The problem with no releases....

Thanks for all the replies. Here is the rawpedia page with broken links to nightly builds.

The “development build” link simply points back the the same page.

@Barugon You are right, but how do you get to this page specifically? This is a page in the ‘es’ (Spanish) portion of the wiki, however it is written in English and is not linked to by any “official” English page as far as I am aware.
@XavAL do you know more about this?

@Thanatomanic:

Well, that’s somehow my fault…

There was a time when people couldn’t easily find proper dev releases and there was a thread where I linked to that specific RawPedia page. Note I wrote it myself, but was not near «official», thus there is no link to it anywhere at RawPedia. Think of it as a temporary patch. In such thread it was suggested that the links should be updated periodically, but honestly I completely forgot about it. Sorry.

I will update links on that «hidden» page, and I will edit my post in the old thread, so hopefully there will be no more misunderstandings.

EDIT: work done. Well, it seems that I can’t edit my old posts, so the old thread remains untouched. Hopefully with the edited RawPedia page, things will be a little clearer.

1 Like

I typed “nightly build” into the search field on the left.

@Morgan_Hardwood @Thanatomanic @jdc @heckflosse @Lawrence37
In the near future there seems to be no possibility to get out a 5.9 release of RT. If this is true: would it be an idea to skip the 5.9 release and work on a 6.0 release. Since last release there is a lot of new stuff in dev and there are some PRs that could be merged. in my opinion things like locallab, favorite tab and a lot of additional addons and cleanups legitimate a 6.0 release.

Doesn’t matter whether it’s a 5.9 or a 6.0 - same problem. There isn’t really anything significant that is currently being impacted by the 5.9 feature freeze to my knowledge at the moment (IIRC, the feature freeze does not affect camera support, so anything related to camera support is still going through, such as the LinearizationTable fix for some DNGs), so there isn’t a benefit to “unfreezing”.

Pretty much everything tagged for 6.0 in Pull requests · Beep6581/RawTherapee · GitHub is either a minor UI enhancement or an under-the-hood cleanup, and it would be better to get a release out with what’s currently merged or tagged for 5.9 than to “unfreeze” and wait for those 6.0 items to stabilize. Especially since “unfreezing” doesn’t solve the inevitable problem that 6.0 may eventually have.

I don’t know if @Thanatomanic or @heckflosse have reached out to @Morgan_Hardwood privately to try and get permission for at least a temporary handoff of release responsibilities to them - he’s not really responding to any tagging of his public accounts, I don’t know if they have a better way to reach him. I can understand burnout or wanting to take a break (been there done that), but a handoff to a delegate like Roel or Ingo would’ve been nice (I’ve been 100% retired from a large project I co-founded for years, but in my last year or two on the project I did a lot of focusing on trying to train/mentor potential replacements, and I like to think I did a good job because the project still seems alive and healthy.).

5 Likes

Yes I did. He hasn’t seem to have logged in to Pixls since mid-January though. No replies to any post he has been mentioned in. I sure hope he is all right…

3 Likes

For me, especially regarding “LA”, everything is day with 5.9 (of course we can always improve…). The remaining problem is the evolution of the GUI which is not my competence.

We are in an imbroglio where we wait for an answer from someone who doesn’t answer, and where the actors present are all full of good will, but who don’t have each (a priori) all the competences , or rights (unless I’m wrong) to realize the update towards 5.9

What is the way out?

  • Stay on 5.8 and make official the “rotating” updates (which a priori work). In this case, it would be enough (I don’t know how to do it) to change the Rawtherapee Web home page, with the announcement of the new features, the links to the “rotating” versions (Windows, Linux, MacOS) , whitout changing tag .
  • Or something else?

But this situation is not normal, everyone (developers, contributors, users,…) is waiting, leading to annoyances, disinterest, etc. RT’s image is fading .

Jacques

3 Likes

We can do that if that is what the team wants. Might be interesting to support a rolling release only.

3 Likes

@paperdigits

I think it’s a way out, maybe transitory…But it would have the merit of doing something.

Moreover the “Rolling Releases”, associated with a regular update of the website, would show to the media and users (and of course also to the team) that RT is active and that it is a very good product.

What others think about this suggestion (developers, contributors, users…) ?

Jacques

8 Likes

Word of the day. :nerd_face:

1 Like

I have push rights for the website repository, so I can make changes. However I haven’t managed to get a local version working yet for testing. I will try again soon. After that it should be straightforward to place a news item with some explanation.

12 Likes

Love this idea. I think this is the best solution given the circumstances. :slight_smile:

1 Like

I do too. He’s mentioned a house remodel - maybe that got hosed up by supply chain issues. I know lumber prices spiked horribly over the past few years, and other stuff is screwy too.

I wasn’t sure if you or @heckflosse (some of his closest co-contributors on the project) had a more direct method (alternative email, phone number) for reaching out. I’ve found one possible phone number on a rather outdated website, but I’m definitely NOT calling that number myself, and I’m guessing you’ve probably already seen that number at some point too…

I think at this point we don’t need to rehash the story about our absent friend or @ him but look toward what we could do moving forward. Imagine him visiting one day and seeing 1000 mentions or @'s. I would nope back out. :stuck_out_tongue:

1 Like

It’s a delicate thing - because ideally, out of respect for him, I don’t think anyone wants to take a project that he’s been the lead of for a long time against his will. But there is a need to make the project go on, and people with the motivation to go ahead.

Ideally someone (preferably someone he has a really good working relationship with) could reach him and get two simple questions answered:

  1. Is he OK? Some of us are worried about him!
  2. Since he’s obviously got a lot on his plate, and hey, he deserves a break, would he be OK with someone else like Roel taking over the release manager role for 5.9?
2 Likes

I will leave the decision about the release strategy to others because I have only been a contributor since recently. However, I want everyone to be aware of the challenges of adopting a new release schedule and be prepared accordingly.

A rolling release, for example, makes stability less clear. We have a dev branch that is usually more stable (in terms of bugs) because of bug fixes. Occasionally though, there will be a significant change (think exiv2) or an unexpected regression. We also have high standards for pp3 compatibility between stable releases. This needs to be reevaluated for a rolling release.

I hope we can figure things out and get a new version of RT out to the general public soon. There have been so many great improvements since 5.8.

7 Likes

Yes. A rolling release is only slightly different from the already-existing dev builds. (I am, in fact, having difficulty conceptualizing anything other than occasionally version-bumping dev without a formal release process…) Also has a greater risk of “wasting a translator’s time” which seems to be a concern. Although at least one translator has gone ahead anyway recently, maybe it’s time to proceed at risk with that regard? If it’s beneficial I’m thinking of writing a script to look for orphaned strings. Something like Weblate, IMO, should wait for post-5.9

Very good point on the PP3 compatibility concern.

As far as release process goes, I’m questioning whether or not there is a benefit of doing a “release candidate” build prior to the formal release. I think there have been platform-specific glitches (primarily Windows IIRC?) in both of the past two releases from what I recall. On the other hand, the fact that Windows and Linux are automated (is MacOS automated?) with CI now should hopefully mitigate this concern.

Perhaps it’s time to introduce a “release” branch, where all the non-breaking fixes get merged to, so some of the “6.0” PRs can get merged? There is nothing more frustrating than investing hours into a proper PR and then it goes nowhere. It might also help to get releases out faster as features are out to the public earlier and can be tested during the weeks (months/years…) of releasing.