Certain MM patch not applying #24
-
|
This mm patch does not seem to work: Details@part:HAS[@module[ModuleB9PartSwitch],#&/ModuleB9PartSwitch/baseVolume&>0]:FINAL } All other MM patches seem to function |
Beta Was this translation helpful? Give feedback.
Replies: 49 comments 87 replies
-
|
[scratch that] |
Beta Was this translation helpful? Give feedback.
-
|
Ok. Let's try again. You want to add Well, I can't tell if the thing is working as intended only with the MM log. I need the ConfigCache too, so I can inspect whatever is there to see why this patch didn't worked as intended. Obviously, this patch will not be applied on parts without the For example, the following patches WILL NOT work are intended: Because the first patch will be ran first for sure, when the B9PS patch below wasn't executed yet. |
Beta Was this translation helpful? Give feedback.
-
|
Forgot to mention, the patch is in the My_Patches file, but I have also tried it in a separate file, and when I did that, the file did not show up in "MMPatch.log". When I checked the ConfigSHA file, it was recorded by MM. Here is ConfigCache and ConfigSHA |
Beta Was this translation helpful? Give feedback.
-
|
Regarding order of patches, I have the near future mods which use B9PS. Those have part definitions, so it should apply before the :FINAL right? Those parts have functional B9PS switchers. |
Beta Was this translation helpful? Give feedback.
-
Attach the whole patch file too. I think things are slightly simpler than it looks. If I would write that pach, I would write it as: Case matters. |
Beta Was this translation helpful? Give feedback.
-
|
As well the location of the file. Tell me the full patchname of the patchfile too. |
Beta Was this translation helpful? Give feedback.
-
|
Full patch: |
Beta Was this translation helpful? Give feedback.
-
|
Well, the best way to see what happens is running it on my rig. If course no patches will be applied, as I don't have anything installed on it, but I will be able to at least see them being deleted, what I didn't found on your log. I will be back to you in half an hour give or take. |
Beta Was this translation helpful? Give feedback.
-
|
I will try your updated patch soon |
Beta Was this translation helpful? Give feedback.
-
|
Weird... I installed B9PS on my rig just to see the patching happening and... you appears to be right. MM is silently ignoring the patches mentioning I think there's a small syntax error somewhere on this file. MM is absolutely tragic on handling errors. It's really shameful. |
Beta Was this translation helpful? Give feedback.
-
|
Uhh... WAIT... Why are you using
|
Beta Was this translation helpful? Give feedback.
-
|
The patch looks right (once you use $ instead of &). I had toyed a bit with this on Forum once. However, it is being ignored on my rig too. |
Beta Was this translation helpful? Give feedback.
-
|
I wrote the patch after checking the Sarbian MM wiki, where it said to use &. Must have misinterpreted something. So I should replace & with $? I'll test that when I get home. |
Beta Was this translation helpful? Give feedback.
-
|
Your edited patch, using $ instead of &, is also ignored. B9PS still works. Any idea what the issue is? |
Beta Was this translation helpful? Give feedback.
-
|
well, the MM unit tests tells us differently: See? When the value is on the current context, not even The correct syntax would be: Not even $ is used on And yes, it's confusing and MM doesn't makes thing easier by just ignoring errors like this one. It's the reason I gave up trying to be "smart" and gone brute force, patching part by name, one by one, exactly like I want them to be. Don't fool yourself, if something goes wrong on the user's machine later, you will be as lost as you were now. :/ Attached the fixed file: Fixed_Patches.cfg.txt |
Beta Was this translation helpful? Give feedback.
-
|
Well, I simulated again (for speeding the tests) the target environment using the following synthetic patch: This will add a ModuleB9PartSwitch to every PART in the game, and then will add another one only for I managed to reproduce the user's error reported above. Now I have something to work with. |
Beta Was this translation helpful? Give feedback.
-
|
Is it possible to use MM to search for the largest baseVolume value among all B9PartSwitch modules in a given part? |
Beta Was this translation helpful? Give feedback.
-
|
I think it's possible, but using a hell of a kludge simulating FSM using Right now I'm trying to just fetch the first B9PS module with a |
Beta Was this translation helpful? Give feedback.
-
|
New Synthetic Test Environment. I'm using this to check the patches while I (fail to) build them: |
Beta Was this translation helpful? Give feedback.
-
|
Oukey, finally some results. The following patch will look for all B9PS modules and will use the last one's This IS NOT what you want yet. The next step is to convince this piece of crap to only assign the I think we can simplify a bit this thing, but this is also another fight to yet another day!! 😜 |
Beta Was this translation helpful? Give feedback.
-
|
The worst part is that besides being a FSM (Finite State Machine), the order in which some operations are done are not what you expect. Sometimes something coded later in the patch is executed first, and then you end up chasing your tail pursuing errors where they are not. For example, if I try this: You will get an error because the patch will not find the Sometimes, you also fail to access a value inside a NODE because you are still editing the current node, and any new values are stored on the buffer, while trying to access values will always reach the original still to be replaced by the buffer once the NODE is completed. This is incredibly annoying, because inside an editing NODE you read values from a place and write them into another - an incredible source of tears, blood and teeth grinding. |
Beta Was this translation helpful? Give feedback.
-
|
This sort of makes sense, but it is incredibly convoluted. I guess that shows MM's deficiencies; this would be way easier in a programming language. I was not aware that MM had if statements at all? |
Beta Was this translation helpful? Give feedback.
-
|
I agree. Trying to maintain add'ons based on these stunts on the long run is near suicidal. 😆 And, well... It's not exactly It's again FSM black magic and, so, pretty counterintuitive and cumbersome. There's a reason I avoid complex MM patches on TweakScale like the plague (besides maintaining my own fork of MM!) |
Beta Was this translation helpful? Give feedback.
-
|
Revisited solution. Tried to get rid of renaming the module's name, but got stuck on something that may be a bug on MM, or perhaps another idiosyncrasy due reading values from the original buffer, and writing on the new. I will need to check the MM code to be sure. In a way or another, follows the thing for historical purposes. This version of the patch have exactly the same results of the previous one (i.e., the last B9PS module's |
Beta Was this translation helpful? Give feedback.
-
|
I suggest toying with this thing on a pretty clean KSP installment to save time on loading and testing. We are not testing the KSP or the results of the patching, we are only checking the patching itself. As a matter of fact, a butchered KSP installment with only two parts in the whole system would be enough... |
Beta Was this translation helpful? Give feedback.
-
|
Using
Without something like I'm prone to consider this a MM functional bug (as the uncountable ones on the upstream's bugtrack - I had forked this thing for a reason). But even if I'm right, by fixing it I will deviate from canon behaviour. What good it would come from this? |
Beta Was this translation helpful? Give feedback.
-
(sigh) if this is a bug, someone will need to fix it sooner or later. |
Beta Was this translation helpful? Give feedback.
-
|
@ProgrammerFailure , I released a new Module Manager /L with a shitload of new log messages to screw your filesystem's I/O. In order to activate them, you will need to edit I hope this will help us a bit on the problem at hands. |
Beta Was this translation helpful? Give feedback.
-
|
Here are the log files after editing KSPe.cfg: |
Beta Was this translation helpful? Give feedback.
-
|
I am currently testing what happens with 2 modules, #1 with baseVolume = x and #2 with no baseVolume. |
Beta Was this translation helpful? Give feedback.
-
|
Seems like no baseVolume reads as 0. Feels like a contradiction from before, where not having a baseVolume has no bearing on baseVolume > 0. Edited cfg and the resulting cfgcache: Detailscfgcache: UrlConfig DetailsPART // Attachment // Tech // Info // Parameters } |
Beta Was this translation helpful? Give feedback.
-
|
Bleh. As expected, RealLife™ kicks our ass no matter how smart we think we are. The current draw back of the present solution is that the "last module" wins the election - always. B9PS modules without a It's the reason I still need to find a way to retain only the biggest value - or, at very least, the first non zero value that it's found. Right now, I'm trying to figure out a way to do exactly that - but the operations MM gives me right now appears to have a bug (as using or replace it to some other clever (and incredibly clumsy) trick, we have the problem partially solved. If we manage to have something like this working: Then we have the final solution for the problem. |
Beta Was this translation helpful? Give feedback.
-
|
(need a pause now, still at DayJob©) |
Beta Was this translation helpful? Give feedback.
-
|
You said that the problem is that: The attribution on the inner patch tries to get the baseVolume from the first ModuleB9PartSwitch module - and then borks because exactly this one doesn't have the damned value. So if I set the :HAS clause to search only the first ModuleB9PartSwitch for baseVolume, then the error would be gone correct? Just not sure if it would still apply to everything it applies to right now. |
Beta Was this translation helpful? Give feedback.
-
I tried that, but the thing didn't worked. However, I may be biased to "my solution" and could not break free from it. Try it yourself, let's see if you can find a way different from mine. In time, I had a situation on DayJob©, I'm still writing reports. As soon as I figure out what happened there, I will come back to this. |
Beta Was this translation helpful? Give feedback.
-
|
Hi! Did you had any different outcome from this? I (hopefully) will be able to tinker with it again this weekend. Would be good to have something new to work with! |
Beta Was this translation helpful? Give feedback.
-
|
I haven't tested it yet; I messed around with my computer settings a little too much and had to trouble shoot that. I'll be testing today or tomorrow. |
Beta Was this translation helpful? Give feedback.
-
|
Well, you got luckier than me. I had messed up a whole server this week!! 😝 |
Beta Was this translation helpful? Give feedback.
-
|
Oukey, back to basics. The following patch: Is not working on this step: Because the Now I need to figure out if I'm using this Evidences: MM Patch Log: MMPatch.log noteThis log was made using the current dev branch, the code is not released yet. |
Beta Was this translation helpful? Give feedback.
-
|
KRAPPP. There's no index for
|
Beta Was this translation helpful? Give feedback.
-
|
I'm feeling extremely dumb now, |
Beta Was this translation helpful? Give feedback.
-
|
I'm working on a solution. Managed to implement part of it, but the week end ended too soon! |
Beta Was this translation helpful? Give feedback.
-
|
@ProgrammerFailure , I managed to simplify the patch (now only one However... I think that this patch should be applying the last >0 Evidence: excerpt from ConfigCache (terribly verbose, that Excerpts from the MM Patch Log: |
Beta Was this translation helpful? Give feedback.
-
The index is not on the
Yep. But it's a starting point, I documented it here to avoid forgetting it. The good thing is that you don't mind from where the biggest value came, all that matters is the value itself. So populating a temporary data storage with them and then iterating on the data is enough - essentially the I think I have a very crazy (or very stupid, still to determine) to solve this but it will need nested |
Beta Was this translation helpful? Give feedback.
-
|
Well, testing failed... because my computer crashed. Windows fixed itself, thankfully. I get off early tommorow, so hopefully I will have time to test/play properly. Is this correct?: DetailsAlso, presumably the MYDataStorage section would be omitted, and ia and ib would be the baseVolume values from different B9PartSwitch modules. |
Beta Was this translation helpful? Give feedback.
-
|
My last exercise was.. well... an exercise. I needed to find a way to solve the "bigger" problem for two numbers without creating extensions for MM (what I was kinda coding already, but then I realized it would not solve your problem, only mine). You are right, What I'm planning to do is try to use arrays to solve the problem, something like this: Then interate on This will left |
Beta Was this translation helpful? Give feedback.
-
|
I finally completed a pet project of mine; syncing all ksp data across two pc's. Now I can actually use my newer, halfway decent gaming pc for testing stuff. |
Beta Was this translation helpful? Give feedback.
-
|
I will have some time to burn this (local) holidays too. Some planning I had for the company's infra gone South as we will have clients operating on the holidays too. 😞 Oh, well... Now I have time to play. 😆 |
Beta Was this translation helpful? Give feedback.
-
|
Creating another "thread" for a proposal of a MM patch to find the biggest value of a list of values. Now that we established that looking for the biggest of two values is feasible, I want to extend the solution to many ones. The initial setup is The idea is to iterate on Caveats:
Working on this now. |
Beta Was this translation helpful? Give feedback.
-
|
I think this is roughly what the patch will look like, correct me if I'm wrong: and we will require a filter for how many "ModuleB9PartSwitch" modules there are. I am unsure of the syntax for deletion so I put in my best guess. I abridged the @module[WBIConvertibleStorage] for clarity. Correction: |
Beta Was this translation helpful? Give feedback.
-
|
Ok, used the following patch: Results: All parts got a ModuleB9PartSwitch as per the PART[*] ConfigCache was not generated, because of the following error: Best guess, I messed up the syntax for accessing values in a list; I couldn't find the syntax for accessing an indexed value in a list, where the list itself is referred to via index. this is the config node for trussAdapter, obtained via partInfo in PAW mod. Seems like it's down to a small syntax issue. |
Beta Was this translation helpful? Give feedback.
-
|
Possible solution would be having multiple TEMPSTORAGE modules, each with its own "v =" section, and decrementing each v. This gets around having nested indices. Something like this: However, I am unsure how to stop decrementing when there is only 1 Which is a necessary step to finding the winning v,2 basically we need |
Beta Was this translation helpful? Give feedback.
-
|
I will come back to this Soon™ - I'm having a hell of a week at DayJon©! |
Beta Was this translation helpful? Give feedback.
-
|
Oukey... My initial attempt of sorting 2 numbers was plenty of syntax errors. The correct one follows: And it works, as the ConfigCache tells me: Now proceeding to the multiple (3 or more) values stunt. |
Beta Was this translation helpful? Give feedback.
-
|
This is my initial proof of concept for getting the biggest of an arbitrary quantity of numbers: it just decrements every first index 10 times. It works: |
Beta Was this translation helpful? Give feedback.
-
|
KRAPP!!! I keep forgetting... There's no way to use indexes on |
Beta Was this translation helpful? Give feedback.
-
|
Ok, this line of research is kaput. Trying something else. |
Beta Was this translation helpful? Give feedback.
-
|
Oukey, found a way! And it ended up being simpler than @ProgrammerFailure attempt. and it gives me: |
Beta Was this translation helpful? Give feedback.
-
|
Yep, it's confusing and this time is my fault. I answered that on a hurry (still DayJob©). On my original patch, you will find: See how there's an "inner" try: |
Beta Was this translation helpful? Give feedback.
-
|
Ok, this patch: With this test environment: Generates this output: The important part being: Which suggests that MM_PATCH_LOOP still isn't working. I couldn't find the documentation for MM_PATCH_LOOP, so presumably the syntax is incorrect? |
Beta Was this translation helpful? Give feedback.
-
|
Another unforeseen issue: if any of the first three ModuleB9PartSwitch doesnt have a baseVolume, it throws an error. Can be rectified by |
Beta Was this translation helpful? Give feedback.
-
|
This is getting terribly confusing. You are fighting more than one front on this battle. I suggest you forget about B9PS and WBI for while and focus on the problem at hands. For starters, I don't know B9PS neither WBI, so I just can't tell when something wrong happens on a patch involving these two things. I suggest we make a prototype for the solution using a synthetic, simpler module dataset and make the solution work there. THEN, once we managed to make the damned thing to work, you can adapt it to B9PS and WBI. |
Beta Was this translation helpful? Give feedback.
-
|
For while, this is what I'm trying to do: The problem is the Phase 2, where the LOOP just don't behave as I expect. Apparently, the inner node editing is not being fed with the "HASLIST" (the list of data that I suspect I need to check the source code to see how the "HASLIST" is being transferred to the inner nodes being patched. The way the thing is behaving is hinting me that this list is only available to the "main node" patches, not to the inner node ones. |
Beta Was this translation helpful? Give feedback.
-
|
I found a way of selecting the biggest of a series of numbers, what satisfies the initial problem. From where the numbers came, what to do with the biggest one and how to do the needed house cleaning after is left as an exercise to the reader. In a final thought, this doesn't worth it. We spent weeks bashing our sorry arses on a problem that could be solved quickly using an external script building "dumb" patches for an specific add'on. I don't want to even think on how many man/hours were spent on this stunt, neither what I could do with them instead. Complexity piles up quickly, you are going to get an unmaintainable zeitgeist way sooner than you expect. Been there, done that. I gone hard coded way with https://github.com/TweakScale/Companion/ for a reason! |
Beta Was this translation helpful? Give feedback.
-
|
In time, this is the exact contents of a directory called "Hacking" where all my working prototypes were tested. I think you are misunderstading something about where to shove these things. These patches are working for sure. I posted evidences above. |
Beta Was this translation helpful? Give feedback.
-
You are on linux, I'm right? How are you starting KSP? On MacOS, I use: Without the
Yes, this thing reloads everything, including the DLLs - what a lot of Assemblies of mine (and I think some others too) can't withhold and then causes the problem you are facing. I did it a long time ago (4 years, I think), when my knowledge of both KSP and MM intrinsics wheren't at their best. The commit I linked above wasn't published yet. I will publish a new MM release with this fix sometime this week. |
Beta Was this translation helpful? Give feedback.
-
|
My variable search was in fact in error, works now, but MM_PATCH_LOOP still just shows up as a subnode: I'm going to try reinstalling MM. This feels like some small issue that I am missing. What if MM acted exclusively on a cache, such that it read files from the filesystem? Anyway, I don't know much about KSP internals, or KSP in general. |
Beta Was this translation helpful? Give feedback.
-
|
Oh also, I am on windows. I put the -mm-dump into the KSP debug menu after opening the game and hitting Alt-F12 |
Beta Was this translation helpful? Give feedback.
-
|
That's explain the problem. You need to use But I'm going to publish an Alt-F12 with a working dump option soon. Perhaps even today, depending on how my day is going to be. |
Beta Was this translation helpful? Give feedback.
-
|
@ProgrammerFailure , I think this release will interest you: https://github.com/KSP-ModularManagement/ModuleManager/releases/tag/RELEASE%2F4.2.3.4 |
Beta Was this translation helpful? Give feedback.
-
|
I completely forgot about this. Sorry. I have some unrelated issues to carry on this WeekEnd, but I will do my best to come back to this ASAP. |
Beta Was this translation helpful? Give feedback.
Oukey, found a way! And it ended up being simpler than @ProgrammerFailure attempt.
and it gives me: