All times shown according to UTC.
| Time | Nick | Message |
|---|---|---|
| 03:23 | rnmissionrun joined #evennia | |
| 05:40 | jaggiJ joined #evennia | |
| 07:42 | DiscordBot | <Viktyr> |
| 07:42 | Viktyr-1 | ... dude. |
| 07:56 | Cloud_Keeper | Fae, you're speaking you yourself |
| 07:56 | To yourself* | |
| 07:57 | Viktyr-1 | I'm testing the fact that HexChat and therefore IRC supports emojis which I did not know and would not have guessed. |
| 07:59 | vincent-lg joined #evennia | |
| 07:59 | vincent-lg | hi Evennia! |
| 08:23 | Griatch | Cloud_Keeper: The question I got was from a guy who came from the LPMud world and was wondering about the need to reload the server to see code changes. I answered with a discussiom about the issues with making a Python sandbox and a bit about our philosophy about modifying code using external tools vs on the command line. |
| 08:24 | I did a show of hands and about 80% of the audience knew of MUDs and some 30% had played one at a time. | |
| 08:27 | The comments I got later were mostly about the talk being cool and that it was 'unexpected' (but really fun, people had a lot of stories to share) to hear about MUDs in today's day and age. There was some talk of using Evennia for the Py-girls program and a guy was thinking about introducing it to his kids to maybe get them interested in coding. | |
| 08:28 | Overall, it felt very appreciated. | |
| 08:31 | DiscordBot | <Viktyr> ❤ |
| 08:31 | Viktyr-1 | Wow, that looks entirely wrong. |
| 08:31 | <3 | |
| 08:42 | Griatch | :) |
| 08:43 | vincent-lg | Griatch: sounds great! |
| 08:49 | Griatch: I think someone succeeded in adding a test for some task that needed some time to execute. I could post tests on the delay utility on the same principle. Do you remember how he/she did it? I don't even remember who or if it were just an idea or actual code | |
| 08:54 | Griatch | Twisted has its own test parents for this if you really want to test the delay. Otherwise one should probably mock it out |
| 08:56 | vincent-lg | Griatch: but Evennia has its own test parent too, inheriting from Django. I guess using one would prevent using the other one |
| 08:57 | it's worth investigating though, I think if we can test delaying we will be able to increase testing even more as some operations depend on the tool. Well... that's what I figured anyway. I'm still trying to find areas where I can add tests :D | |
| 09:01 | Griatch | vincent-lg: you can mix them; I do so on some of the tests as I recall. |
| 09:02 | But we should probably mock more | |
| 09:09 | vincent-lg | yeah, I agree. Well, I'll see this. Do you have any priority in testing? Areas where you see the number of tests is really low and should get higher, for instance? I browsed the Evennia source code really quickly but found the test coverage pretty good, despite the number we see |
| 09:16 | Griatch | I don't know how accessible it is, but https://coveralls.io/github/ev[…]nia?branch=master lists the coverage of individual files. I'd say things in utils/ would be among the more important ones to properly test since they are used so much elsewhere. But also stuff in server/ . |
| 09:17 | vincent-lg | okay. I noticed DefaultObject doesn't have specified test coverage but it might not be very useful to add either |
| 09:17 | I jot that down! | |
| 09:18 | Griatch | Some lines are covered 'by accident' since they are used by other tests; that's why coverall is interesting. |
| 09:35 | vincent-lg | yes indeed |
| 09:37 | when I coded my equipment handler I made sure to test both my equipment methods, and my commands, even though the latter would have been enough. It's testing the same feature at different complexity levels, but it's important to have both I guess. I'll read coverall's report | |
| 10:28 | Viktyr-1 | Taking a break from the dead trees, hitting up the interactive tutorials online. |
| 10:29 | ldlework joined #evennia | |
| 10:37 | Griatch | Viktyr-1: What kind of concepts are you learning about now? |
| 10:37 | Viktyr-1 | Griatch: Hello, World! |
| 10:38 | Griatch: I am a little bit more advanced than that in reality, but only just. | |
| 10:40 | I had a C++ course in high school, but it was literally just mimeographed worksheets that we were expected to type in by hand, compile and run. With no other guidance, and I doubt the instructor could have debugged the programs if they'd been faulty. | |
| 11:05 | Griatch | I see :) |
| 11:06 | So you are aiming to learn Python from the ground up? | |
| 11:06 | Did you go through the Evennia Python introduction tutorial as well already? | |
| 11:10 | Viktyr-1 | No, I reckon that is my next step. |
| 11:23 | DiscordBot | <BadSectorGaming> Is there an easy way to ensure that a script is unique on a single object? It can run on many objects at once but only once on each. |
| 11:26 | Viktyr-1 | Okay, this is doing my head in. |
| 11:27 | DiscordBot | <BadSectorGaming> What's up? |
| 11:28 | <Viktyr> Trying to do Python tutorials. And I'm just not processing. | |
| 11:29 | Griatch | BadSectorGaming: Scripts are defined per object if you assign them like that. What you may be looking for is rather the tickerhandler I'd think. |
| 11:29 | DiscordBot | <Viktyr> Maybe I just need a break. |
| 11:31 | <BadSectorGaming> Griatch: You may be right, thank you | |
| 11:32 | <BadSectorGaming> @Viktyr that always helps me when I get blocked | |
| 11:54 | <Zironic> Conundrum of the dawn of programming. Do you spend infinity hours redactoring code to satisfy your ideas of elegance potentially breaking it and causing countless subtle bugs, or do you just leave it as is because it works and silently seethe to yourself? | |
| 12:31 | Griatch | Zironic: You make unit tests and can then refactor without needing to worry about any subtle bugs or regressions! :D |
| 12:34 | ldlework | tests assume you know the proper design ahead of time |
| 12:34 | a big part of refactoring/iteration is moving code around and seeing "ah, this would be best over here" etc | |
| 12:34 | s/tests assume/tdd assumes | |
| 12:36 | Griatch | If you refactor so old tests fail, you know that your interfaces have changed and have a better chance of cathing unexpected changes. |
| 12:46 | vincent-lg | I don't know if tests presuppose a constance in the inner working of your code |
| 12:46 | I personally have tests at different levels. Tests on commands obviously test the interface and they don't really care how such a thing is coded, as long as it keeps on working the way it should | |
| 12:48 | that's the one that truly matters to me, though I have lower tests that focus on specific handlers... and yeah, these ones are supposed to remain somewhat static and true (if the interface evolves, chances are the test suite will evolve with it) | |
| 13:07 | DiscordBot | <Zironic> If the code you're messing with isn't written according to a standard, it can be really hard to know if there's random pieces of code somewhere calling other pieces of code without going through an interface. |
| 13:13 | jaggiJ | vincent-lg: very interesting riots you got in your country |
| 13:27 | vincent-lg | jaggiJ: yeah :( |
| 13:27 | DiscordBot | <Zironic> Specifically I've been spending the past week or so doing minor UX improvements for Arx and strings are generated in the most arbitrary places. For instance coins in bank is generated by the inventory command by straight up checking assets but coins in hand are generated by the container mixin. I could make this consistent by moving all the string formatting functions to mixin but the end result of that would be many hours of work just |
| 13:27 | to make the code more elegant and no real benefit to any player. | |
| 13:37 | vincent-lg1 joined #evennia | |
| 13:40 | vincent-lg1 | Zironic: I guess that's what refactoring usually is: no benefit for users, if not additional bugs, but cleaner code for developers. Worth considering when it's needed or not of course! |
| 13:41 | gevrik joined #evennia | |
| 14:05 | DiscordBot | <Zironic> Cleaner code if you're lucky. Sometimes you refactor only to decide you actually want a completely different paradigm |
| 14:06 | <Zironic> Also sometimes architectualy clean code can be super convoluted to work with. Kind of like database normalisation |