I guess I should have added an emoji.
Something is happening in Catalina that wasn’t in Mojave. With both 2019r1.1 and 2019r2.1.
I got it, but we need a [sarcasm] … [/sarcasm] tag in Esotalk 
You can also try pre-unzipping the plugins to make it one less thing that the IDE has to do at launch.
There is a thread about this somewhere on the forums, but basically, rename the plugins to pluginname.zip and then uncpomress them letting the resulting folder replace the old zipped .xojo_plugin file.
Followup to this out of interest: now that my expensive fairly new (a few months) 2TB Samsung T5 USB SSD failed (won’t be buying Samsung again, tho I wonder if APFS or Apple’s encryption could be at fault? Can’t be sure. The Samsung still shows up with diskutil list, so maybe not a complete hardware fail?), I temporarily moved to a vanilla Seagate external USB HD, and Xojo starts up and loads in 30 sec.
2 differences:
- the Samsung was APFS vs Seagate is HFS
- the Samsung was encrypted vs Seagate is not: I suspect, but can’t be sure, that this is the issue.
[quote=467917:@Bob Keeney]The problem is that Xojo doesn’t use plugins (except for the few DB plugins) and therefor doesn’t see the pain many of their users go through with the very slow startup and plugin compile times. It would be great if they practiced what they preach and spawn helper processes to load and compile plugins.
Anyone have a Feedback report that asks for this?[/quote]
We are well aware of the pain and already have a plan to address this which is to allow plugins to be project-specific. This is part of the plan to have plugins written in Xojo.
I hope we get it sooner as Android …
This mirrors my experience as well. I was quite dissappointed with the 16". I’d done my homework before hand and was expecting a 1.9x increase in single core and 2.4x increase with multi-core, compared to a 2012 Retina MacBook Pro running macOS 10.14.
Simply starting up this machine was only 1.25x faster and that was with a clean install of Catalina. Other day-to-day tasks didn’t gain a noticeable improvement. Compiling was quicker, but also not by as much as I’d expected.
The machine had multiple issues with Catalina, most were caused because I’d used the Migration Assistant during set-up to copy across my data, which Apple Support said they know it causes issues and I should do a slow migration instead. But the discovery of a pretty major bug (to the task I was working on at the time) and spending 11 of my 14 days (to return the machine) working with Apple to get them to “see” the bug, forced my hand to return it.
Downright the worst experience I’ve had with an Apple product in 26 years.
There’s this article as well.
https://sigpipe.macromates.com/2020/macos-catalina-slow-by-design/
@Sam Rowlands: very soon you won’t be able to start applications without access to the internet. Great blog post.
We use exactly the same plugins in different projects so having project specific plugins wouldnt improve startup times for us. Surely there must be a way to make this process more efficient.
Project specific plugins sounds good but in every day practice it wouldn’t do much for me because it’s just as easy to have a minimum set of plugins. Right now when you switch plugins it takes FOREVER to get up and going again because Xojo has to recompile all of the plugins. If there was a way to store those compiled caches that would save a ton of time.
[quote=489760:@Sam Rowlands]This mirrors my experience as well. I was quite dissappointed with the 16". I’d done my homework before hand and was expecting a 1.9x increase in single core and 2.4x increase with multi-core, compared to a 2012 Retina MacBook Pro running macOS 10.14.
[/quote]
I’m still running Mojave on a non retina 2012. I just haven’t seen any compelling reason to upgrade. The way things are going with Apple I keep getting this feeling that my next computer may run linux. My son uses a 2007 iMac Core 2 Duo with Manjaro and it runs much better and with more current software than it does running the latest OS X it can run. At this rate it may last another 3 or 4 years.
I hope Apple gets its act together with its computers, but it may just be that my next Apple “computer” won’t really be a computer. It might be an iPad and my computer will be linux.
Sadly some games (even single player games) are like this and it drives me nuts!
Does this imply that opening projects may take longer?
Or will the IDE still cache what had been opened for a specific project, and if another project uses some of the same plugins, they won’t need to be reopened? So they time saved on loading the IDE will only be partially offset as projects first reference a plugin?
And won’t this mean that classes won’t auto complete until you have added specific plugins to the project? Right now, I load the complete MBS and Einhugur plus others with the IDE launch. And then I can start using any plugin function without carrying for what plugin(s) it needs.
[quote=489760:@Sam Rowlands]There’s this article as well.
https://sigpipe.macromates.com/2020/macos-catalina-slow-by-design/ [/quote]
This is interesting: not infrequently, my app takes a long time to just start up (seconds to even 10-20 sec) seemingly doing nothing: no excessive disk, no excessiveCPU, no excessive network traffic (resolution of aliases to remote files under File > Recents excluded; this is another MAJOR PIA that i have not found a solution to, with that dreaded Server at IP addr cannot be accessed or some such message, after staring at an unchanging screen for a whole minute or 2).
Could this stall-on-launch be what’s reported in that article? High Sierra.
[quote=489820:@Peter Stys]This is interesting: not infrequently, my app takes a long time to just start up (seconds to even 10-20 sec) seemingly doing nothing: no excessive disk, no excessiveCPU, no excessive network traffic (resolution of aliases to remote files under File > Recents excluded; this is another MAJOR PIA that i have not found a solution to, with that dreaded Server at IP addr cannot be accessed or some such message, after staring at an unchanging screen for a whole minute or 2).
Could this stall-on-launch be what’s reported in that article? High Sierra.[/quote]
I don’t know for sure.
Personally I would think that the first thing to do would be see if it’s problem with the launch sequence of the app. Adding a simple “system.debuglog currentmethodname” as the first line in your app.open event, will write to the log (reading it is another challenge) before it does something else. Another another debug message in the code that will fire when your application is ready to use. A timer with a 1 ms delay should do it…
To read the log under Sierra or newer, you must open console and do a search for your application, before you launch the application. Then launch your app, and see.
You can use the terminal to disable GateKeeper (temporarily) and see if that helps. Restarting your Mac will clear the caches and the VM, so I’d also give that a try.
@Peter Stys: the article is about Catalina and not High Sierra. Do an analysis with Instruments. See https://www.mothsoftware.com/blog_page.php?permalink=testing-with-instruments to get started.
It may help if Xojo could manage not just one set of plugin cache files, but allow multiple and switching.
e.g. hash the plugin folder data as they do currently and then make a folder for each hash in the cache order.
If you switch between two plugins folders, don’t delete the other cache, but keep them for a week, so you don’t need to recreate them often.
My final post on the original topic: long startup. Turns out my Samsung T5 SSD died but rose from the dead: while 1.2 TB of data disappeared because the encrypted APFS volume refused to mount, it did reformat without a problem and an extensive R/W test shows that all is well with the hardware it seems. So I don’t know whom to blame for this: Samsung or Apple’s APFS, or the combo.
What does this have to do with the original thread you’re asking? After resurrecting the SSD, whether I formatted as HFS+ +/- encryption, or APFS +/-encryption, Xojo startup with a full set of MBS plugins completes in <30 sec.
So I am at a loss to explain the previous slowdown. The only diff is before, the SSD was formatted as APFS+encr under high Sierra, vs now I moved to Mojave. So it wasn’t the encryption, or HFS vs APFS.
I am still left wondering whether APFS is what pooched the SSD in the first place…Has anyone had issues with APFS? I am reluctant now to use anything but HFS.
I think many people are using APFS with no issues.
The last I heard about APFS, is that it’s not recommend for use on external drives, spinning drives (like “Fusion” drives), something to do with it’s design is for soldered on chips of certain capacities, as opposed to being for all drives like HFS+ is/was.
I don’t think the recent OS updates give you a choice in the matter for the boot drive, whereby if it’s not APFS, they auto convert it to APFS. Many people attribute the recent OS slowdowns to APFS as well.