Survive the Wild late March update

We’re getting closer, and closer…
Hello everyone! It’s been about 2 weeks since the last update, so it’s again time for the accounting of all that’s been done since I last posted in regards to Survive the Wild development. We’ve made good progress! A few things came up and then were resolved, some items in the todo list were completed, and as always random bugs have continued getting solved while tiny features and changes were added here and there. Again this post may get technical for some of you as I wish to describe everything in as much detail as anyone who is interested in such information may want, and so I apologize if the post gets a little wordy for a subsection of readers. Anyway without further ado, lets get into it!

The previously mentioned admin UI panel is now usable

For a couple of posts now, there has been some mention of a giant user interface panel for admins, the one that lets them modify character data instead of using chat commands to do it. I’m happy to say that this panel is now functional! Admins have been spending time going through it and finding the bugs in it, which I have then been fixing. I have a few controls to add, but they won’t take very long. For example jailing players is still done with a chat command, we want to make a jail function in the character and account panels. I’d say this panel is pretty much 95+% complete, on top of being increasingly already well tested and perfectly usable at this point! At the time of writing, I can only think of 4 additional controls that I’d like to add to it, and I expect it to be an easy thing to do. The controls I have left are the jail control, the warn control, a control that lets an admin modify a character’s quest data, and a non-critical control that would let admins modify some miscellaneous character variables that are usually of little consequence. If you’re wondering how such important controls as warning and jailing got saved for last, it’s because when I was figuring out what all functions the dialog needed, I was looking at the list of variables that make up a character and deciding whether a control needed to exist for that variable. Jailings don’t actually save directly in the character data, instead the jail information is tracked elsewhere. Similarly warnings also did not used to save in the character data, but instead account history. In short, up to this point I was working on the controls that directly modified the data of a given character. Now, I need to ad the couple of controls that perform an action on a character but without actually modifying that character’s data. The quests modification control didn’t get added yet only because I’m considering changing some internal structuring for quests, and will add this control either after this minor restructuring is complete or I’m sure I want to save it for later.

Support tickets are functional again, now with categories

Well, even after all the Angelscript fixes (described below), I still had to dirty my codebase a bit to get them working, at least temporarily. However, I was able to get support tickets functioning perfectly within the last few days! After that, I expanded them to allow tickets of different types as talked about in previous posts. Now bug reports and suggestions are simply a type of ticket, so we can actually ask you for more info on your bug or idea in the same thread now if we need! Further more, the bugs and suggestions are now accessible from within a sorted menu in the game rather than a txt file where they were previously being written to. Bug reports and suggestions can be marked as critical, which will thrust them up to the top of the queue so that they will be attended to quickly, and they can be viewed by scripters and worldmakers as well so that they can implement the content that players want to see first! Overall, I expect these developments in support tickets to be a massive leap forward in player to staff communication. To anyone who has ever submitted bugs and suggestions that never received implementation/fixing or even so much as a response prior to the introduction of these systems, we seriously apologize and assure you that do to these tickets, future efforts to contact the game staff in this manner will not ever be in vein again!

The remote user interface system is now shared with worldmakers

To everyone who has ever wished that actual objects in the game could have their own user interface panels when appropriate, I’m very glad to say that this wish is now granted! While I was modifying the remote dialog layer to include extra functionality for the character administration panel, I had a sudden and random epiphany that unexpectedly gave me the ability to share the entire remote menu and dialog classes with the worldmaker’s scripting module in only an hour or 2 of work! After I’d done this I of course wanted to see it working not to mention worldmakers needed an example of how to create a dialog, so this caused me to spend a couple more hours rewriting vending machines in the shared scripting module which luckily a worldmaker had already started on. In the end, it all worked great! Now if you use a vending machine with an empty hand, you’re presented with an actual dialog with a list of items to purchase, a balance read only text field, a return currency button etc. No more having to press random letters in a makeshift and awkward key hooking screen to purchase an item from a machine! I’m sure you’ll see more such panels appear in objects that require or could use them in the future. Though we haven’t had a situation to test this yet, worldmakers can now also create entire submenus anywhere in the options menu list. So for example, a worldmaker can add a “configure extra settings” item to the arena menu for a host, which would bring up a configuration dialog that the worldmaker also coded when clicked on, or a submenu of configuration options etc. In short, worldmakers can now create user interfaces in the game for any reason that are just as complicated if not more so than that massive admin UI panel I’ve been talking about. I’m excited to see what they do with it!

Three fixed Angelscript bugs

In short, I’m happy to report that the Angelscript developer has been very responsive and helpful, having so far fixed pretty much everything I’ve reported to him! I still have a bug I need to report to him, but at this point, at least in our situation, it’s more minor (though I must not forget to tell him about it before the bug slips my mind because of it not effecting me). This section gets a bit long, technical and contains tangents, so the tldr of it is that even though there were some difficulties including one so bad that it caused a halt in development for around 3 days, the server is no longer being inhibited by Angelscript bugs! If you do not care about the details of this, feel free to skip to the next heading.

Anyway, the first bug was the most critical and was what stopped me from even compiling the client or server using the latest Angelscript, because the compiled bytecode wouldn’t load. The second one was more minor, having to do with shared property accessors. I was having to hack around this bug for a few months, and now no longer need to do so. The third bug? Oof that was yet another that had to do with property accessors in a much more brutal way, which really did hit me like an unexpected sock to the gut.

Though I immediately recompiled the client with the fixed bytecode, it was a few days between when that bug got fixed and when I tried updating the server with the latest Angelscript. This is because I was actively working on the character administration panel at the time, and planned to update the server once that panel was ready for testing. So after days of writing this huge dialog and debugging for the Angelscript developer (something I’m still kind of new at) and getting excited to test out not only the character panel, but also the newly shared remote dialogs and more, the day came where I was ready to update the server. So I rebuilt the latest Angelscript for Linux, ran several tests with smaller scripts that all succeeded, and finally tried updating the server with the new Angelscript.

Can you guess what happened? The server update failed miserably! Now if I tried launching the server, it simply exited with an assert failure being generated by the latest Angelscript. It seems that some time between the January 2022 version I was using and the latest Angelscript with the fixed bytecode, a new bug had been introduced with Angelscript involving compound property accessor assignments to an object by handle. This meant that as soon as the server tried compiling all the code the worldmakers wrote from a shared Angelscript module, the server just crashed with an error message that I immediately knew was going to take me yet another couple of hours to debug enough to contact the Angelscript developer about, who was then going to have to take an unknown amount of time to continue debugging and then fixing himself.

I’m going to be completely honest with you guys, this discovery just completely wiped me out for the next 2 to 3 days. I’d been so excited to try this new Angelscript on the server with the dialog panel I’d spent days coding, and pretty much any extra delay in this update development process stresses me out just in general of which we’ve had many. Over the last months I’ve had to learn 2 different debuggers (one on windows and one on Linux), for that matter I had to learn Linux enough to build NVGT with it, I’ve had to learn how to use so many random libraries I lost count, and have had to solve many random hangs and crashes with next to no debug info, some of them not triggered by me but instead by a scripter in the game who is only trying to do their best in adding new content or in this case by Angelscript bugs. Right now, I know that every single unexpected issue that pops up amounts to even more time that my players will remain disappointed about the lack of a release, making the issues many times worse. This random error that I didn’t cause popping up out of no where simply turned out to be the last straw for my spirit at that point, so that for a couple days, my will to do pretty much anything was more or less completely depleted. My production was therefor entirely halted between last Monday and last Wednesday, and the beta server was down during this time.

Ok so like, I do of course take time for myself just like any other human. The problem is that in general, my work-life balance is sadly terribly underdeveloped. So I’ll get excited about something, maybe the RTig to Angelscript conversion or compiling on Linux or whatever you want. I’ll code and code at it for days with little break, stopping for only the minimum amounts of time required to keep going much of the time. Obviously a person can only go like that for so long before something gives out, which is eventually what happens to me. This causes the effect that I don’t stop coding because I want to, but because my body and overworked brain forces me to. Sometimes I keep going anyway until what I’ve been feverishly coding barely works, before finally somewhat willingly allowing myself to crash really hard at a point where I’m not being badly bothered by some remaining bug in the system. When I wake up (sooner than I should after such an extended session), all of the remaining bugs in what I’ve been working on are still flitting around my brain as well as ideas for solutions to fix them, accept by that point my drive to code is pretty much gone and won’t return properly until I take a break where I do something other than coding while I’m fully awake. So then I let myself read a long novel or something. The problem is that most times even when I’m trying to enjoy my novel, after a few days I begin chiding myself for not getting any work done and all of the bugs I know I need to fix will not get out of my brain, causing me to pause my book every 45 minutes and look at some code I want to fix before rediscovering that I still don’t really have the energy to continue yet and resume my book. Some time later, the cycle generally completes itself. I’ll get my coding drive back, just completely burn myself out in one several day little sleep extended coding session, and then feel anything from bad to irresponsible after that for not being able to continue working for days afterword. Every cycle of this causes a bit more learning which makes the next one just a bit less brutal, but I still have a lot of work to do in that area before I can work for several hours, stop at a good point, and then knowing I’ve done a full days work, unashamedly spend a few hours relaxing/gaming/reading before sleeping and doing it all again the next day. Wow so that was indeed one little tangent, why did I mention all of the defects in my work schedule? Mostly just to explain the positive effects of the true break that this unexpected Angelscript bug caused. Yes, it certainly added yet another 3 days of delay to this update, which frustrates me just as much as you guys I’m sure. The thing is though, for once in honestly quite a long time I can say I’ve had a genuine, complete, full break. This new bug, or really I think the unexpectedness and bad timing of it, got me into a state where I new on a very instinctive level that I absolutely 100% could NOT problem solve a single iota more. This had the result that for a few glorious days, my brain wasn’t constantly assaulting me with loads of potential solutions to this or that bug while I was trying to relax, and I wasn’t randomly running to my keyboard at any given time of the day to check if this or that idea or solution would be effective. For the first time in a while no part of me felt bad for not getting something done while I was relaxing, as I knew instinctively beyond a shadow of a doubt that I genuinely couldn’t if I tried, just like a person certainly knows sometimes when they physically cannot run any farther. The result was enjoying some of the best sleep I’ve had in a while, and a resulting rejuvenation of my coding energy on the likes of which I haven’t seen in weeks, a couple months?

Anyway, I started coming back online last Wednesday much refreshed, and began getting back into things by doing some badly needed sound design for a few hours, which slowly transitioned into some minor UI tweaks and other little fixes, as well as most importantly, an email to the Angelscript developer with a code sample that would reproduce the bug that was stopping the beta server from launching. He’d fixed it by that Friday! The beta server came back, and I was able to test that character panel and a bunch of other queued up code that I was finally able to run again. Sadly there still seems to be some sort of weird bug with opImplConv methods that I thought was fixed, causing me to have to call explicit functions in various areas of code to cast database objects to strings. I haven’t reported it yet because even though I can cause it, I don’t know enough details yet to tell the Angelscript developer about it and still need to do some debugging first. Who knows, I could very well find out that it was me doing something wrong during my debugging, which is why I never prematurely contact the dev unless I know for sure that the problem isn’t my fault. I don’t think the assert failure I am still able to generate at the time of writing this post is my fault though. So that’s pretty much all of what’s been going on with Angelscript. Though it even managed to knock me out mentally for a couple of days, I’m glad to report that while I still know of one more unresolved bug that exists in the library which will take a couple hours to report, it doesn’t effect us much. As such, at least in regards to Angelscript, I’m glad to say that the server currently appears completely stable!

Status on RTig to Angelscript conversion

I said in the last post that I would give more details on the worldmaker’s efforts to convert all of our old RTig scripts to the new shared Angelscript module system, so am doing that now. As I said before I have not been as involved with this part of the process as I’ve been working on the hardcode (the part of the code that only I have access to) for a while now. But from what I can tell, it’s going pretty great! I think the only old animal that hasn’t been converted yet is the polar bear. All of the food scripts that could be generalized are converted, loads of crafting recipes are done, the arena control scripts are 90% complete, a good portion of the weapon scripts are converted, scripts for finished new quests are converted, much more. There is still some left to get done, mostly just random use scripts and a number of somewhat more obscure interacts, the description scripts (such as pressing enter on a death free card in the inventory), a couple weapons, and the biggest thing is organization. During the early stages of this conversion, there were a bunch of cases where a script would get converted, but then the old script wasn’t deleted encase it needed to be referred to in the future or for any other edge case. It turned out that this started creating confusion over time, and the chance that one worldmaker could convert a script that another one had already converted. Obviously that can’t continue, and so is now being dealt with. Worldmakers now delete old scripts after conversion of course, so this little mistake and resulting confusion should dissipate quickly. There are a bunch of old RTig scripts having to do with specific game events that will not be needed for the present, such scripts are being skipped over for now. Maybe worldmakers who are more involved in this process would give me a slightly different number and if so I apologize for the invalid estimation, but from what I see here, if I had to put a number on it, I’d say maybe the process is a good 80% complete or somewhere around there. It would probably be a bit higher accept for the fact that as you’ve likely read above, the beta/development server was down last week because of the Angelscript bug and it’s effects, making it so that the only one who could work on any game code was me for that week when I was able. I fully expect this script conversion process to continue to go well, and of course will give another update on the situation again in the next post!

Other minor improvements

As with most times, in the process of working on these larger systems I’ve mentioned, there is the usual little slew of much more minor improvements and smaller bug fixes. In this case, to get tickets working I had to fix an endless loop scenario in the database object class involving tables with non-autoincrementing primary keys, that one took a couple hours to stomp out. Then later when testing the converted vending machines, another rare endless loop scenario in the map system showed itself that I had to attend to. Endless loop scenarios are when the application just hangs infinitely do to a logic error that causes the same section of code to repeat over and over again forever doing nothing useful until the process is killed. Aside from the fixed endless loops, I also fixed a problem we’ve been dealing with since the SQL conversion where character’s clothing items weren’t saving properly when the server restarted. Though I didn’t get to it much over the last 2 weeks, I just barely touched a couple things in the object broadcasting system regarding lag reduction (I think I’m gonna hit lag/stability hard again in the next few days). There were some other random UI improvements (mostly for admin panels), but also a couple other features were added to the remote dialog layer which will be used in general such as the ability to remotely change the label/caption of a dialog control whenever needed, fixing an issue with dialog layering and previously focused elements, a few minor things like that. As I mentioned briefly above, I got some sound design done as well. Me and the masters have started configuring rank settings so that moderators only have access to the features of the new administration panels that they need to, though this process isn’t complete yet. I also converted the logs for point and credit spending over to SQLite, as I hadn’t gotten to those yet (beta members aren’t usually spending points and credits so they weren’t priority for a few weeks). As I’ve said before once we begin talking about changes that are this small, it sometimes becomes difficult for me to remember them enough to write about them, for example it took several minutes for me to remember what one of the endless loop scenarios above even was. That being said though, at the time of writing I cannot think of any other minor bugfixes or small changes that have happened over the last 2 weeks (though I’m sure there are a couple). If I remember any, you’ll of course hear about them in the next post!

The todo list as it currently stands

To read more about some of what’s mentioned here in more detail, please see the previous 2 posts for more information. The remaining things we need to get done that have already been mentioned before include the libgit2 integration and the improvements to the tracking system as well as finishing the remaining 20 odd percent of RTig script conversions. Also, though it hasn’t been mentioned as much recently, it’s imperative that we establish a well planned and properly configured system for data backups, including transferring backups to multiple remote sites at the same time. A previous commenter on this blog recommended this duplicity software as a start. Well, since I don’t really have much knowledge of what exists for automatic backups on Linux anyway and was just planning to code my own system instead, I decided to look into this thing and indeed it looks cool (thanks Ordrin)! It already seems to have support for several remote backends, from Dropbox to Google/one drive to Amazon s3 and more. It’s also in python which I’m already familiar with, so likely I can get this thing interfacing with the game’s server application so that administrators can begin and view the status of backups at any time. I’ve already begun setting up some of the infrastructure, but I’d say the task is only about 5 to 10% complete so far. I think once I actually sit down and give the issue my full attention for a few hours though, I’ll get it working pretty quick. We must have this system working before public beta, I refuse to allow us to lose any more data do to the lack of good backups! As mentioned above, I have to add those 4 final controls to the character panel, and the masters and I also must finish the rank metadata, then we can actually train moderators on all of this shiny new administration UI!

One new thing that we’ve known about for ages but has always been set aside in favor of fixing other issues… the player watching system is like, severely busted at present. I can’t even list all of the ways how, but to say the least it just flat out doesn’t work right now. I’m pretty sure beta members know many more details on this than me, and I will be asking them/debugging it myself as soon as I give it my full attention. All I know is that I’ve been told several times that the system is pretty much useless right now, and it usually gets brought up after some sort of weird server crash or fluke that was caused by the look system’s testing. As I run out of other stuff to deal with, this issue will certainly rise much higher on the priority list and will be attended to. The only other thing that I can think of right now, and it’s quickly rising in importance, is the massive amount of English text I need to write. I mentioned most of it in previous posts, the rules and the website touchup, but something else is becoming a huge problem, and that is worldmaker/staff documentation in general. When we promoted new staff members last June and July, we had quite a bit of it written (though even at that point certainly not enough and it was already getting old). Then a bunch of new systems got added or changed and we dropped RTig in favor of Angelscript, and suddenly even more of our docs were out of date! The result is that worldmakers have been using existing code and a lot of questions to learn how to script things in the game, rather than properly written documentation with examples. The original plan was to sit down and focus on these docs after public beta. Unfortunately though, I’ve now seen several situations where a worldmaker codes something in a difficult way in order to accomplish their goal, only to find that they could have had a much easier time of it had they known about this or that function/utility that I’d already provided them, but that they didn’t know about because no documentation really described it properly. Usually it’s smallish things as the worldmakers have done really well just using the example code provided, but it’s still becoming increasingly clear that it may be very helpful to the games development if I drop everything for a day or 2, and just pound out several pages of highly needed worldmaker documentation that will let staff use the available systems to their full potential. On top of that, documentation also needs to be written for the new administration UI panels so that new moderators have some static learning material. For reference, this one blog post has taken oh I don’t know, maybe 5 hours at least for me to compose? That’s just a guestimate, but it has certainly taken several regardless. The amount of text I have to write for the game is certainly much longer than this blog post, probably by several times and so I do expect it to take at least 1 or 2 days of dedicated writing. So I’m just going to apologize ahead of time, because I know for a fact that I’m forgetting something. I was proof reading this post and thought of something I wanted to add to this section, but by the time I got back to it, whatever I remembered had completely slipped my mind, despite the 10 minutes I’ve been sitting here trying to remember it. So annoying! Aside from whatever it is I can’t remember though, that should be a pretty full accounting of the major hurtles we have left!

A note about the question of release times

Understandably, the question that every single reader of this post has on their minds is “when will this update be released!” TLDR? If I answered this question, all of the players would be far more sure of the date that I gave than I would be myself! I am fully aware that I have not given a time estimate in this post. If you wonder why, I’d ask you to please read between the lines. In both this post and the last, you have seen accountings of multiple unexpected situations, these are things that 2 or 3 weeks ago, we either didn’t know about, or misjudged the difficulty of implementing or fixing. Then if you look at the comments of the previous blog post, you will notice that toward the bottom, some people were already starting to get anxious because this post was not made *EXACTLY! 2 weeks or sooner after the last one even though multiple times I clearly said that I’d post approximately every 2 weeks, not *exactly 2 weeks on the dot. Now, if you put those 2 things together, just imagine what would happen if I gave a release date for the game, and then 2 days before the release date, a beta member finds some critical security vulnerability that takes me 4 or 5 days to research and fix. First, The amount of disappointment from players who were looking forward to an exact date or time would be so palpable and overwhelming that it would effect my ability to continue working, causing me to make terrible rushed coding decisions to try to appease the disappointed players as soon as I could. Of course, such rushed coding decisions would come back to bite us probably not just later, but soon. Secondly, this would damage people’s trust in what I write. Even if I use words like approximately, hoping for, maybe etc, either do to the language barrier or over excitement, a large portion of people tend to take release dates/times as a promise when it’s not intended as such. To such people, if I then don’t deliver, I’ve then broken my word even if the reason for not delivering the release on a given arbitrary date was completely out of my control unless I wanted to reduce the quality of the final product just to stick to a point on the calendar. I don’t want anybody to develop the idea that what I write here may not be true just because some unexpected bug or issue stopped be from being able to stick to a certain release time that was only supposed to be a hope. Giving a release date, or even estimate, is only a wise thing to do if I am absolutely certain that I can stick to it. Because if I’m not, and I can’t, the consequences are such that it is not even remotely worth the risk of basically pulling a date out of thin air and being like “yep I think I can do it by then!” even though really I have no idea whether I could actually do so. For these reasons and more, my current policy is to never publicize hopes of release dates to my games unless I have next to no tasks remaining before the release is to take place. I want to be clear, I am stoked beyond words that people care enough and are excited enough about my game to even ask such a question, and I am not at all mad or frustrated at anybody for doing so! I’m simply trying to explain why answering this question is not a healthy decision either for me, for the game, or even for your excitement. Instead, the promise that I can make is that there WILL be a new blog post every approximately two weeks that will show, in detail, that development is actively continuing on the project. This way, you know as much as I do about when this game will finally be released. I genuinely appreciate and thank every one of you for your understanding here, and assure you that though I can’t say when, this game is certainly going to come back online and it’ll be better than ever before when it does!

And that’s a wrap

As I’m writing this last paragraph I’m racking my brain to see if there is anything else I forgot to talk about here, but presently, nothing is coming to me. My hope is that the next 2 weeks involve as little of the unexpected or unplanned for as at all possible. Aside from early last week, things are going pretty well here and I look forward to providing you guys with more positive news and developments in the next post. As always, me and the staff thank everybody for your continuing patients, support, kindness, and simple trust in the fact that we *WILL get this out to you guys even if we don’t know exactly when! As for me and the staff, I can tell you that there isn’t much that we’re looking forward too more than watching all of our players react to and enjoy all of our work, and we want that day to come just as much as all of you do! Well, the more time I spend yammering on here, the more time I’m spending not writing the code that will finally get this thing released, so with that being said I’ll get to proof reading this thing and will see you all again in around 2 weeks!