is flaky on macOS and ARM64.

The problem is in the limited range for mcode chunk's
addresses on ARM64 and in the macOS strategy for `mmap`
to ignore user's hint. See [1] for details.

After the solution for the address range limitation problem
will be applied, the test must be reenabled.

1. https://github.com/tarantool/tarantool/issues/13261

Part of #13261

--
Best regards,
Evgeniy Temirgaleev


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [Tarantool-patches] [PATCH luajit] test: disabled flaky test for OSX/ARM64
  2026-09-29 10:31   ` Evgeniy Temirgaleev via Tarantool-patches
@ 2026-09-29 11:22     ` Sergey Bronnikov via Tarantool-patches
  2026-09-29 13:10       ` Evgeniy Temirgaleev via Tarantool-patches
  0 siblings, 1 reply; 5+ messages in thread
From: Sergey Bronnikov via Tarantool-patches @ 2026-09-29 11:22 UTC (permalink / raw)
  To: Evgeniy Temirgaleev; +Cc: tarantool-patches

Hi, Evgeniy,


On 9/29/26 13:31, Evgeniy Temirgaleev wrote:
> Hi, Sergey! Thanks for review. Please, see my answers below.
>
> The branch is force pushed with new commit message.
>
> 28.09.2026 15:45, Sergey Bronnikov пишет:
>> Hello, Evgeniy!
>>
>> Thanks for the patch! LGTM with minor comments below.
>>
>> Sergey
>>
>> On 9/28/26 09:42, Evgeniy Temirgaleev wrote:
>>> From: Temir Galeev <temir.galeev@bk.ru>
>>>
>>> The test `fix-mips64-spare-side-exit-patching.test.lua`
>>> is flaky on OSX and ARM64.
>>>
>>> The problem is in the limited range for mcode chunk's
>>> addresses on ARM64 and in the OSX strategy for `mmap`
>>> to ignore user's hint [1].
>> The commit message should contain a trailer "Closes tarantool/ 
>> tarantool#13261", right?
>
> None. The ticket is more general. It's about mcode allocation problem
> due to address range limitation.
Then you can add "Relates to tarantool/ tarantool#13261".
>
> Updated the commit description.
>
>
<snipped>

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [Tarantool-patches] [PATCH luajit] test: disabled flaky test for OSX/ARM64
  2026-09-29 11:22     ` Sergey Bronnikov via Tarantool-patches
@ 2026-09-29 13:10       ` Evgeniy Temirgaleev via Tarantool-patches
  0 siblings, 0 replies; 5+ messages in thread
From: Evgeniy Temirgaleev via Tarantool-patches @ 2026-09-29 13:10 UTC (permalink / raw)
  To: Sergey Bronnikov; +Cc: tarantool-patches

Sergey, thanks for review!

The commit message is fixed:

-Part of #13261
+Relates to tarantool/tarantool#13261

The branch is force pushed.

29.09.2026 14:22, Sergey Bronnikov пишет:
> Hi, Evgeniy,
> 
> 
> On 9/29/26 13:31, Evgeniy Temirgaleev wrote:
>> Hi, Sergey! Thanks for review. Please, see my answers below.
>>
>> The branch is force pushed with new commit message.
>>
>> 28.09.2026 15:45, Sergey Bronnikov пишет:
>>> Hello, Evgeniy!
>>>
>>> Thanks for the patch! LGTM with minor comments below.
>>>
>>> Sergey
>>>
>>> On 9/28/26 09:42, Evgeniy Temirgaleev wrote:
>>>> From: Temir Galeev <temir.galeev@bk.ru>
>>>>
>>>> The test `fix-mips64-spare-side-exit-patching.test.lua`
>>>> is flaky on OSX and ARM64.
>>>>
>>>> The problem is in the limited range for mcode chunk's
>>>> addresses on ARM64 and in the OSX strategy for `mmap`
>>>> to ignore user's hint [1].
>>> The commit message should contain a trailer "Closes tarantool/ 
>>> tarantool#13261", right?
>>
>> None. The ticket is more general. It's about mcode allocation problem
>> due to address range limitation.
> Then you can add "Relates to tarantool/ tarantool#13261".
>>
>> Updated the commit description.
>>
>>
> <snipped>

-- 
--
Best regards,
Evgeniy Temirgaleev


^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2026-09-29 13:10 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-09-28  6:42 [Tarantool-patches] [PATCH luajit] test: disabled flaky test for OSX/ARM64 Evgeniy Temirgaleev via Tarantool-patches
2026-09-28 12:45 ` Sergey Bronnikov via Tarantool-patches
2026-09-29 10:31   ` Evgeniy Temirgaleev via Tarantool-patches
2026-09-29 11:22     ` Sergey Bronnikov via Tarantool-patches
2026-09-29 13:10       ` Evgeniy Temirgaleev via Tarantool-patches

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox