Reminder: The Master branch will be closing for the 1.77.0 release on WEDNESDAY 4-August

newer
...

older
Re: [boost] Reminder: The Master...

Marshall Clow

2 Aug 2021 2 Aug '21

3 a.m.

Release calendar here: https://www.boost.org/development/index.html — Marshall

Show replies by date

Robert Ramey

2 Aug 2 Aug

4:41 a.m.

On 8/1/21 8:00 PM, Marshall Clow via Boost wrote:

...

Release calendar here: https://www.boost.org/development/index.html

— Marshall

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Hmmmm - shouldn't https://www.boost.org/development/tests/master/developer/serialization.html have a date later than last May before we do a release? Robert Ramey

Tom Kent

11:34 a.m.

On Sun, Aug 1, 2021 at 11:41 PM Robert Ramey via Boost < boost@lists.boost.org> wrote:

...

On 8/1/21 8:00 PM, Marshall Clow via Boost wrote:

...
Release calendar here: https://www.boost.org/development/index.html

— Marshall

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Hmmmm - shouldn't https://www.boost.org/development/tests/master/developer/serialization.html have a date later than last May before we do a release?

Robert Ramey

Yes. I just checked the teeks99-* jobs and everything is still running. The linux ones are all cycling and happy. The windows one is crashing very quickly, I'll debug that here in a minute and see if it is my side or something coming from the repo. However, the test matrix doesn't seem to have updated in months!?!? Was this perhaps running on travis-ci and didn't get moved when they stopped running OSS jobs? Tom

Niall Douglas

12:32 p.m.

On 02/08/2021 12:34, Tom Kent via Boost wrote:

...

However, the test matrix doesn't seem to have updated in months!?!? Was this perhaps running on travis-ci and didn't get moved when they stopped running OSS jobs?

I reported this to Slack #boost last week. The test matrix stopped updating on 19th May. It's likely a cronjob which isn't firing any more. In any case, fixing it before we do a release would be super useful. Niall

Marshall Clow

8:49 p.m.

On Aug 2, 2021, at 5:32 AM, Niall Douglas via Boost wrote:

...

On 02/08/2021 12:34, Tom Kent via Boost wrote:

...
However, the test matrix doesn't seem to have updated in months!?!? Was this perhaps running on travis-ci and didn't get moved when they stopped running OSS jobs?

I reported this to Slack #boost last week. The test matrix stopped updating on 19th May. It's likely a cronjob which isn't firing any more.

In any case, fixing it before we do a release would be super useful.

My current theory is that this commit to filesystem caused the tools to stop building https://github.com/boostorg/filesystem/commit/88c2a2df8c23e28732bfab81226129... The docker image that the tools get run from doesn’t have “linux/magic.h” More as we discover it. — Marshall

Marshall Clow

10:42 p.m.

On Aug 2, 2021, at 1:49 PM, Marshall Clow wrote:

...

On Aug 2, 2021, at 5:32 AM, Niall Douglas via Boost wrote:

...
On 02/08/2021 12:34, Tom Kent via Boost wrote:

...
However, the test matrix doesn't seem to have updated in months!?!? Was this perhaps running on travis-ci and didn't get moved when they stopped running OSS jobs?

I reported this to Slack #boost last week. The test matrix stopped updating on 19th May. It's likely a cronjob which isn't firing any more.

In any case, fixing it before we do a release would be super useful.

My current theory is that this commit to filesystem caused the tools to stop building https://github.com/boostorg/filesystem/commit/88c2a2df8c23e28732bfab81226129...

The docker image that the tools get run from doesn’t have “linux/magic.h”

Michael Caisse has updated the docker image in question, and we’re starting to see results posted. Thanks to Michael. — Marshall

Дмитрий Архипов

3 Aug 3 Aug

6:05 a.m.

Current release notes for Beta haven't been updated since July 14th and there 3 website PRs have been merged since. Can someone please update the website? вт, 3 авг. 2021 г. в 01:43, Marshall Clow via Boost :

...

On Aug 2, 2021, at 1:49 PM, Marshall Clow wrote:

...
On Aug 2, 2021, at 5:32 AM, Niall Douglas via Boost <

boost@lists.boost.org> wrote:

...
...
On 02/08/2021 12:34, Tom Kent via Boost wrote:

...
However, the test matrix doesn't seem to have updated in months!?!?

Was

...
...
this perhaps running on travis-ci and didn't get moved when they stopped running OSS jobs?

I reported this to Slack #boost last week. The test matrix stopped updating on 19th May. It's likely a cronjob which isn't firing any more.

In any case, fixing it before we do a release would be super useful.

My current theory is that this commit to filesystem caused the tools to stop building

https://github.com/boostorg/filesystem/commit/88c2a2df8c23e28732bfab81226129...

...
The docker image that the tools get run from doesn’t have “linux/magic.h”

Michael Caisse has updated the docker image in question, and we’re starting to see results posted.

Thanks to Michael.

— Marshall

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

John Maddock

2 Aug 2 Aug

4:31 p.m.

With sincere apologies for spotting this so late, Math has a rather grievous bug right at it's core, see: https://github.com/boostorg/math/pull/670. Fortunately the fix is relatively trivial. If possible I'd like to cherry pick this into master before the deadline. CI has all cycled. Thanks, John. -- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

Marshall Clow

4:52 p.m.

On Aug 2, 2021, at 9:31 AM, John Maddock via Boost wrote:

...

With sincere apologies for spotting this so late, Math has a rather grievous bug right at it's core, see: https://github.com/boostorg/math/pull/670.

Fortunately the fix is relatively trivial. If possible I'd like to cherry pick this into master before the deadline. CI has all cycled.

Go ahead. — Marshall

John Maddock

5:36 p.m.

On 02/08/2021 17:52, Marshall Clow via Boost wrote:

...

On Aug 2, 2021, at 9:31 AM, John Maddock via Boost wrote:

...
With sincere apologies for spotting this so late, Math has a rather grievous bug right at it's core, see: https://github.com/boostorg/math/pull/670.

Fortunately the fix is relatively trivial. If possible I'd like to cherry pick this into master before the deadline. CI has all cycled. Go ahead.

Thanks, done. With thanks to some sharp eyes from Peter Dimov, this duplicate definition fix should also probably go in once CI has cycled: https://github.com/boostorg/math/pull/671 Thanks again, John. -- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

John Maddock

3 Aug 3 Aug

7:50 a.m.

...

With thanks to some sharp eyes from Peter Dimov, this duplicate definition fix should also probably go in once CI has cycled: https://github.com/boostorg/math/pull/671

Ping? CI has now cycled clean. Thanks, John. -- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

Marshall Clow

1:44 p.m.

On Aug 3, 2021, at 12:50 AM, John Maddock via Boost wrote:

...

...
With thanks to some sharp eyes from Peter Dimov, this duplicate definition fix should also probably go in once CI has cycled: https://github.com/boostorg/math/pull/671

Ping? CI has now cycled clean.

Thanks, John.

Go ahead. — Marshall

John Maddock

4:43 p.m.

On 03/08/2021 14:44, Marshall Clow via Boost wrote:

...

On Aug 3, 2021, at 12:50 AM, John Maddock via Boost wrote:

...
...
With thanks to some sharp eyes from Peter Dimov, this duplicate definition fix should also probably go in once CI has cycled: https://github.com/boostorg/math/pull/671

Ping? CI has now cycled clean.

Thanks, John.

Go ahead.

Many thanks for your patience on this Marshall, done. Best, John.

...

— Marshall

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

-- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

Emil Dotchevski

12:01 a.m.

Technically not a Beta bug fix but I think it would be nice if these fixes make it in the next release, both reported after the Beta lock: https://github.com/boostorg/qvm/commit/46797f11f71c625420217ef9043d6ee4a12cf... . https://github.com/boostorg/qvm/commit/36ab7dfd000d8100727ceee98df233dc7e5de... (Issue 33). This also fixes GHA errors due to having to install GCC. On Sun, Aug 1, 2021 at 8:01 PM Marshall Clow via Boost < boost@lists.boost.org> wrote:

...

Release calendar here: https://www.boost.org/development/index.html

— Marshall

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Marshall Clow

12:12 a.m.

On Aug 2, 2021, at 5:01 PM, Emil Dotchevski via Boost wrote:

...

Technically not a Beta bug fix but I think it would be nice if these fixes make it in the next release, both reported after the Beta lock:

https://github.com/boostorg/qvm/commit/46797f11f71c625420217ef9043d6ee4a12cf... .

This one looks worrisome to me; that’s a significant change. Let’s get the bots back up, look at the results, and talk tomorrow or Wednesday. [ I’m not against merging it, I’m just concerned ]

...

https://github.com/boostorg/qvm/commit/36ab7dfd000d8100727ceee98df233dc7e5de... (Issue 33).

This one is fine. Go ahead.

...

This also fixes GHA errors due to having to install GCC.

On Sun, Aug 1, 2021 at 8:01 PM Marshall Clow via Boost < boost@lists.boost.org> wrote:

...
Release calendar here: https://www.boost.org/development/index.html

— Marshall

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Emil Dotchevski

1:46 a.m.

On Mon, Aug 2, 2021 at 5:12 PM Marshall Clow via Boost < boost@lists.boost.org> wrote:

...

On Aug 2, 2021, at 5:01 PM, Emil Dotchevski via Boost <

boost@lists.boost.org> wrote:

...

...
Technically not a Beta bug fix but I think it would be nice if these

fixes

...
make it in the next release, both reported after the Beta lock:

https://github.com/boostorg/qvm/commit/46797f11f71c625420217ef9043d6ee4a12cf... .

This one looks worrisome to me; that’s a significant change. Let’s get the bots back up, look at the results, and talk tomorrow or Wednesday.

Perhaps misunderstanding? The bots are up, tests all pass: https://github.com/boostorg/qvm/actions/runs/1054919685. They weren't passing before this commit because github had removed some gcc versions from the base ubuntu image used by GHA, so now they need to be installed manually. To clarify, this change touches many files and looks scary but it is all along these lines: We had: T const mag=sqrt<T>(m2); Now we have: T const mag=sqrt(m2); The sqrt function template is defined in, and invoked from, namespace boost::qvm. The difference is that now we're using ADL, which could bind some other sqrt function (m2 may be of user-defined type). The same is done for other math functions, e.g. sin, cos, etc.

...

[ I’m not against merging it, I’m just concerned ]

Yeah, I'm on the edge for this one too. Seems simple and safe though. :)

...

...
https://github.com/boostorg/qvm/commit/36ab7dfd000d8100727ceee98df233dc7e5de...

...
(Issue 33).

This one is fine. Go ahead.

I'd rather not merge this one if the other one isn't merged in also. Thank you!

Marshall Clow

1:57 a.m.

On Aug 2, 2021, at 6:46 PM, Emil Dotchevski via Boost wrote:

...

On Mon, Aug 2, 2021 at 5:12 PM Marshall Clow via Boost < boost@lists.boost.org> wrote:

...
On Aug 2, 2021, at 5:01 PM, Emil Dotchevski via Boost <

boost@lists.boost.org> wrote:

...
...
Technically not a Beta bug fix but I think it would be nice if these

fixes

...
make it in the next release, both reported after the Beta lock:

https://github.com/boostorg/qvm/commit/46797f11f71c625420217ef9043d6ee4a12cf... .

This one looks worrisome to me; that’s a significant change. Let’s get the bots back up, look at the results, and talk tomorrow or Wednesday.

Perhaps misunderstanding? The bots are up, tests all pass: https://github.com/boostorg/qvm/actions/runs/1054919685.

I was thinking of these bots. https://www.boost.org/development/tests/develop/developer/summary.html

...

They weren't passing before this commit because github had removed some gcc versions from the base ubuntu image used by GHA, so now they need to be installed manually.

To clarify, this change touches many files and looks scary but it is all along these lines:

We had:

T const mag=sqrt<T>(m2);

Now we have:

T const mag=sqrt(m2);

The sqrt function template is defined in, and invoked from, namespace boost::qvm. The difference is that now we're using ADL, which could bind some other sqrt function (m2 may be of user-defined type).

The same is done for other math functions, e.g. sin, cos, etc.

Yeah, I got that. My concern is wholly about finding “other" sqrt functions.

...

...
[ I’m not against merging it, I’m just concerned ]

Yeah, I'm on the edge for this one too. Seems simple and safe though. :)

...
...
https://github.com/boostorg/qvm/commit/36ab7dfd000d8100727ceee98df233dc7e5de...

...
(Issue 33).

This one is fine. Go ahead.

I'd rather not merge this one if the other one isn't merged in also.

That’s your call. — Marshall

Richard Hodges

4 Aug 4 Aug

9:29 a.m.

Boost.Beast release 319 pushed to master. Thank you. On Mon, 2 Aug 2021 at 05:01, Marshall Clow via Boost wrote:

...

Release calendar here: https://www.boost.org/development/index.html

— Marshall

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

1194

Age (days ago)

1196

Last active (days ago)

List overview

Download

17 comments

8 participants

participants (8)

Emil Dotchevski
John Maddock
Marshall Clow
Niall Douglas
Richard Hodges
Robert Ramey
Tom Kent
Дмитрий Архипов