Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Use Docker consistently #3884

Merged
merged 7 commits into from
Feb 29, 2020
Merged

Conversation

greg0ire
Copy link
Member

@greg0ire greg0ire commented Feb 28, 2020

Q A
Type improvement
BC Break no
Fixed issues not an issue, but #3883 (review)

Summary

When a Docker version of the build is available, use it. Also, remove jobs that test the same version as a docker job.

@greg0ire
Copy link
Member Author

greg0ire commented Feb 28, 2020

No idea if this is actually going to work, let's see…

Spoiler alert It doesn't

@greg0ire greg0ire force-pushed the use-docker-for-all-rdbms branch from f46dc27 to 7fb5938 Compare February 28, 2020 21:15
@greg0ire
Copy link
Member Author

Ok it works now. Please review.

@greg0ire greg0ire requested a review from a team February 28, 2020 21:28
SenseException
SenseException previously approved these changes Feb 28, 2020
Copy link
Member

@SenseException SenseException left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A few minutes ago, I wanted to write something about 7.4snapshot, but your recent commit is handling that. 😁

@morozov
Copy link
Member

morozov commented Feb 28, 2020

I’m all for this with the exception that it will increase the likelihood of #3478 which is already PITA. The tests against Docker-based MySQL (and perhaps MariaDB) images fail from time to time because, despite the readiness probe passing, all subsequent queries fail with "2006 Server gone away".

Before we go all Docker, I want this issue to be addressed or at least mitigated.

Maybe instead of doing \q, we should execute some real query like SELECT NULL FROM information_schema.something in:

sudo docker exec -i mysql80 bash <<< 'until echo \\q | mysql doctrine_tests > /dev/null 2>&1 ; do sleep 1; done'

@greg0ire
Copy link
Member Author

@greg0ire
Copy link
Member Author

Not sure if mysqladmin is available in the image, let's see…

@greg0ire
Copy link
Member Author

I think I have something more to do to force waiting for the container to be healthy.

@greg0ire greg0ire force-pushed the use-docker-for-all-rdbms branch 4 times, most recently from 8f5b3b4 to b18ec2a Compare February 28, 2020 22:43
@greg0ire
Copy link
Member Author

Done, and it seems to work. @morozov what now? Let's restart the build many times?

@morozov
Copy link
Member

morozov commented Feb 28, 2020

Let's restart the build many times?

I thought about the same. Maybe 5-10 times to be on the safe side. But before that, since we're generalizing the dependencies, why don't we generalize the install-*.sh scripts for MySQL and MariaDB?

Instead of copy/pasting the code, we could introduce a required CLI argument that would take a Docker image name to be spun up like: docker-run-mysql-or-mariadb.sh mysql:8.0.

@greg0ire
Copy link
Member Author

I'm on it. I'm going to use an environment variable to specify the image, that way you can still see what the version is in the overview of the jobs.

@greg0ire greg0ire force-pushed the use-docker-for-all-rdbms branch from 073e994 to e029896 Compare February 29, 2020 10:52
@greg0ire
Copy link
Member Author

greg0ire commented Feb 29, 2020

Not 100% sure but I think I just made these lines useless:

dbal/.travis.yml

Lines 16 to 17 in 96d10bd

before_script:
- if [[ "$DB" == "mysql" || "$DB" == "mysqli" || "$DB" == *"mariadb"* ]]; then mysql < tests/travis/create-mysql-schema.sql; fi;

I think they are not in use for (some?) jobs that end with .docker, correct? Why is that?

@greg0ire
Copy link
Member Author

Ah that's correct, I thought global and local before_script were merged, they aren't, it's an override.

@greg0ire greg0ire force-pushed the use-docker-for-all-rdbms branch 4 times, most recently from 04dbe80 to e39698d Compare February 29, 2020 15:10
@greg0ire
Copy link
Member Author

Green build again :)

Do you think I should get rid of the .docker part in DB env variables and xml filenames?

@greg0ire
Copy link
Member Author

There is an infinite loop now, many containers are unhealthy…

@greg0ire greg0ire force-pushed the use-docker-for-all-rdbms branch from e39698d to 37ec53c Compare February 29, 2020 15:35
@greg0ire
Copy link
Member Author

I added a check that will fail the build on unhealthy container. I noticed docker is run with sudo, why is that?

@greg0ire
Copy link
Member Author

greg0ire commented Feb 29, 2020

Ok so it's unhealthy because I badly implemented the addition of the extra command option, no weird transient issue :)

@greg0ire greg0ire force-pushed the use-docker-for-all-rdbms branch 2 times, most recently from 8f00e8b to 1215beb Compare February 29, 2020 16:21
@greg0ire greg0ire force-pushed the use-docker-for-all-rdbms branch from d9612c3 to 398f23e Compare February 29, 2020 19:05
@greg0ire
Copy link
Member Author

🤔 no Scrutinizer job now…

@morozov
Copy link
Member

morozov commented Feb 29, 2020

I'm really not for dropping things as soon as possible, but supporting platforms that are not supported by the vendor themselves seems to be another extreme.

I agree and I'd rather drop their support in a minor release than be puristic about semver and keep dragging this legacy. Although, this PR doesn't drop support for legacy platforms but instead stops testing still supported code.

It seems mostly related to IBM DB2… I did touch that job to remove the services: [docker] section, but can't see how that would explain this drop in coverage.

It may be a Scrutinizer bug (because it's buggy as hell) but maybe not. If we don't see this issue on other PRs, then most likely this one is causing the trouble.

@greg0ire
Copy link
Member Author

It may be a Scrutinizer bug (because it's buggy as hell) but maybe not. If we don't see this issue on other PRs, then most likely this one is causing the trouble.

I found a setting about tracking branches, and added 2.10.x as a tracked branch. let's trigger a new build.

@greg0ire greg0ire closed this Feb 29, 2020
@greg0ire greg0ire reopened this Feb 29, 2020
@greg0ire
Copy link
Member Author

Although, this PR doesn't drop support for legacy platforms but instead stops testing still supported code.

This was done in #3883 actually. How do we usually drop support for legacy platforms?

@morozov
Copy link
Member

morozov commented Feb 29, 2020

How do we usually drop support for legacy platforms?

AFAIK, we only do that in the next major as required by semver. Doing it earlier will require at least the team’s approval. Apart from that, the community is quite serious about maintaining the BC promise even if there's no practical sense.

@greg0ire
Copy link
Member Author

greg0ire commented Feb 29, 2020

Ok so this time, no coverage drop. Since 2.10.x had no Scrutinizer job (because of settings), there might have been an unnoticed coverage drop there. 😓

@greg0ire
Copy link
Member Author

Apart from that, the community is quite serious

This tweet is unavailable

@greg0ire
Copy link
Member Author

BTW, apparently, it's possible to specify distro on the per-job basis in Scrutinizer.

Did you mean Travis?

@greg0ire
Copy link
Member Author

greg0ire commented Feb 29, 2020

Anyway, there's probably no need to change distros if we use Docker, right?

EDIT: Ah but that would mean using Docker for Postgres…

They are no longer supported by their vendors, but until we do support
them, we should continue testing them.
@greg0ire greg0ire force-pushed the use-docker-for-all-rdbms branch from d1aaf21 to 6e505de Compare February 29, 2020 20:21
@morozov
Copy link
Member

morozov commented Feb 29, 2020

This tweet is unavailable

Weird. The tweet is available (Dec 29, 2019) but the link seems to work only if clicked from the twitter feed, not from another website.

Did you mean Travis?

Yes, Travis.

Ah but that would mean using Docker for Postgres…

Yes. We can do it later if needed.

@greg0ire greg0ire merged commit f089a93 into doctrine:2.10.x Feb 29, 2020
@greg0ire greg0ire deleted the use-docker-for-all-rdbms branch February 29, 2020 20:49
@greg0ire
Copy link
Member Author

@SenseException well that was harder than expected :P

@SenseException
Copy link
Member

Seems like it.

@morozov
Copy link
Member

morozov commented Feb 29, 2020

[…] well that was harder than expected :P

You shouldn't have requested my review 😄

rgrellmann added a commit to Rossmann-IT/dbal that referenced this pull request Apr 22, 2020
Release [2.10.2](https://github.com/doctrine/dbal/milestone/75)

2.10.2
======

- Total issues resolved: **4**
- Total pull requests resolved: **19**
- Total contributors: **10**

Improvement,Static Analysis
---------------------------

 - [3964: Mark every exception as immutable](doctrine#3964) thanks to @greg0ire

CI,Improvement,Static Analysis
------------------------------

 - [3961: Stop relying on the master version of Psalm](doctrine#3961) thanks to @greg0ire
 - [3951: Setup static analysis with Psalm](doctrine#3951) thanks to @greg0ire
 - [3799: Upgrade to PHPStan v0.12](doctrine#3799) thanks to @lcobucci

Improvement,Logging,Test Suite
------------------------------

 - [3957: Reworked LoggingTest to be able to test Statement::executeUpdate()](doctrine#3957) thanks to @morozov

CI,Code Style,Improvement,Strict Typing
---------------------------------------

 - [3955: Remove baseline](doctrine#3955) thanks to @greg0ire

Bug,SQLite,Schema Introspection,Schema Managers
-----------------------------------------------

 - [3937: Column comment incorrectly introspected on SQLite](doctrine#3937) thanks to @morozov

Bug,Documentation,Prepared Statements,Query
-------------------------------------------

 - [3896: Updated documentation for QueryBuilder::execute() return value type](doctrine#3896) thanks to @morozov

Bug,Prepared Statements
-----------------------

 - [3894: Make sure that the $types array has the same keys $params](doctrine#3894) thanks to @morozov
 - [3893: Ensure the constructor arguments are passed to custom classes](doctrine#3893) thanks to @duncan3dc
 - [3843: Fix unquoted stmt fragments backslash escaping](doctrine#3843) thanks to @morozov

Documentation,Improvement
-------------------------

 - [3886: Update readme](doctrine#3886) thanks to @greg0ire
 - [3834: Fix docblock typos in DriverManager docs](doctrine#3834) thanks to @CHItA

CI,Improvement,MariaDB,MySQL
----------------------------

 - [3884: Use Docker consistently](doctrine#3884) thanks to @greg0ire
 - [3478: Improve readiness probe stability for containerized databases on CI](doctrine#3478) thanks to @morozov

 - [3883: Fix broken build](doctrine#3883) thanks to @greg0ire

Bug,Documentation,Query,Query Limit/Offset Modification
-------------------------------------------------------

 - [3842: Fixed the QueryBuilder::setMaxResults() signature to accept NULL](doctrine#3842) thanks to @morozov

Bug,Query
---------

 - [3832: Fix JOIN with no condition bug](doctrine#3832) thanks to @BenMorel

Bug,PostgreSQL,Schema Introspection
-----------------------------------

 - [3821: &doctrine#91;pg&doctrine#93; fix getting table information if search&doctrine#95;path contains escaped schema name](doctrine#3821) thanks to @linniksa

Documentation,Improvement,Logging
---------------------------------

 - [3812: Fix DebugStack#queries docblock type](doctrine#3812) thanks to @ostrolucky

Bug,Regression,Schema
---------------------

 - [3790: fixed unqualified table name of fk constraints when using schemas](doctrine#3790) thanks to @stlrnz and @Alarich

# gpg: Signature made Mon Apr 20 19:59:36 2020
# gpg:                using DSA key 2C3A645671828132
# gpg: Can't check signature: public key not found

# Conflicts:
#	README.md
@github-actions github-actions bot locked as resolved and limited conversation to collaborators Aug 4, 2022
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants