nahuelarjonadev

Hi, as a quick update we are moving back to SVN for now. The concept of Plastic sounds great but our team considered it is not ready to handle big projects with binary files well. Definitely not a big Unreal Engine project as ours is. I gotta say the Plastic team was always blazing fast to answer our questions, so thanks for that. We may consider giving Plastic another try in the future, if features like traveling locks or a better integration with Rider get implemented. I'll keep a close eye on Plastic as the idea is so promising, but due to its current limitations we ended up having to use single branch development without distributed repos. I'll keep an eye of how it develops.

I'm cross-posting from old post this to keep all conversations about distributed locks in the same place: Hi @calbzam, thanks for your quick answer! After reading this post, I now see that the one issue with my trigger-based solution is that someone can request a valid yet non-useful lock from a local repo. If user Alice pushes a change to file A therefore releasing the lock for A, and then user B requests the lock for file A without pulling changes first, user Bob would acquire the lock successfully despite not having the latest A version. I guess the before-checkout operation should verify both 1) that the file is not locked on the central repo, and 2) that the requesting user has the most up-to-date version of the file in comparison to main. Any ideas on how I can implement a before-checkout trigger that accomplishes these 2 checks? Assuming everyone with a local server uses the same triggers pointing at the same central server, I think that would cover my team's needs. Am I missing something? Regards, Nahuel

Edit: Moved answer to

Hello there! Sorry for reviving this really old post, but it was the first that showed up when googling the topic of exclusive checkouts in a distributed environment. Is there a way to make exclusive locks work on distributed environments right now? E.g. something similar to this workflow: I request a Lock and Checkout on a file on my local server. That operation, before running locally, triggers a Lock and Checkout operation on the central server. The central server sends back a response stating if the Lock and Checkout operation on the central repo succeeded. If it succeeded, then the file is also checked out and locked locally. If it failed, the user is warned that someone else already has that file locked and the local operation doesn't go through. Similar functionality would happen for checking in files. Effectively, "local exclusive checkout/checkin" would become a mirror of "server exclusive checkout/checkin". Motivation for this ask: We are using Plastic Cloud (paid customers), and several of us work on both binaries and text files, binaries rarely being less than 30% of the work. Not being able to easily checkout binaries exclusively is enough of a reason to ditch the distributed workflow entirely, and be forced to use Gluon with a centralized repo workflow. This kind of defeats the purpose of us having switched over from SVN, as one of the (if not THE) biggest reasons we decided to give Plastic a try was to get a distributed workflow for "mergeable" files as a possibility. So, even if this is not officially supported yet, could you hint me about ways to achieve this via some workarounds? Using triggers maybe? How can I kick operations on the central repo from a script or something like that? Thanks for your understading!