Fast-RTPS external dependency Quality Declaration

[Blast545]

This is a quality declaration for eProsima Fast-RTPS external dependency for ROS2 packages as described in the proposed REP-2004.

This declaration represents some thoughts regarding the current state of this external dependency and how this affect Quality Level 1 for ROS2 packages. Open to discussion, any thoughts are welcome.

Signed-off-by: Jorge Perez jjperez@ekumenlabs.com

[brawner]

It looks like this needs to be updated for our most current style.

[ahcorde]

please review typos and links

[ahcorde]

LGTM

[JaimeMartin]

Please do not close this PR yet. In a quick review, I see there is missing information, for example, we do have public coverage reports.

@MiguelCompany , could you help on this one?

[Blast545]

@JaimeMartin Sure! Let me know if there is anything missing before merging, I opened a similar PR in #360 that maybe you'd like to give a check as well.

[Blast545]

@JaimeMartin @MiguelCompany are you checking on this? Please let me know, I'd like to get this merged without missing information.

[MiguelCompany]

@Blast545 Yeah, I have some comments on this.

We are now enforcing DCO on the PRs. You can see it, for instance, on PR 1190. We still have to add this to our contibuting.md, but we will have it before 2.0.0 is released.

Regarding coverage, our nightlies do have coverage reports, see an example here

Even though we have committed mistakes in the past, we do try to follow Semantic Versioning. We have encouraged all the team to be specially careful on the reviews about changes on public headers, and will not allow mistakes from 2.0.0 onwards. API breaks should only be allowed between major version changes, if an ABI break is required, it should be done between minor versions and be clearly stated on the release notes.

[ahcorde]

@MiguelCompany what's about the public coverage reports that @JaimeMartin mentioned? Can you provide us a link or suggest the changes?

[MiguelCompany]

I accidentally published my comment unfinished. It's been edited.

[ahcorde]

do you have these lines explicity written somewhere in your documentation?

API breaks should only be allowed between major version changes, if an ABI break is required, it should be done between minor versions and be clearly stated on the release notes.

[MiguelCompany]

do you have these lines explicity written somewhere in your documentation?

It is here

[EduPonz]

Hi @Blast545 @ahcorde @brawner ,

At eProsima, we have been working on Fast DDS Quality level and we'd really appreciate your input in a collection of pull requests that in our opinion qualify for a Quality Level 2 once merged. Those are:

• Fast DDS Quality Declaration: Add Quality Declaration and update README [9371] eProsima/Fast-DDS#1400
• Fast CDR Quality Declaration: Add Quality Declaration [9390] eProsima/Fast-CDR#85
• Foonathan Memory Quality Declaration: Add Quality Declaration for foonathan_memory external dependency [9398] eProsima/Fast-DDS#1404

Please mind that this is a work in review, some links in the documents do not work yet. For those cases, the PR descriptions link to the relevant documents. Feel free to comment and make suggestions in any of the related PRs. Once we have those sorted out, we'll be ready to update this PR to a Q2 for Fast DDS.

Thanks!

[EduPonz]

Hi @Blast545 @ahcorde @brawner ,

All our quality declarations have been merged now, with the exception of Fast DDS' (eProsima/Fast-DDS#1400), which @brawner already approved. If it's OK with you guys we'll merge that too, and then this PR should be updated to point to the QD of its dependencies directly (instead of stating them here). All in all, Fast DDS qualifies now for a Quality Level 2, and so this this package.

@Blast545 Could you update the PR to reflect those changes? IMHO it's the only thing left to merge the PR.

Thanks!

[chapulina]

Fast DDS qualifies now for a Quality Level 2

That's great news, congrats!

point to the QD of its dependencies directly

Agreed, we don't need a QD for FastDDS in this repo, we should link to the one upstream.

[EduPonz]

Due to Fast DDS reaching QL1 (eProsima/Fast-DDS#1610), I think this PR can be closed