Recompile switch9/20/2023 ![]() ![]() I've been using Unity since the 1.x days, and past a very early prototype project with maybe a scene or two, something breaks during a hot recompile. Is this for some non-game use (configurable customizer or something)? I've asked around over the years, and I've never found a single project that is compatible with it, let alone uses it as a fixture of their workflow. I don't understand why it would be removed at all.įurthermore, I don't understand which projects are even capable of surviving a recompile while in play mode. I just noticed this and posted on Twitter, and saw this thread after a search there: Prevent Recompile During Play Mode (TRUE/FALSE)īut to be honest, I'd much rather have just kept the original "Script Changes While Playing" option determine this because it's worked that way as long as I can recall and it lead to much confusion when it quietly changed.Auto Refresh (same as current dropdown option).My personal workflow is to always restart play mode after recompile.įor what it's worth, if we have to direct this choice through the new Auto Refresh option then my ideal solution would be to have the existing auto-refresh option for assets, and then have an additional option for code changes. In fact, I never want to recompile code during play mode. There are times when I would want to reimport assets (such as art/audio) but NOT recompile code changes, and may need to reimport one but not the other. The new auto-refresh option is OK, but still not ideal. I came looking for others experiencing this issue and just wanted to add a +1 to the original post. Unfortunately I can't figure out how to add a poll here as I'm interested if we're the only ones who miss the feature. If I change a 2 gig Photoshop file, I may not want to reimport it immediately. If I change a script, I want to recompile. what we really want is to have option to have "only" scripts to be auto refreshed outside playmode, but not the all other assets." I later found this entry by which makes way more sense: It has been replaced with a new auto refresh feature in `Preferences -> Asserts Pipeline -> Enabled Outside Playmode`" When activating the "Enabled outside playmode", my scripts aren't compiled at all anymore. "This features has been depricated due to it not working as intented. Now maybe I didn't get it, but the recommended "solution" is not working the same as before, so I wonder if someone can explain to me how I can get the old behaviour again? It was one of the *best* features added in the past and decreased our iteration times big times! I assumed this to be a bug and searched for it and then found the issue in the bug tracker: Without this "Compile after playing" feature our iteration times will increase by a lot. It’s a shame you won’t be in the UK for SQLBits this year.I switched from Unit圓d 2019.4 to 2021.3.9f and found that the "Recompile after finished playing" feature has been removed from the editor. I really shouldn’t be try to have a technical discussion with the likes of Brent after this many beers. Would it not be a benefit if the optimiser did the same and invalidate the plan after the fact? Or do you think we’d end up in the same sort of flip-flop that you’ve seem with AMG? We look for queries where the estimated and actual number of rows are significantly out. But for a more mixed environment, would I really want my reports to keep recompiling? | I’ve got my ETL queries that run every hour, and I don’t give a hoot about plan re-use, the data volume pattern is all over the place. Is it just high cost queries that are the issue, or is it a function of cost and frequency of execution. It feels that a Cost Threshold for recompiles isn’t the right tool. ![]() The rise of OLAP, better parallel plans, and smarter more tolerant people saying “Cost Threshold for Parallelism” is the setting to use changed the Best Practice of the herd. MAXDOP = 1 was the wrong setting of course. And any reporting queries that could benefit from parallelism just wasn’t worth the effort. I think we were all running OLTP systems and we didn’t want to deal with parallelization. Over 70% of us were setting to MAXDOP = 1, and the rest couldn’t be bothered to put their hands up. You could feel the mood of the crowd turn ugly. I was at a London conference sometime post-SQL2000, pre-SQL2005, and the presenter started talking about the benefits of parallelization. ![]()
0 Comments
Leave a Reply.AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |