Acren

Hi Héber, I believe OP is asking about using the API with Plastic Cloud specifically, not running their own API server. I'm interested in this as well, but the link you posted does not mention anything about Cloud, unless I'm missing something. Do you have any more information on this? Thanks

Hi @blue, after having this pop up some more I now believe this also happens without the Unreal plugin ever having been used on the machine either. Just a clean Plastic Cloud install downloaded straight from the website, today's current version. No previous versions, no Unreal plugin. DNS error blocks usage, and config has: <WorkspaceServer>*@cloud</WorkspaceServer>

@Wolfram Thanks for the information but it seems like this happens for us even when using mostly short task branches, although we can't get rid of long running branches completely because they are required for our workflows. @KikeYeah, I'd definitely agree with your idea of desired functionality there. For us, the outer repo should know about the inner repo that it contains, and not the other way around. The inner repo logically doesn't need to care about the outer repo's changesets, branches, merges, or any other kind of history. That's sort of the whole idea of component libraries in the first place. I'm not sure I see how the current behavior can be considered intentional at all though. If it were just creating branches and merges for actual changes to the inner repo that were submitted via the outer repo, at least I could sort of understand how that might be intentional even if it's not ideal. But, the current behavior of pushing all of the outer repo's branches and merges into the inner repo I don't think would make sense in any scenario. Of course we can only really speculate... Take a look at our inner repo's branch explorer: There are now 45 branches and 156 changesets on the inner repo that were automatically created by the outer one. How many actual changes are there to the code on the inner repo? Just a single one. Thanks a lot for pushing this with the Plastic team, hopefully they can fix or improve this.

@blueThere aren't any existing installation Plastic installations on the machines, it's just a fresh download of Cloud Edition from https://www.plasticscm.com/download. Possibly the Unreal Engine plugin if they have downloaded the project but not connected it yet. The value was *@cloud by default as the above reply by Oleg states.

Hi @blue All of my team members get this error after installing Plastic Cloud. It's a bit strange and annoying to have everyone on the team coming to me with this problem, to which I have to say to manually find and update the client.conf file as suggested above. In all cases, the organization appears correctly everywhere in the client except for in the client.conf file. After updating the client.conf this issue is fixed. Do you have any idea why this would be a recurring issue? Thanks

To clarify: The main issue is that when a branch on the parent repo is merged, it appears to also create a new branch and empty merge on the xlinked repo, even though the contents of the xlinked repo was not changed at all. This creates a lot of confusion and clutter. After some more research I've discovered this thread, which seems to be the exact problem we're facing. But the last post was ~2.5 years ago. Has there been any development to address this? Or any workarounds discovered? Thanks.

I've been trying to use the new code reviews with my team so I'll collect some feedback here and update when I find more things: - Initial impressions are it looks much cleaner and a step up from the built-in interface. - I share the confusion above about Unity dashboard and separate accounts. However I can probably guess that the goal is to move everything to use this environment and deprecate the old one. If that's the case it might make sense eventually. - What we are missing the most from GitHub pull requests and other similar systems is the ability to paste images directly into comments. This is really handy when needing to reference something visual in a game by screenshot, or show part of a visual scripting graph. Are there any plans to add this functionality? - We would love email notifications from code review activity like GitHub and competitors. We use plastic triggers for this currently but it's clunky, buggy and confusing. e.g. Deleting a comment currently sends a trigger to our slack with "New comment to the review 'Review Name' > Under review" which is wrong... - It would be helpful to have some sort of "closed" state for code reviews. In some cases we've wanted to abandon the changes proposed in a code review, so "changes required" is not applicable. We also did not want to delete the code review because we want to preserve the conversation to look up in the future. - It unfortunately feels quite slow currently, the front end is quick but it waits a very long time on the backend to do basically anything. For example, it takes about 15 seconds after a code review loads for the comments to appear. - Bug? When the "reviewers" of a code review is set to a group, it doesn't work at all. Nobody can review it even if they are in the group. - Bug? The "files changed" isn't actually files changed in the branch. It appears to be a straight diff of the branch against the target branch. Which means that any changes on the target branch will also show up there inversely, which doesn't seem correct. I think maybe what we should see is a dry-run merge, showing exactly what would be changed on the target branch if it were merged. - Maybe a bug? I would like to @ mention other developers on my team in my plastic code review comment, however the list of users it suggests after typing "@" are not from my organization and I have no idea where they come from. - Bug? The number inside the "Changed Files (n)" tab changes when you select different tabs. - Another bug? Members of my organization can all see the same code reviews, but we cannot link code reviews to each other. The organization ID in the link is different for all of us and we all get error 403 when trying to access a link given by another team member. - Switching to another tab in the code review discards the comment you are writing without warning. This hurts when you are writing a comment and you need to switch to another tab to see the changes and the comment is lost. With something like GitHub, the comment would be preserved when you switch back. - This one is less technical, but with the code reviews, the "rework required" icon being a thumbs down feels a bit negative like it's saying "bad job" when in reality it mostly means "good work but please change this". I'd almost be averse to using it, as I think feedback on changes should be treated as a natural part of the process and not like something is wrong. I hope it helps. I think there's quite a way to go still, but it's definitely in the right direction.

Hi there. We've started trying out xlinks to manage our component repos, and it seems to work well apart from one thing: The branches on the parent repo getting created on the child repo is really undesirable from our point of view, because these branches only make sense in the context of the parent repo and mean nothing to the child, they are also all just empty merges and don't contain anything meaningful. So we also get all of these automatically created empty branches cluttering the child repo and creating triggers and notifications where we don't want them to, and I can only see this getting much worse when there are more xlinks as well. Can we disable the auto expansion and keep the xlink writable? The desired behavior would be to always use the /main branch on the child and not create other branches. Essentially this would be the same thing as having an expansion rule for every branch on the parent repo mapped to /main on the child repo, but I'm not sure how to specify this, maybe I'm missing something because I'm still new to xlinks. Thanks!

Thanks for the advice Carlos, How can I use the cm diff command to diff a specific changeset with the local working copy?

Hello, I'm trying to get a list of files modified after a certain date, but excluding modifications in any changesets which are not present in the working copy. I'm currently trying to use cm find revisions to get this working, and filter by date, but I can't figure out how to filter out the revisions that aren't in the working copy. It returns revisions which are made in different branches and not merged yet, which I don't want. How can I achieve this? Thanks

Hey there. I currently have a job configured on Jenkins to poll a single branch for changes, then of course checkout the changes and run the job as you'd expect. However I want the job to run on every branch that has changes, not just a specific one. How can I achieve this? Much thanks.