Strong point. It puts the dev in a tough position because its hard to discriminate between bugs that MIGHT affect users and those that definitely will.
Not really. Mod makers generally fix
all real bugs that come to them. But there's the second layer who are put in the tough position. Those are the server admins. They must decide if a particular update is important/critical enough to update on their server. That decision should be solely up to the opinions of the server admin, not the modmakers. And the users should not get punished for a difference in opinion between the two. Yes, some servers run bleeding-edge. Reika probably likes them. But not all should be
required to. I'm more old-school and prefer long-term stability, meaning fewer updates unless features are added or issues are encountered by the players on
my server. That's the course taken by enterprise solutions for
very understandable reasons.
Be mindful of regression scenarios. You're inadvertently creating the image that the dev runs in circles creating bugs from bugs, whereas in reality 99% of the bugs are based on brand new mechanics.
Naw. Not what I'm meaning. Many are based on new mechanics. Some are inadvertently created by an unexpected interaction from fixing a previous bug, something far more common than many would admit. Also not something I ever blame a dev for. Code is unwielding and complex, y'know?
If they exist, they will affect your server. They may not be doing so now, but would you rather not update before someone crashes it and makes it temporarily unloadable by placing a GPR, corrupts it by Teleposing a pylon, or gets 6500000 blaze powder from an exploit?
You assume good admins don't read changelogs and aren't capable of making judgement calls for themselves. Please don't do that.
No. All but a very few bugs of mine get introduced in major versions, as a result of content addition or system redesign. Minor versions almost never have those kinds of changes, and even when they do, they are not the same severity as major version bugs. Seeing as the two week period is longer than any minor version will be later than its parent, this is a non-issue: You can go straight from v5d to v6f without issue.
Almost never != never.
Or you could add all the mods at once, seeing as you often have to do that anyways given API issues? I maintain a server and do that too. It takes maybe an hour a week.
No. To properly test it takes well longer than an hour if you run a modpack of decent size. And then you must not only test it client side, but on a 'test server' as well as some issues only manifest when played on servers.
You missed the part about only telling the pack dev, or that is insufficient for you.
If I can permanently turn it off for my users then that's sufficient for me. I keep them on on my own version. But it must be a
permanent disable, not one that will be overridden at anyone
else's whims. All my
users need to know is that they need to run the newest version of my pack and to report all issues to
me. I, then, verify against the newest version of the mods, yes sometimes newer than what's in the pack, and only pass it on to the devs if it still persists then.
And you also missed the command, apparently.
Last I checked from every source, including yourself, that only persisted for x number of major versions, not permanently. That isn't sufficient for me.
This is what I mean. I do not mean to target you personally, but you are doing exactly what I am talking about when I say people miss half the tools at their disposal and then complain as if they do not exist.
I haven't missed them. All information provided tells me that it isn't permanent and
will be overridden once they become too old for
your discretion. Your judgement calls on what version I must run doesn't matter. Just as I shouldn't tell
you what you should run or do with your mod, you shouldn't tell
me what version I need to be running.
As I said above, I do the same - and make the damn mods - in less time. Also, judging from your and other's posts, it is not that two weeks is insufficient. I could make it a month, two months, six months - totally destroying the point of an update checker, given my major versions are usually about 1.5 months apart - and you would still complain.
Good on ya. I have other things to do with my free time than
just administer my modpack and server. I have other obligations and interests that take a much larger majority of my time, including helping to lead a sizable RP guild in a MMO. Minecraft is just my side hobby to my main hobby.
Here's the thing, Reika. I think having a checker is a good thing. A
very good thing. But having it irremovable is a bad thing. A very
bad thing. There is a reason every single piece of decent professional software out there has it on by
default but allows you to turn it
off. If a user wants to run an old version for
whatever reason they choose the should be allowed to without harassment. This is doubly true for servers that desire long-term stability. They should
also not expect to have their ancient bug reports listened to, though.
Having the ability to turn off a version checker doesn't destroy the point of it. It just gives people an option if they're fine running old versions that have known, and fixed in later versions, bugs. Sometimes the bugs you know are more desirable than the ones you don't, you know?
I should also note that I believe that last update to
any modpack before they go unsupported should turn
on all version checkers. Because at
that point it becomes the user's prerogative.