WinXP - Win 8 --- something I need to know ?!?

[tester]

This thread is too long to read, but I see you are aving some issues with saving/loading prims?

Few tips.

1) When using "afterload" primitive to load something - add whatever delay/timer to afterload primitive. It had to do with order of loading fhe filesystem versus everything else in the compiler.
2) Trigger order. Always keep the trigger order, especially when writing files, where "confirmation" trigger is required before loading something that is being saved. If you use a trigger from string directly (because you can), then it is sent just after the filepath is sent. If you use "done" trigger from primitive - it is sent after the job is done. My guess is - this may be needed when delays in operations occur, or larger text files are processed.
3) Relative paths. A real headache. Sometimes triggers are sent sometimes not. But in the past, with Trog - we noticed, that using "start folder" primitive may help. But this prim behaves differently in schematic and differently in exported app. In exported app - it follows the path where app is, while in schematic - it follows the path, where the exported app will be. It has input for path, but I'm not sure whether this is used only in schematic or in exported app too; you may need to play wit this little fellow, to find out. But after I started using it it - everything works fine here (file operations and triggers match).
4) Rewiring the environment around save/load and other primitives, so that you are sure, that everything is as it should be. Recently I posted "extension manager" examples, where I did few stupid lookong things, but these things - finally made the prim work correctly in all cases. So play with triggers in order to force the check on saving/loading routines. In a few days I may post something on it, since this is also somewhere around me right now.
5) Ruby - don't use it if you don't need to. Take care of the triggers around ins and outs, when using it.
6) Using the extension manager I posted somewhere around - may be also useful.
7) Measure the amount of triggers in crucial points.

//edit: avoid ambiguous trigger situations. There are scenarios in which trigger flow seems to be normal, until you start seriously using the app. Then the events start to go out of phase with each other, and triggers start to miss correct order. When it happens? When you combine independent flows of triggers.

[RJHollins]

All excellent tips !

Yet ... I've hooked a direct trigger to a text load routine and it could still fail to load the file.

The PreSet Manager design has been here for a long time ... why would this now fail to read its own file ??

REAPER, a very popular DAW is no longer able to have a VST install on the Master Track and load in a user config file, nor the PreSet Manager file. This specific issue is with XP through 8.1

I can easily accept that I've done error in my schematic. But even special test files written by experienced programmers exhibit the same problem.

Very few seem troubled by any of this ... I don't understand why !

Is it a BUG ? ... bad programming ? .... where ?!?!

I would love for this topic to be closed with a solution. or an answer.

[Drnkhobo]

Thanks Tester for the heap of tips there. Really appreciated!

I have a few gripes like Mr Hollins in that why did the same schematic work in previous FS versions & not in the latest. I think I am going to revert back and try some tests with the old version.

I am emailing support about this. . .

[Drnkhobo]

Hows this for weird. . . .

user loads plugin and registers. Plugin saves a file to disc. Triggers sent to load file, successful and unlocks plugin. Then if you load up another instance in the same project, the plugin is locked! wtf, if the file was written and loaded upon registering AND unlocked cause the file was read, then WHY does it not load the file when you restart the plug?

Weird. . .


Hey tester could you explain more on. . .

avoid ambiguous trigger situations. There are scenarios in which trigger flow seems to be normal, until you start seriously using the app. Then the events start to go out of phase with each other, and triggers start to miss correct order. When it happens? When you combine independent flows of triggers.

Independent flows of triggers?

[tester]

Hows this for weird. . . .

user loads plugin and registers. Plugin saves a file to disc. Triggers sent to load file, successful and unlocks plugin. Then if you load up another instance in the same project, the plugin is locked! wtf, if the file was written and loaded upon registering AND unlocked cause the file was read, then WHY does it not load the file when you restart the plug?

Weird. . .

Not necessarily. There are few things coming to my mind. There are two parts in this process. "Loading the file" and "loading the content". One of them seems to be a crap.

For example - I would add an empty keyfile, that is being replaced by correct one during the registration process. Thus - your app would load either empty or correct one, but always something. Then I would indicate in options (file or internal) to search for keyfile, so that when it's not there - "demo" notice should come in also. Your second instance in the same project - is probably not refreshing something.

Second - I would rearange the way of looking whether the key file is present, and loading empty string if it's not there. Some sort of refreshing using "last" or "select" switch and trigger order (to refresh the content). You could use "find files" prim, to knw whether (empty or not) key file is present.

Hey tester could you explain more on. . .

avoid ambiguous trigger situations. There are scenarios in which trigger flow seems to be normal, until you start seriously using the app. Then the events start to go out of phase with each other, and triggers start to miss correct order. When it happens? When you combine independent flows of triggers.

Independent flows of triggers?

In one of my projects I have an unsolved situation like that. I cheated to make it work, and it works "almost" perfect. I have a loop, in which green values/triggers are circulating (refreshing all "last" switches), and undoredo system. I have no way to inform the undoredo system whether certain trigger was sent or not, but I must trigger the undoredo system anyway, in certain section of the loop. So in order to make this work - I added a time gate (the only "bullshit" scenario with trigger separation, that i would agree with MyCo), to pass through the triggers from the main loop. It works. But... It works only for 1 unit. When the project operates on multiple clones - trigger order of the loop, and timing of the gate get out of phase, and in some clones undoredo works, and in other - don't (different order). Independent trigger flows.

Sometimes, it may happen, that your measurement of an event - should be before something else, but becomes after that, and you almost have no way of knowing that. That's why I would use some tools like trogz tools, to monitor the trigger order.

[Drnkhobo]

Sorry tester, I was ambiguous with my explanation but I do have the plugin load a empty file on startup. It is then in demo mode. Once the file gets overwritten its unlocked the plugin.

Thanks for the info I will go find trogz tolls now

Hey I just thought, why dont you post ur schm for saving reading files and we can all see if it works independently of systems???

[tester]

I was just going to ask you the same thing. Could you extract the routine (only the routine) you are using? (better post it in separate topic, that will focus just on that). That would help to find out the issue, and maybe something will clear out earlier.

As for my schematic - well - I use various parts of what I wrote in various schematics, and some concepts still wait to wire up, but I see what I can do.

[tester]

BTW, meanwhile there are two things you could do.
1) Check inside txt whether the file is really written before load (save in one instance, check the text file from inside, load second instance). It may happen, that ulnocking goes internally (again - trigger order), and saving somewhere later or due to some second retrig.
2) Add some button/area to gui, to retrig manual loading of the licence file; if it works - it will give you the idea of some missing trigger or wrong trigger order. If it isn't working - it will give you some idea on the "location" of that second instance in relation to original filepath (maybe something else happens in win8 than in other wins?).

[Drnkhobo]

Thanks tester, I will look into it.

Got a really nice support email from the higher powers so hopefully things will get sorted quick. Will keep you posted.

[RJHollins]

Thanks tester, I will look into it.

Got a really nice support email from the higher powers so hopefully things will get sorted quick. Will keep you posted.

Thank-You Drnkhobo !

I'll hold comments on 'testers' suggestions till then.