Plugins Broken Via Version Update

Thank you Rubenator and Taint (what I was trying to say - cellphone typing for the win). I didn’t realize that about the // in my init.txt. I will modify the file and this should solve my issue.

However, it did work last week and for the years that I used it before, now after the update it no longer works so has something to do with the issue and is just another data point for reference in the analysis.

That worked great ty!

There's no need.

If you, today, want the ability to enable/disable windower auto-updating on launch, you can do that with windows firewall. Not permanently, obviously; You will need to update eventually, but it's super easy to toggle that rule on or off as you see fit so you can update according to the schedule you want, without changing the default behavior for everyone else or adding yet another click in a login process that has plenty already.

A config setting or launch flag would make it slightly more convenient but have exactly the same result.


Nope. I temporarily blocked all outbound on mine so I could do group stuff with working binds last night; It works just fine with no hanging or timeouts.

I'll get to your suggestions below, first I want to expand on our situation a bit. Our core goals of working on Windower are the following:
1. Keep the end user experience high
2. Keep our workload low

We failed the first one yesterday, and the reason for that was inadequate testing. Unfortunately, increasing the test load by ourselves is simply not an option, since it directly goes against goal #2. We put in pretty decent test work on this, as all members of our team (as well as several other people) were using the newest version of Windower, with all its bugs, for months. Most of them weren't noticed, because they are all particular to using certain commands or binds, which weren't part of our usual play style.

So the question is how to increase the testing capacity without further inconveniencing ourselves. Our solution to this was to introduce a Windower dev branch, which people could opt into and experience newer features as soon as they were released. This way, bugs could be discovered sooner, and fixed, before an update is released to the masses. And if a bug was so bad that it disrupts people's experience, they could switch back to the stable branch, and enjoy the stable experience while we work in comfort to fix the dev branch.

This is exactly what we did, and the reason all these bugs weren't discovered sooner is because simply not enough people are using the dev branch. There are a couple dozen people who do, and only one of them has reported only one single issue. That's why we assumed the current code to be stable, and pushed the updates out to everyone. And then the reports started coming in.

The problem is not as easily solved as just telling people to use the dev branch. We've been telling that to people for years, and many won't do that, and we can't blame them, because the current experience on the dev branch is not very user friendly. I mentioned before, that you can just switch back to the stable branch, but that's not easily done, due to the way it's currently set up. You either need to manually replace many files, which is annoying, or have a completely different folder for each branch, with a copy of all your settings, and every time you adjust something in one, you also need to adjust it on the other.

We want to make that experience better, and we have some ideas, but it will take a few weeks to implement, at least. Ideally, switching between the dev and stable branch should just be a button click inside the launcher. So many people can play on dev, and switch back easily at any time, if we introduce a bug that's preventing them from playing the game properly. If we release a bad update, like yesterday, we could even remotely flag the dev branch, so everyone who launches Windower who was on the dev branch, automatically starts on the stable branch instead, unless they override it.

Another idea is to make rollbacks possible. This works to a limited degree even now, but it's a lot of manual effort, which violates goal #2. This is mostly due to how our system is set up, and can be fixed. We actually just got a good suggestion for implementing rollbacks on our Discord server, and I'll try to implement that tomorrow or the day after. Unfortunately it won't help us with the current situation, although the worst of it has been resolved, but would help with goal #1 a lot in the future.

These are just some ideas that I'd like to see implemented soon, and we welcome any other ideas that keep both of our goals in mind. And with that, let me address your suggestion.

The reason I don't like a prompt is the same reason we forced updates in the first place - we don't want people on older versions, and some people would categorically decline updates, because they like "everything as it is", and because they've had bad experiences with updates in the past, like yesterday. This will ultimately result in a lot of work for us. You mention in your second possible con, that people will complain if their stuff is outdated, and our response to that can just be "update" and close the ticket. I really, really wish it was that simple. This is not a hypothetical, I've been around since before we had automatic updates, and the heavy burden customer support for outdated files placed on the dev team was the very reason unskippable updates were introduced, and why we have, several times, declined to implement an option to skip updates. The issues ranged from even identifying that someone was using an outdated version ("I assure you, I am on the latest!") to figuring out if an issue was even caused by an update, or a real issue that's been around, but unnoticed, or maybe just due to their local setup, etc. It's almost never a simple situation that we can just close and move on. This is *still* affecting us today, many of the issues we have to investigate and put work into today are because an update simply failed.

That said, let me address the three reasons you mentioned for this feature, because I think all of them are valid. The first two both boil down to not interrupting actively logged in instances. This can be fixed on our end by making launcher detect running instances, and if it finds some, skip the update step. This could be combined with a message upon login to notify people they are out of date. That way, they would only update the next time they log in all of their characters.

Your last point is understandable, but misguided. We really don't want any unwilling guines pigs, even just for a selfish reason, because unwilling guinea pigs will complain, which puts more burden on us, see goal #2. We specifically set up the dev branch for *willing* guinea pigs, and the plan is only to release updates once we're confident that things run smoothly. As I mentioned before, the only problem was, that we didn't have enough guinea pigs this time. I also mentioned what we plan to do to address this.

I welcome all suggestions that keep both our goals in mind. While goal #2 may sound selfish (and it definitely is), we are not young anymore. I used to work on Windower for ten hours a day back in my glorious youth, and I wish I had the option to still do that, because, despite everything, I enjoy working on it. But life goes on, everyone on the dev team is an adult now, with annoying adult responsibilities, we simply cannot afford to put much work into it anymore. The big change that broke stuff this update has been implemented over two months ago. We haven't published it to the stable branch sooner, because we wanted to make sure that it's safe, so we let it live on the dev branch for that long, to give our few testers all the more time to test. Just to avoid a situation like yesterday, that would involve everyone panicking and us having to divert attention from our busy lives to this.

One final thing, about the firewall - this is not a solution, unless you play on a client-locked private server. It simply means your game will break the next time resources are updated, which is very likely the next monthly FFXI update.

@Lews
One of the changes that was made was refactoring the command processor to be less spaghetti.
So please lose your own command spaghetti and we'll go from there?

While it technically worked it was never a proper or intended way to deal with it

So i kinda half figure out the failure to paste

It's triggering on release of V, but it has to be held long enough to work.

That's one of the things fixed, right? It triggering on release instead of press. The whole setkey up and down reversal etc.

Per the Windower discord #announcements, all keybind issues should be fixed aside from certain bind commands for backslash `\`, which we are debating about internally
edit: oh and semicolon `;` binds

This is the first I'm hearing about "The whole setkey up and down reversal etc." -- no one has come to the discord to discuss such a thing

I appreciate the time you took to write that out Arcon. I understand your perspective on the matter, but I still dont comprehend the reasoning, hear me out.
You say you want to ensure that everyones on latest and greatest to reduce the amount of bug reports from people using outdated version, fine I get that. But then you go on to say
Just based on this, it seems highly unlikely someone would be able to play the game on windower while declining the updates forever. This is very reason why I dont immediately log in after maintenance (that and its 1:30am and I likely have work the next day), I know windower files are gonna be changed and need updating. If I try to log in the next morning and I dont get the file update spam, I'll hold off on logging in until the afternoon. Its pointless knowing I'd have to relog all my chars, and logging into this game is a PITA.

Honestly this would be a huge boon.

If you can't be assed to relog your accounts more than once a month, maybe twice if you jump the gun, that seems like a you problem. 99% of the time you can simply open a launcher after update day to let windower update itself while already logged in on your actual accounts. They goof once with a large update and all of a sudden you insist on an OK button to push new updates? Seems like extra work just to please your individual situation, which is the very thing they are trying to avoid, as was stated.

EDIT: Nevermind. Can disregard this post from here.
(I misunderstood what the percentage meant, it was the cumulative chance of opening the chest, not the chance the next guess would be correct.)

Pretty minor but I tried out the boxdestroyer addon but the displayed success percentage for the next guess shows (100%) a lot of the time when it isn't. It seems like the logic for which number to guess next is good though as it's helped me open a lot of boxes correctly in a way that seems to make sense.

Fwiw it used to show the next percentage and was changed to show final percentage a few years ago

super common misconception, so you are in good company :D

we have discussed how to word it better but nothing makes it "clearer" without looking like a silly legal statement that becomes too long

While I said it would "break" their game, the truth is, most of the time it's only about resource issues, which may not even noticed until way into an update cycle, or even the next one. Just a while ago we had someone complain about an issue, and it turned out their resources hadn't correctly updated for several months. But they only noticed the issue once they obtained a newly released item. This was an isolated incident, but the more people don't update their files, the more this would occur. And

And I think you underestimate how many people don't want to update. This is partly on us, as we've made the experience pretty bad in the past. Regardless, people definitely try not to update, even across game updates. And that would result in avoidable issues. There used to be a simple workaround way to skip updates in the past, and people exploited that, and continued to do so through game updates, and then came to us to complain. Again, this is not hypothetical, we have a pretty strong stance on this issue, due to past events.

That said, I think the idea to not update while active instances are running seems like it would fix the main issue. Thinking about it some more, I think we might be able to implement a feature that allows people to only update when fixes are available, and not for feature updates. Obviously we can't fully separate the two in one file, but we could probably implement a mechanism, that only downloads new files when a fix was reported. This wouldn't be trivial to implement, but might be worth the effort. I'll think about this some more over the weekend.

It's a passion project. This is obvious to everyone but you. Literally a free service that people do for no money so that a 20 year old game is more playable for the majority of the player base. These people are already being more than generous with their time.

Again if you don't like the features of Windower, you have the option to play without it...These people owe you nothing, the opposite is not true and your lack of humility is why people consider stop doing passion projects.

This isn't a good mindset to have, and I doubt anyone working on it wants you to think of it that way either. Windower has become a very important part of the game, and it was made to and kept to a high standard.

Everyone appreciates the work that is done on it and understands the time constraints, but you can complain or ask for stuff without it intending to be disrespectful. Doesn't mean you'll get it but people can still ask.

Free is irrelevant to most people, good is what matters and that's why people use it.

I'm not Windower staff, but I'm going to respond to this the same way I respond to all of the people asking to pay me to write an addon or donate because of my efforts.

I work on this stuff because I enjoy it and I want to bring happiness to people. I don't want to take on an obligation to do so, that would undermine my own motivations. I'm not going to get rich off of what people will pay for it voluntarily, and I don't want to deny the average user the opportunity to use things for free by forcing a payment.

I can't speak as to whether Windower people feel the same, but there is a gigantic valley between providing something for free and being paid enough to prioritize that thing. It's not easy to navigate that area, and you should be glad they're still willing to provide updates.

So they dont want feedback and suggestions to make their pet project better, gotcha


Yes, I'm fully aware without Windower FFXI subs would plummet.

Thats cap.
There are paid bots in the wild and the popular ones are the free ones built upon eliteapi, and they arent that great.

I think plummet might be bit of a exaggeration. You might not like it but there are alternatives to Windower, being Ashita.

Your feedback was wanting control over when you update. It was received, favorably I might add, and yet here you are still whining. What more exactly do you want?

No it wasnt.
A fallback solution was offered for multiboxers to detect if other instances of POL are open because the update process will fail if files are locked. And by fail I dont mean "ugh I have to relog everyone", I mean files are locked by the PC because they are in use by other instances and the update will fail.
I had one char disconnect overnight. I tried to log it back in, opened windower, downloaded the updates just to watch them not install, then the client crashed.



It would continue to crash on load until I logged everyone out. Once everyone was out, I could open windower, but nothing redownloaded so I wasnt confident that it would run properly. I did the whole "delete resources and updates folders" to force an update and be safe.


And I can already tell your response is gonna be some stupid ***like "aw boo hoo you have to relog your chars, this is their pet project, dont use it if you dont like it" just like Iocus dumbass comment. I dont want to put words in the Dev's mouths, but I'd like to think they want to avoid situations where the updates fail based on this:

I saw some addons updated. After logging in, my issues remain. I'm guessing no one else has reported the problem I have? Is there any additional information I can provide that might help? I'm on windows 10 for what it's worth. The crash log shows nothing, which doesn't surprise me as nothing appears to be crashing. I can't get any information from the console as it no longer displays.

What issue would that be? For faster troubleshooting, going to the Windower Discord might be an idea.

You are not the only to report the console not opening, assuming you are not the person who came to the discord to discuss that yesterday.

As for the addons not displaying there was an addon dev who noticed that as well and pointed out that adding a library to the addon (bluguide in his case) got the display working again. So we are still looking into that.

^ this is always great advice we got a lot of helpful people and there is almost always someone around to troubleshoot issues.

Since the update, any addons that display anything outside the chat log don't display anything at all. Also, the console text does not display. So while say Gearswap and Findall work perfectly, TParty and Pointwatch don't appear on screen. Console commands appear to work, but there's no visual feedback as the console text does not. TParty and Pointwatch appear to load and run just fine, no crashes, but nothing appears on screen, rendering them useless.

This is all the result of one update that didn't quite go as planned as far as I can tell. Updates in the past have been pretty seamless from my experience. SE breaks ***all the time on the client and we just laugh it off as spaghetti code. All I'm saying is maybe check the tone of your feedback. You can give constructive criticisms without being an *** about it.

Thank you for letting me know.

That's strange and sounds frustrating. I assume you mean stuff that uses the texts elements (aside of the console window). I have several addons loaded post update and displays fine.

Please point out where I'm being an *** here. Educate me on not being an ***.

Nynja's post was fine, we were not attempting to ignore him, which it might have looked like based on all the replies we made to other posts in the thread after his.

He made valid arguments that we needed to stew on for a bit, we do not take any offense to the way he communicated his suggestions.