EMBA (Emacs Gitlab CI build server) is now available

EMBA (Emacs Gitlab CI build server) is now available

[Ted Zlatanov]

You can access EMBA at https://emba.gnu.org

It will build Emacs commits on some pre-defined branches like master and
emacs-26.

If you need a signup invite, please contact me (for now).

Thank you
Ted

Re: EMBA (Emacs Gitlab CI build server) is now available

[Michael Albinus]

Ted Zlatanov <tzz@lifelogs.com> writes:

Hi Ted,

> You can access EMBA at https://emba.gnu.org
>
> It will build Emacs commits on some pre-defined branches like master and
> emacs-26.
>
> If you need a signup invite, please contact me (for now).

Looks like I need an account. But why is it password based? I'd prefer a
public key, or a certificate.

I also believe, that account-free access shall be possible to check test
run results.

> Thank you
> Ted

Best regards, Michael.

Re: EMBA (Emacs Gitlab CI build server) is now available

[Ted Zlatanov]

On Mon, 31 Dec 2018 18:54:03 +0100 Michael Albinus <michael.albinus@gmx.de> wrote: 

>> You can access EMBA at https://emba.gnu.org
>> 
>> It will build Emacs commits on some pre-defined branches like master and
>> emacs-26.
>> 
>> If you need a signup invite, please contact me (for now).

MA> Looks like I need an account. But why is it password based? I'd prefer a
MA> public key, or a certificate.

Gitlab, the software EMBA uses, supports passwords and 2FA. I don't know
if it supports others.

MA> I also believe, that account-free access shall be possible to check test
MA> run results.

I agree, but I don't think the server is ready for full public access.
I've allowed your signup, at least for now.

Ted

Re: EMBA (Emacs Gitlab CI build server) is now available

[Michael Albinus]

Ted Zlatanov <tzz@lifelogs.com> writes:

Hi Ted,

>>> If you need a signup invite, please contact me (for now).
>
> MA> Looks like I need an account. But why is it password based? I'd prefer a
> MA> public key, or a certificate.
>
> Gitlab, the software EMBA uses, supports passwords and 2FA. I don't know
> if it supports others.

According to <https://docs.gitlab.com/ee/topics/authentication/index.html>,
OpenID is supported as well. This might be an alternative, I own an
OpenID account.

> Ted

Best regards, Michael.

Re: EMBA (Emacs Gitlab CI build server) is now available

[Ted Zlatanov]

On Tue, 01 Jan 2019 13:08:23 +0100 Michael Albinus <michael.albinus@gmx.de> wrote: 

MA> According to <https://docs.gitlab.com/ee/topics/authentication/index.html>,
MA> OpenID is supported as well. This might be an alternative, I own an
MA> OpenID account.

Unfortunately we're running the free Community Edition, we're limited on
many pay-only features.

Ted

Re: EMBA (Emacs Gitlab CI build server) is now available

[Michael Albinus]

Ted Zlatanov <tzz@lifelogs.com> writes:

> MA> According to <https://docs.gitlab.com/ee/topics/authentication/index.html>,
> MA> OpenID is supported as well. This might be an alternative, I own an
> MA> OpenID account.
>
> Unfortunately we're running the free Community Edition, we're limited on
> many pay-only features.

In User Settings -> Applications, I could enable GitLab as OpenID
provider. Nice.

The other way, using an OpenID provider for login into GitLab, doesn't
seem to bee offered. At least I haven't seen.

> Ted

Best regards, Michael.

Re: EMBA (Emacs Gitlab CI build server) is now available

[Justin Joseph Kaipada]

[-- Attachment #1: Type: text/plain, Size: 731 bytes --]

Is it just me or https://emba.gnu.org/ is down already ?

On Tue, Jan 1, 2019 at 11:51 AM Michael Albinus <michael.albinus@gmx.de>
wrote:

> Ted Zlatanov <tzz@lifelogs.com> writes:
>
> > MA> According to <
> https://docs.gitlab.com/ee/topics/authentication/index.html>,
> > MA> OpenID is supported as well. This might be an alternative, I own an
> > MA> OpenID account.
> >
> > Unfortunately we're running the free Community Edition, we're limited on
> > many pay-only features.
>
> In User Settings -> Applications, I could enable GitLab as OpenID
> provider. Nice.
>
> The other way, using an OpenID provider for login into GitLab, doesn't
> seem to bee offered. At least I haven't seen.
>
> > Ted
>
> Best regards, Michael.
>
>

[-- Attachment #2: Type: text/html, Size: 1332 bytes --]

Re: EMBA (Emacs Gitlab CI build server) is now available

[Michael Albinus]

Justin Joseph Kaipada <justinjoseph0007@gmail.com> writes:

> Is it just me or https://emba.gnu.org/ is down already ?

Works for me. What does "down" mean? No page returned?

Best regards, Michael.

Re: EMBA (Emacs Gitlab CI build server) is now available

[Michael Albinus]

Ted Zlatanov <tzz@lifelogs.com> writes:

Hi Ted,

> You can access EMBA at https://emba.gnu.org
>
> It will build Emacs commits on some pre-defined branches like master and
> emacs-26.
>
> If you need a signup invite, please contact me (for now).

Wouldn't it be possible to move the emacs/emacs project visibility to
"public"? By this, people w/o emba credentials could look on the
results, and files like <https://emba.gnu.org/emacs/emacs/-/jobs/446/raw>
could be accessed easily from within Emacs.

> Thank you
> Ted

Best regards, Michael.

Re: EMBA (Emacs Gitlab CI build server) is now available

[Toon Claes]

[-- Attachment #1: Type: text/plain, Size: 420 bytes --]

Michael Albinus <michael.albinus@gmx.de> writes:

> Wouldn't it be possible to move the emacs/emacs project visibility to
> "public"? By this, people w/o emba credentials could look on the
> results, and files like <https://emba.gnu.org/emacs/emacs/-/jobs/446/raw>
> could be accessed easily from within Emacs.

We plan to do so in the near future. There are some small things we want
to fix before we do that.

-- Toon

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Toon Claes]

[-- Attachment #1: Type: text/plain, Size: 608 bytes --]

We have made emba.gnu.org *public* available. So it's no longer needed
to log in to see the project or CI results.

The failed builds are mailed to emacs-buildstatus, so you can follow that
mailing list if you want to get notified about failures.

GitLab configured so it sends plain-text mails and the code is patched so
the mail notifications include a direct link to raw log file (no need to
run any javascript, set up sessions/cookies/etc)

If you want an account on the GitLab server, get in touch with Ted,
Michael, or me to set it up.

Many thanks to Ted and Michael for the collaboration!


-- Toon


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Robert Pluim]

Toon Claes <toon@iotcl.com> writes:

> We have made emba.gnu.org *public* available. So it's no longer needed
> to log in to see the project or CI results.
>
> The failed builds are mailed to emacs-buildstatus, so you can follow that
> mailing list if you want to get notified about failures.
>
> GitLab configured so it sends plain-text mails and the code is patched so
> the mail notifications include a direct link to raw log file (no need to
> run any javascript, set up sessions/cookies/etc)
>
> If you want an account on the GitLab server, get in touch with Ted,
> Michael, or me to set it up.
>
> Many thanks to Ted and Michael for the collaboration!

It looks pretty cool, however if I visit
https://emba.gnu.org/emacs/emacs/activity , it looks like we should
blame Ted for every breakage. Would it be possible to display the
commit authors there?

Robert

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Michael Albinus]

Robert Pluim <rpluim@gmail.com> writes:

> It looks pretty cool, however if I visit
> https://emba.gnu.org/emacs/emacs/activity , it looks like we should
> blame Ted for every breakage.

It's not a bug, it's a feature :-) I really appreciate this, given that
I have caused several breakages these days ...

> Would it be possible to display the commit authors there?

Don't know. Likely not, because it reflects the git clone on the emba
server itself, which is triggered with Ted's credentials.

However, if you follow the git commit ref links, you will see the
committer. And the mails to the buildstatus ML contain the real
committer as well.

> Robert

Best regards, Michael.

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Toon Claes]

[-- Attachment #1: Type: text/plain, Size: 613 bytes --]

Michael Albinus <michael.albinus@gmx.de> writes:

> Robert Pluim <rpluim@gmail.com> writes:
>
>> Would it be possible to display the commit authors there?
>
> Don't know. Likely not, because it reflects the git clone on the emba
> server itself, which is triggered with Ted's credentials.

Yes. I suggest we create a "bot" user that pushes the changes. This
should reduce the confusion.

> However, if you follow the git commit ref links, you will see the
> committer. And the mails to the buildstatus ML contain the real
> committer as well.

Yes. When we've set up a bot, this should be more obvious.


-- Toon

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Michael Albinus]

Toon Claes <toon@iotcl.com> writes:

>> Don't know. Likely not, because it reflects the git clone on the emba
>> server itself, which is triggered with Ted's credentials.
>
> Yes. I suggest we create a "bot" user that pushes the changes. This
> should reduce the confusion.

Good idea. This shows also clearly the difference, when somebody has
started a pipeline manually.

> -- Toon

Best regards, Michael.

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Ted Zlatanov]

On Wed, 30 Jan 2019 13:13:16 +0100 Michael Albinus <michael.albinus@gmx.de> wrote: 

MA> Toon Claes <toon@iotcl.com> writes:
>>> Don't know. Likely not, because it reflects the git clone on the emba
>>> server itself, which is triggered with Ted's credentials.
>> 
>> Yes. I suggest we create a "bot" user that pushes the changes. This
>> should reduce the confusion.

MA> Good idea. This shows also clearly the difference, when somebody has
MA> started a pipeline manually.

I think it's correct to show the pushing user, because the push is what
triggered the pipeline and could have contained multiple commits by
multiple authors. So a bot user makes a lot of sense for our current
setup, pure mirroring.

...and I see Toon already implemented it :) Thank you Toon!!

Ted

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Robert Pluim]

[-- Attachment #1.1: Type: text/plain, Size: 472 bytes --]

Toon Claes <toon@iotcl.com> writes:

> We have made emba.gnu.org *public* available. So it's no longer needed
> to log in to see the project or CI results.
>
> The failed builds are mailed to emacs-buildstatus, so you can follow that
> mailing list if you want to get notified about failures.

I should join that list.

Given the recent build failures, Iʼm guessing that server doesnʼt have
IPv6 enabled at all? In which case I plan to push the following:


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1.2: 0001-Fix-network-stream-tests.patch --]
[-- Type: text/x-patch, Size: 1640 bytes --]

From 450f455620a926064f5322af64c3c7b439fd0fae Mon Sep 17 00:00:00 2001
From: Robert Pluim <rpluim@gmail.com>
Date: Tue, 5 Feb 2019 13:47:27 +0100
Subject: [PATCH] Fix network stream tests
To: emacs-devel@gnu.org

* test/lisp/net/network-stream-tests.el
(make-ipv6-tcp-server-with-unspecified-port): Skip if IPv6 is not available.
(make-ipv6-tcp-server-with-specified-port): Likewise.
(echo-server-with-local-ipv6): Likewise.
---
 test/lisp/net/network-stream-tests.el | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/test/lisp/net/network-stream-tests.el b/test/lisp/net/network-stream-tests.el
index 6151c3064c..b8b0a8fd82 100644
--- a/test/lisp/net/network-stream-tests.el
+++ b/test/lisp/net/network-stream-tests.el
@@ -72,6 +72,7 @@
     (delete-process server)))
 
 (ert-deftest make-ipv6-tcp-server-with-unspecified-port ()
+  (skip-unless (featurep 'make-network-process '(:family ipv6)))
   (let ((server
          (make-network-process
           :name "server"
@@ -86,6 +87,7 @@
     (delete-process server)))
 
 (ert-deftest make-ipv6-tcp-server-with-specified-port ()
+  (skip-unless (featurep 'make-network-process '(:family ipv6)))
   (let ((server
          (make-network-process
           :name "server"
@@ -171,6 +173,7 @@ server-process-filter
     (delete-process server)))
 
 (ert-deftest echo-server-with-local-ipv6 ()
+  (skip-unless (featurep 'make-network-process '(:family ipv6)))
   (let* ((server (make-server 'local 'ipv6))
          (port (aref (process-contact server :local) 8))
          (proc (make-network-process :name "foo"
-- 
2.20.1.142.g77556354bb


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Michael Albinus]

Robert Pluim <rpluim@gmail.com> writes:

Hi Robert,

>> The failed builds are mailed to emacs-buildstatus, so you can follow that
>> mailing list if you want to get notified about failures.
>
> I should join that list.

In case you use gmane, you could check it there:
nntp+news.gmane.org:gmane.emacs.buildstatus

Best regards, Michael.

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Robert Pluim]

Michael Albinus <michael.albinus@gmx.de> writes:

> Robert Pluim <rpluim@gmail.com> writes:
>
> Hi Robert,
>
>>> The failed builds are mailed to emacs-buildstatus, so you can follow that
>>> mailing list if you want to get notified about failures.
>>
>> I should join that list.
>
> In case you use gmane, you could check it there:
> nntp+news.gmane.org:gmane.emacs.buildstatus

Yes, thatʼs useful.

Iʼve pushed changes to the failing tests under the assumption that
either that host doesnʼt have IPv6 support in its headers, or it has
no IPv6 stack running, but since I donʼt have any hosts like that, Iʼm
guessing quite a bit.

Regards

Robert

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Ted Zlatanov]

On Tue, 05 Feb 2019 15:58:43 +0100 Robert Pluim <rpluim@gmail.com> wrote: 

RP> Iʼve pushed changes to the failing tests under the assumption that
RP> either that host doesnʼt have IPv6 support in its headers, or it has
RP> no IPv6 stack running, but since I donʼt have any hosts like that, Iʼm
RP> guessing quite a bit.

I think this is an artifact of running inside Docker. It does IPv4 by
default. According to the docs we can give the default docker0 bridge a
local IPv6 address or globally routable IPv6 addresses (which we also
have to confirm with the FSF sysadmins). I'm not sure what's better in
the context of building and testing Emacs?

Ted

Re: EMBA (Emacs Gitlab CI build server) is now *public* available

[Robert Pluim]

Ted Zlatanov <tzz@lifelogs.com> writes:

> On Tue, 05 Feb 2019 15:58:43 +0100 Robert Pluim <rpluim@gmail.com> wrote: 
>
> RP> Iʼve pushed changes to the failing tests under the assumption that
> RP> either that host doesnʼt have IPv6 support in its headers, or it has
> RP> no IPv6 stack running, but since I donʼt have any hosts like that, Iʼm
> RP> guessing quite a bit.
>
> I think this is an artifact of running inside Docker. It does IPv4 by
> default. According to the docs we can give the default docker0 bridge a
> local IPv6 address or globally routable IPv6 addresses (which we also
> have to confirm with the FSF sysadmins). I'm not sure what's better in
> the context of building and testing Emacs?

All the emacs IPv6 tests connect to localhost only, so definitely no
need for a globally routable address. We could probably get away with
just "::1" working.

Robert