Looking for maintainers / developers

[klauer]

Questions

• Do you use caproto at your facility?
• Do you want to continue using caproto at your facility?

Background

• The original developers no longer have much time for maintaining caproto.
• You might see us submitting bugfixes here or there, but there are no current goals for the project.
• You shouldn't expect fast support or turnaround time for issues/PRs.

Who?

• @danielballan, @tacaswell, and @klauer initially developed this library over 5 years ago now
• We have had a few contributions from users here and there - thank you! It's been a fun ride.
• @klauer has been almost solely developing/maintaining the library for the past few years. He has many other projects that require his time and effort. caproto has long since taken a backseat to those other projects.
• NSLS-II will make take care of critical bug fixes and keeping up with upstream changes

Alternatives

There are many better-supported alternatives to caproto these days, you might consider using them instead.

On the server side, some I know of are:

• https://github.com/mdavidsaver/pyDevSup
• https://github.com/klemenv/PyDevice
• https://github.com/paulscherrerinstitute/pcaspy
• https://github.com/mdavidsaver/p4p (PVAccess)

On the client side, there are too many to list:

• https://github.com/pyepics/pyepics
• https://github.com/dls-controls/aioca
• https://github.com/dls-controls/cothread

[vstadnytskyi-FDA]

This is very unfortunate. I really liked caproto. Unfortunately, I do not know if I have expertise to support this project. I just created to new issues that I have no idea how to tackle.

[toqduj]

That is unfortunate indeed, I very much enjoyed its cleanliness and the ease of install... pip install caproto and voila, a useful running epics backend. I hope someone will step up?
(contrast, for example, the installation procedure for pcaspy: https://pcaspy.readthedocs.io/en/latest/installation.html#installation-label which is not for the faint-hearted).

[vstadnytskyi-FDA]

That is unfortunate indeed, I very much enjoyed its cleanliness and the ease of install... pip install caproto and voila, a useful running epics backend. I hope someone will step up? (contrast, for example, the installation procedure for pcaspy: https://pcaspy.readthedocs.io/en/latest/installation.html#installation-label which is not for the faint-hearted).

The caproto underlying code is too complicated for me. I am good with it as a user but not maintainer. Unfortunately, I do not see how I can pick up the slack.

I have success with pcaspy. For example here https://github.com/vstadnytskyi/caproto-sandbox/tree/master/caproto_sandbox/transition/pcaspy, I have an example of a server that sends CPU, Memory usage, etc. and a GUI that visualizes it https://github.com/vstadnytskyi/caproto-sandbox/blob/master/caproto_sandbox/simple_camera/gui.py

On windows 10 I just did pip install...

[tacaswell]

I do not expect carproto to suddenly stop working (at a minimum I'm going to keep it working with new versions of Python because we use it in the ophyd tests) and it is not like the CA protocol is changing.

I do not expect new feature requests to get resolved quickly (if at all) and user support may be limited. but if it works then I expect it to keep working.

At NSLS-II we just ended up with a caproto IOC (https://github.com/NSLS-II/archiver-proxy) on our critical path....

[clintonroy]

I'm happy to put my hand up to help out; while I'm at a synchrotron using epicsy stuff.

[tacaswell]

@clintonroy you should have an invite to join the org in your inbox

[mrakitin]

I apologize, I accidentally unpinned the issue thinking it was my personal config. I pinned it back.

[codedump]

Hello,

I'd like to announce my intention of helping out with the project as much as I can. I'm currently at a "synchrotron doing epicsy stuff" (to put it in @clintonroy's words :-) ), that would be BESSY-II in Berlin.

My time and resources will likely be limited, and to add insult to injury, I'm pretty much caught in the prevalent "postdoc forever / time-limited contract" hamster wheel, so I might not be around for a very long while. But for as long as it lasts, I love caproto, and my colleague Matthias and I have pretty much gone "all-in" on it with our beamline setup. We've used a little bit of pythonSoftIOC and other frameworks along the way, but CAproto just "feels" best and most fun to work with. I found out that I'm generally a fan of the CA protocol itself, unlike the rest of the tooling and philosophy around the "original" EPICS project.

A steady number of CAproto-IOCs is growing, based on the asyncio side of code (see our unofficial beamline repo is here), and with Matthias's support I've just made the conscious decision to ride it out 'till the bitter end. for as long as it lasts. I've made peace with having to fix one or the other bug that comes our way myself. When that's not possible anymore everything might go up in flames at the beamline when I'm gone, but then again... ¯\_(ツ)_/¯

Anyway.

What's next?

How can I help with limited resources?

Is there anyone else around I can play-pretend "the project is not dead yet" with, or am I pretty much on my own and should just as well clone the project into my personal github space and sulk on it there? :-p

Currently I've found a bug that essentially prevents asyncio-based CAproto clients from running in pytest test harnesses. I'll be reporting it shortly, together with a (dirty) workaround.

Cheers,
F.

[tacaswell]

Yes! We are still alive here, just maybe a bit slow.

Please report bugs and PRs will get reviewed (eventually).

[tacaswell]

also, send me an email at tcaswell@bnl.gov and I'll get @codedump and Matthias set up with commit rights (given the xz fiasco I would like to have a zoom call with both of you first 😉 ).

[codedump]

also, send me an email at tcaswell@bnl.gov [...]

Have you received my email? I sent it 3 days ago, but haven't seen an answer from you yet.

If it got lost, I can send it again (?) or you can look me up -- I'm this guy, also available with an alternate email address here.

I had sent you the email from my private address (which I prefer for communication regarding projects that are likely or supposed to survive my affiliation to particular research institutes, because my institute-bound email addresses change like laundry :-p )

Cheers,
F.

[tacaswell]

Yes, sorry just underwater on my end!

[codedump]

No worries, do it in your own time. Just wanted to make sure nothing got lost on the way.