Discourse DMZ
-
Current dev version has a PHP requirement of 5.3.8 and a MySQL of ... 5.1? But it's still mostly legacy codebase
Trouble is, even some of the current dev people have (publicly) said that 2.1 is a version no-one wants to work on because it's going to be effectively obsolete the minute it launches.I also don't want to step up to dev myself because I know what it will do to me on a personal level. It would screw me up trying to do that.
-
I was debating between single and double, I feel single better suits my needs. Double costs too much.
-
I thought they were called floats not singles these days?
-
N...o...?
-
I thought they were called floats not singles these days?
They're both technically floats: single-precision floating point and double-precision floating point. The term you actually use may vary depending on the language you are using.
-
No one likes floaters.
-
I was trying to be funny. You know, double precision numbers versus what used to be called single precision but in most languages these days seems to be called just floats?
Sorry, I'll try to make my humour make more sense next time. I'm hungry and can't eat anything now until tomorrow morning (blood tests)
-
I'll have a milkshake for you with my dinner.
-
Enjoy!
-
Take your PHP nonsense elsewhere, sir.
-
Don't think you'll find it's just PHP... but I'm going already. Got too much code to write and not enough time to write it all...
-
Speaking of coding things,
http://what.thedailywtf.com/t/the-integration-of-things/2854
-
<?php exec('/usr/bin/env/python header.py'); ?>
-
Speaking of coding things,
http://what.thedailywtf.com/t/the-integration-of-things/2854
PM?
-
You're welcome to join in as well.
-
Evidently not. :(
-
No?
It's in TBD because it's not private, I just know @Arantor does/did unity3d stuff. I don't think anybody else spends much time playing with it.
-
I have no idea what TBD is.
-
I don't know what TBD is either :( As for Unity, I know what's going on with it but I wouldn't call myself especially competent
-
I have no idea what TBD
private forum for Programmer Testing > To Be Determined category.
-
Discoverable.
-
Who writes the automated tests? In most cases, it's the same software developer that's developing the code under test.
That's totally true.
Your colleague at the next table, the human tester, might have a different opinion.
Indeed, and analogous situations happen like that all the time. My goal is to write a failing test whenever I start to fix a bug, and have the test passing when I check in the fix. I fail at this a lot, because I'm a lazy shit like most every other human being. But it's good to have goals.
I still catch regressions with the existing test suite, though.
-
-
Sorry I gotta pause for a second. I'm not used to the mental gymnastics required to intepret a straw man that big. I'll reply more later.
You get used to it.
Okay... If you redefine "doing one's best" as "stopping new feature development and fixing all the known bugs" I guess I didn't demonstrably do that!
But this time, he's pretty much right about "doing one's best to finalize a stable release."
-
For me, it's the fact that I spend 5-10 minutes writing up a bug report, working out repro steps, fighting discourse to format it correctly, and then Jeff deletes the bug post, doesn't PM me why, and subsequently deletes any follow up posts when I complain about it.
YES. The passive aggressive bullshit over on meta.d is...bullshit.
-
That's what I meant by my post, kind sir dickbag.
@blakeyrat, please update your long name accordingly.
In fact, to elaborate since you probably did some brain damage by hitting your head against that wall, the WHOLE point was that I'm not denying we have made mistakes about prioritizing work and we're trying to be better going forward.
Aha! This is another thing. We pointed out a zillion times that you guys are sabotaging yourself by shoehorning Discourse into an issue tracker. It's impossible to use it to prioritize and view backlog, plan milestones, etc. This is possibly your biggest best practice sin, because it makes so much stuff hard[1].
Is there anything specific that's being done for prioritizing stuff? A link to an appropriate meta.d topic would be acceptable here. This is a serious question.
[1] That's what she said.
-
EvilTrout Viewpoint: yeah I knew. I just couldn't think of anything more best praticey. I'll try harder next time.
That's cool. Stick around. TDWTF style trolling takes practice for most people. Only a very few are really naturals.
-
private forum for Programmer Testing > To Be Determined category.
Private in that it's kept from google. It's for people to have other people test their shit. So it's semi-private. @PJH can hook you up and put you in the http://what.thedailywtf.com/groups/programmers_testers group.
-
Oh, it's not visible to everyone?
That... actually explains a lot. I thought it was just a different forum like the bug category with minimum level 1 permissions.
-
I guess I'll have 2 things to PM @PJH about (applying those styles to mobile and joining that group, maybe?) unless he doesn't gloss over this topic since it blew up after the point I gave the styles to make "new" stand out again.
Unless this notification catches his eyes.
-
eviltrout said:
That's what I meant by my post, kind sir dickbag.@blakeyrat, please update your long name accordingly.
I second the motion!
-
-
2^2
-
I object to "kind".
-
if (kind is System.Object) {
-
Some of y'all are a bunch of toxic dickbags.
I'm not. I'm a fucking cunt, an imma cutcha.
@eviltrout said:I wonder what could have possibly happened to breed that kind of negativity.
There's 2 things.- Your high priest's attitude, both here and on meta.d
- The introduction of Discurse gutted the community; certain members objected to the fundamental misdesign issues it represents so much they upped and left, c.f. Lorne_Kates and others.
- Discussion of Discord's multiplicity of failings is still, nearly 4 months on, drowning out what this site is about.
Prioritizing stuff is always hard.
And it's a fuck sight harder if you're trying to do your bugtracking on a piece of forum software. Do you, perchance, use Gimp to do your time recording?
@eviltrout said:we'll try to do better in the future.
I doubt it. If you don't understand what blakey's getting at when he starts talking about best practice, instead claiming it's a massive straw man, you will never do better in the future except, possibly, by accident. Because you don't understand what "better" is.There's a joke around these parts. It goes like this:
Days since last Discourse bug : 0
but it's not funny any more; the number has been static since mid-May.
I don't mind good honest bugs, but Discourse is, seemingly, made of race conditions, bad practices and performance issues. But the team are seemingly focussing on whatever is itching <a @twatwood's arse at the moment, and all he seems capable of getting itchy about is indicator colours.
You want something priority to work on? Performance. Make it so scrolling the "all" page of your profile doesn't crash your browser after a few pages. Make it so typing this message doesn't slow my quad-core i7 down so much it misses fucking keypresses if I type at normal speed.
Work on race conditions. Work on rate-limiting the enter key. And above all, get yourself a decent bug tracker. If Jeff doesn't understand why that's needed, hold him down and beat him with sticks until he does.
-
> blakeyrat said:
The problem is you're demonstrably not doing your best right now,Okay... If you redefine "doing one's best" as "stopping new feature development and fixing all the known bugs" I guess I didn't demonstrably do that!
Sorry I gotta pause for a second. I'm not used to the mental gymnastics required to intepret a straw man that big.
I don't know in what (or how many) companies you have worked before, but SOP in software development, even in open source development, is defining milestones (like "version 1.0") and something called a "feature freeze" in the run-up to that milestone.The reason for a feature freeze is simple: stop introducing new bugs. And free resources so that more existing bugs can be fixed before the release.
Yes, it's more fun developing new features. Yes, it's more fun to play around with UX instead of fixing bugs. But you are co-founder of a company that wants to make money with the product. You should have a vested interest in delivering a solid product, and not in doing the fun stuff. Because the bugs won't go away.
You have been lucky so far to have users like us who care to report bugs. Just consider this: what if the email leaking bug hadn't been reported on meta.discourse.org, but on The Register instead?
-
what if the email leaking bug hadn't been reported on meta.discourse.org, but on The Register instead?
The next one might be. And slashdot, and ars, and - well - everywhere.
-
This post is deleted!
-
Personally, I think every company should hire a "chief Troll" for every team.
Responsible for...
Telling all devs they are doing it wrong
Causing internal conflict between all devs
Posting cat picturesImagine how much better all software would be.
I don't mind good honest bugs, but Discourse is, seemingly, made of race conditions, bad practices and performance issues. But the team are seemingly focussing on whatever is itching @twatwood's arse at the moment, and all he seems capable of getting itchy about is indicator colours.
Hired, you are far more effective than @blakeyrat or @Arantor
also (talking about not updating).
see trust level badges here: https://meta.discourse.org/badges working through the rename, it takes a while, renaming the badges to whatever is 100% supported (cc @PJH) ... rest of the places in the UI is going to come, probably faster if we are trolled more aggressively.
-
probably faster if we are trolled more aggressively
You fool! Do you realise what you've done?! Don't say you weren't aware of the consequences…
-
I suggest a slight electroshock to be administered to Discourse devs every time a TDWTF meta bug tread gets 10 likes. It would at least get @blakeyrat involved.
-
I think a lot people here were hating on Discourse way before that ever happened.
No. It was Jeff's attitude that riled everyone up in the end.
But hey, what else should I expect from a community that's entirely based on making fun of other people's code?
Try "community that works in software development" and you'd be a bit closer.
This community was (ab)used as a testing ground for software that was - at the time - deemed suitable for release back in May. Within days XSS flaws were found along with lots of other bugs. 1.0 was pushed back until September.
Maybe the experiment with us worked, maybe it backfired, but egregious bugs got discovered and fixed because of this community. If the reaction back is "waahhh - no fair - they're complaining" coupled with nail festooned mod-sticks then of course you're going to get bitten back.
Normal posters on any forum are advised to lurk before posting for a reason. Giving Mod to a developer who has no respect for the community or any idea how it works was always going to Not Work. And it didn't. And it backfired - badly.
but I wonder what could have possibly happened to breed that kind of negativity.
See above.
I wonder how much of it was getting off on the wrong foot.
Most of it, to be honest.
applying those styles to mobile
Which ones? I'll admit that while I frequently use the mobile version, I haven't been updating the CSS for that as much as the main site.
and joining that group,
Done.
-
This is bizarre - I couldn't imagine having no human QA.
Automated tests are great for a big number of things, and I do like to automate even the UI tests (using Selenium) but there is nothing but nothing that can beat a human tester for finding both bugs and regressions.
If I had had your experience, I would be ranting exactly like you do.
-
it takes a while, renaming the badges to whatever is 100% supported
Why aren't the names of trust levels configurable per instance?
-
I do,
not masses, but some. I am probably too close to a non-developer unity3d user though, because I keep trying to use the unity editor to make what I want rather than anything else.
-
Personally, I think every company should hire a "chief Troll" for every team.
Now excuse me while I go talk to my boss about my new job description.
-
My experience has mostly been "no automated testing" -- not even unit testing. Not even when it's simpler to do unit tests than to manually repro a bug -- I've seen developers dig into the database and alter the data so they have a test case they can reproduce because they don't know how to do a unit test. So that might be why I take the opposite tactic to Blakey, screaming the benefits of automated testing from the rooftops: because in my company, manual testing isn't going away anytime soon, so I don't have to argue for it.
-
Which ones?
These:
[code]
.badge-notification.new-topic {
background-color: #6cf;
color: #fff;
font-size: 11px;
font-weight: normal;
}
.topic-list .badge-notification.new-topic {
padding-left:3px;
top:-2px;
}
.badge-notification.new-topic:before {
content:'';
}
.badge-notification {
min-width:10px;
padding:5px;
}
[/code]
-
My experience has mostly been "no automated testing" -- not even unit testing.
The problem with Discourse, though, is no testing at all.
Yes, I know, @eviltrout says "we use unit tests, we use UI testing frameworks, blah blah blah". But if you look at the fixes for bugs, they do not contain tests to detect regressions of that bug. And given that bug "tracking" is done using pissforce, there's no easy way of even knowing what bugs have supposedly been fixed, or how to duplicate them manually. Especially given Atwood's penchant for deleting threads.
Hired, you are far more effective than @blakeyrat or @Arantor
Yeah, ha ha ha. Dismiss serious points in a flippant manner. You're almost as funny as your boss. Fucking twat. Next serious security bug I find? It goes public.
