diff options
Diffstat (limited to '')
| -rw-r--r-- | 2023/captions/emacsconf-2023-test--what-i-learned-by-writing-test-cases-for-gnu-hyperbole--mats-lidell--answers.vtt | 1406 |
1 files changed, 1406 insertions, 0 deletions
diff --git a/2023/captions/emacsconf-2023-test--what-i-learned-by-writing-test-cases-for-gnu-hyperbole--mats-lidell--answers.vtt b/2023/captions/emacsconf-2023-test--what-i-learned-by-writing-test-cases-for-gnu-hyperbole--mats-lidell--answers.vtt new file mode 100644 index 00000000..bc1bfff5 --- /dev/null +++ b/2023/captions/emacsconf-2023-test--what-i-learned-by-writing-test-cases-for-gnu-hyperbole--mats-lidell--answers.vtt @@ -0,0 +1,1406 @@ +WEBVTT + + +00:00:08.740 --> 00:00:09.240 +[Speaker 0]: Do we have any listeners? + +00:00:13.340 --> 00:00:13.840 +It's you and I. I have a question. + +00:00:16.420 --> 00:00:16.640 +How many tests do you have for hyperbole and + +00:00:18.800 --> 00:00:19.279 +How would you rate the test coverage compared + +00:00:21.279 --> 00:00:21.500 +to other packages? Well, + +00:00:28.279 --> 00:00:28.700 +that's a tricky 1. Shall I spell it out loud + +00:00:31.100 --> 00:00:31.600 +and then maybe type it at the same time? + +00:00:36.420 --> 00:00:36.920 +So, I believe it's around like more than 300 + +00:00:43.660 --> 00:00:44.059 +test cases now. But I cannot compare the test + +00:00:45.220 --> 00:00:45.720 +coverage to any other + +00:01:00.020 --> 00:01:00.520 +other package. Maybe I can type that later. + +00:01:01.560 --> 00:01:02.060 +What do you say, Badal? + +00:01:02.660 --> 00:01:02.900 +[Speaker 1]: package. I have no knowledge of any Yeah, + +00:01:03.840 --> 00:01:04.239 +sure, yeah, that's totally fine. + +00:01:05.660 --> 00:01:06.160 +Feel free to just answer them with voice. + +00:01:08.720 --> 00:01:09.220 +[Speaker 0]: Yeah, yeah. There's another question. + +00:01:10.520 --> 00:01:10.920 +1 small suggestion to me, + +00:01:11.440 --> 00:01:11.940 +should means optional, + +00:01:13.660 --> 00:01:14.160 +where shall or must means required. + +00:01:15.940 --> 00:01:16.220 +Not sure if it is too late to make a major + +00:01:17.220 --> 00:01:17.540 +grammar change like that. + +00:01:18.080 --> 00:01:18.580 +Very nice presentation. + +00:01:19.840 --> 00:01:20.340 +So thanks for presentation, + +00:01:24.380 --> 00:01:24.780 +but the package ERT, well, + +00:01:27.920 --> 00:01:28.080 +it's not something that we have come up with. + +00:01:28.920 --> 00:01:29.340 +It's a standard package. + +00:01:32.320 --> 00:01:32.560 +So I believe it has been around for a long + +00:01:37.760 --> 00:01:38.000 +time. So, but please feel free to make + +00:01:39.680 --> 00:01:40.180 +suggestions and maybe you can, + +00:01:43.660 --> 00:01:43.860 +you know, like do a copy or like an alias for + +00:01:46.080 --> 00:01:46.200 +that. If you believe it makes more sense for + +00:01:48.080 --> 00:01:48.580 +your test cases to have that instead. + +00:01:53.540 --> 00:01:53.720 +And then we have another question here. + +00:01:55.540 --> 00:01:55.680 +For your info, you may find this helpful for + +00:01:58.780 --> 00:01:59.020 +running MX test lint both from a command line + +00:02:01.220 --> 00:02:01.720 +and from within MX with a transit menu. + +00:02:03.600 --> 00:02:04.040 +GitHub alpha papa make sure, + +00:02:06.760 --> 00:02:07.100 +yes. It also works on remote CI. + +00:02:08.240 --> 00:02:08.740 +Yeah, thank you, Alpha Papa. + +00:02:10.580 --> 00:02:11.080 +I think I've looked into that, + +00:02:13.440 --> 00:02:13.940 +but we haven't made any use of that. + +00:02:17.920 --> 00:02:18.080 +But maybe you'll inspire me to give it + +00:02:18.400 --> 00:02:18.900 +another look. + +00:02:29.260 --> 00:02:29.760 +[Speaker 2]: Hey guys. + +00:02:34.120 --> 00:02:34.460 +[Speaker 0]: I remember, I recognize that voice. + +00:02:37.160 --> 00:02:37.660 +Hi, Bob. Hey, how are you? + +00:02:40.240 --> 00:02:40.580 +Congratulations, man. Thanks, + +00:02:43.020 --> 00:02:43.320 +Hugh. Thank you. I have another question + +00:02:45.520 --> 00:02:45.900 +here. It is easy to run ad hoc tests inside + +00:02:48.400 --> 00:02:48.600 +an Emacs session given the command line + +00:02:51.180 --> 00:02:51.560 +scripts you need to run to get the batch test + +00:02:54.960 --> 00:02:55.120 +session running? You said it's to run an + +00:03:05.680 --> 00:03:05.920 +ad-hoc test. I'm not sure I understand that + +00:03:14.440 --> 00:03:14.940 +question. Yes, please. + +00:03:15.660 --> 00:03:16.160 +[Speaker 1]: Maybe I can rephrase. Sure. + +00:03:19.900 --> 00:03:20.400 +So I think what I understand is that since + +00:03:22.540 --> 00:03:23.040 +you have to use some of these command lines + +00:03:25.440 --> 00:03:25.940 +scripts to get a batch test session running, + +00:03:28.780 --> 00:03:29.180 +is it easy to run ad hoc tests in an Emacs + +00:03:30.700 --> 00:03:30.900 +session or does that, like in your + +00:03:32.040 --> 00:03:32.540 +experience, has that been difficult? + +00:03:36.820 --> 00:03:37.320 +[Speaker 0]: Well, from the command line, + +00:03:38.660 --> 00:03:38.940 +if you look at the command line, + +00:03:44.160 --> 00:03:44.340 +you'll see that it's only like a few image + +00:03:46.480 --> 00:03:46.980 +functions to call to get that behavior to run + +00:03:55.080 --> 00:03:55.240 +the batch tests. So I think we made some + +00:03:57.100 --> 00:03:57.600 +support function for that in hyperbole. + +00:04:02.800 --> 00:04:02.960 +So it's not, I don't think it's possible out + +00:04:05.540 --> 00:04:06.040 +of the box to do it, but it's not complicated + +00:04:08.060 --> 00:04:08.560 +to do it. + +00:04:12.190 --> 00:04:12.340 +[Speaker 2]: You can define a test anytime, + +00:04:14.780 --> 00:04:15.280 +right? Just like a new function. + +00:04:18.899 --> 00:04:19.240 +So that's ad hoc. You just write your test + +00:04:20.019 --> 00:04:20.519 +and you can run it. + +00:04:22.900 --> 00:04:23.400 +[Speaker 0]: Yeah, yeah, I mean, of course, + +00:04:25.900 --> 00:04:26.180 +but I got the impression it was about running + +00:04:28.620 --> 00:04:29.060 +all your tests like we did with the command + +00:04:35.740 --> 00:04:36.020 +line. Well, so the question is more about how + +00:04:38.260 --> 00:04:38.600 +would you run all your test cases from within + +00:04:44.860 --> 00:04:45.140 +Emacs? And the easy answer to that is + +00:04:48.420 --> 00:04:48.860 +actually you load all your test case files, + +00:04:51.760 --> 00:04:52.080 +and then you run ERT with the T as the test + +00:04:53.600 --> 00:04:53.880 +selector and then it will run all your test + +00:04:53.880 --> 00:04:54.380 +cases. + +00:05:01.780 --> 00:05:01.960 +[Speaker 1]: Right. And I think they have expanded on + +00:05:03.180 --> 00:05:03.520 +their question a little bit as well, + +00:05:04.960 --> 00:05:05.220 +clarifying that. In other words, + +00:05:07.200 --> 00:05:07.360 +can you tweak tests in an Emacs session and + +00:05:08.860 --> 00:05:09.360 +run them right away? Which I believe, + +00:05:11.400 --> 00:05:11.640 +if I understand correctly what Bob was + +00:05:13.820 --> 00:05:14.320 +saying, you can basically define or redefine + +00:05:15.920 --> 00:05:16.080 +functions on the fly and then have them be + +00:05:16.440 --> 00:05:16.940 +run, right? + +00:05:22.200 --> 00:05:22.360 +[Speaker 0]: Yes, yes. You just go into that test case and + +00:05:24.120 --> 00:05:24.620 +you just change it and you run it again. + +00:05:29.060 --> 00:05:29.200 +And either you have to sort of load it or you + +00:05:31.560 --> 00:05:32.060 +can use like the commercial thing I did. + +00:05:36.140 --> 00:05:36.340 +You use hyperbole and just hit meta return on + +00:05:38.560 --> 00:05:38.860 +the test case and it will load it and run the + +00:05:42.240 --> 00:05:42.360 +test case again. So that's of course what you + +00:05:44.220 --> 00:05:44.720 +normally do when you're defining a test or + +00:05:47.440 --> 00:05:47.940 +debug a test case or develop a test case. + +00:05:49.960 --> 00:05:50.460 +Just start with something small, + +00:05:52.700 --> 00:05:53.200 +just make sure maybe you can prepare the test + +00:05:55.320 --> 00:05:55.680 +properly and run it again and again and again + +00:05:56.720 --> 00:05:57.220 +until you're ready with it. + +00:05:59.760 --> 00:05:59.960 +That's a good point. You can definitely do + +00:06:02.800 --> 00:06:03.280 +that and that's part of how I normally + +00:06:06.420 --> 00:06:06.920 +develop the test cases that I mean start with + +00:06:09.160 --> 00:06:09.400 +something small so I can see that I get there + +00:06:12.180 --> 00:06:12.240 +maybe the right input in the buffer that I + +00:06:14.180 --> 00:06:14.340 +want to test on or something and I expand on + +00:06:18.160 --> 00:06:18.400 +that more and more and add more and more more + +00:06:18.460 --> 00:06:18.960 +and more more + +00:06:31.040 --> 00:06:31.540 +[Speaker 2]: tests to it. You might tell them a bit about + +00:06:33.280 --> 00:06:33.480 +how many test cases you have. + +00:06:36.020 --> 00:06:36.440 +I guess you commented on that and like what + +00:06:40.320 --> 00:06:40.820 +happens, you know, with the CICD pipeline, + +00:06:43.780 --> 00:06:44.020 +every time we commit, you know, + +00:06:46.360 --> 00:06:46.560 +across all the versions and what you have set + +00:06:48.760 --> 00:06:49.040 +up there because you know I wish people could + +00:06:53.940 --> 00:06:54.140 +see it. You can go and check on GitHub and + +00:06:57.440 --> 00:06:57.720 +you can see the logs right of any of the + +00:06:59.760 --> 00:06:59.960 +builds and but tell them a bit about that + +00:07:01.080 --> 00:07:01.320 +Mats because I think that's pretty + +00:07:01.320 --> 00:07:01.820 +impressive. + +00:07:07.280 --> 00:07:07.760 +[Speaker 0]: Well, that's part of more the CI, + +00:07:11.760 --> 00:07:12.160 +CD, part of how we developed this using + +00:07:15.460 --> 00:07:15.580 +GitHub and workflows that you get out of the + +00:07:20.740 --> 00:07:20.900 +box from there. So this more than 300 test + +00:07:23.440 --> 00:07:23.720 +cases on our round for I think 5 different + +00:07:26.480 --> 00:07:26.980 +versions of Emacs when we do a pull request + +00:07:33.900 --> 00:07:34.400 +or a commit. So that's a good way to ensure + +00:07:38.040 --> 00:07:38.540 +that it works from version 27.2 + +00:07:42.240 --> 00:07:42.740 +up to the latest master version because + +00:07:45.860 --> 00:07:46.360 +there's some changes in Emacs over different + +00:07:48.940 --> 00:07:49.340 +versions that can affect your functions or + +00:07:49.600 --> 00:07:50.100 +your code. + +00:07:56.580 --> 00:07:56.720 +[Speaker 2]: They all run in parallel and so typically in + +00:08:00.580 --> 00:08:00.780 +under 60 seconds I think you've got all of + +00:08:03.960 --> 00:08:04.460 +them run so you've got pretty extensive + +00:08:08.860 --> 00:08:09.240 +testing which does catch interesting bugs + +00:08:09.760 --> 00:08:10.260 +here and there, right? + +00:08:13.320 --> 00:08:13.820 +[Speaker 0]: Yes, of course it does. + +00:08:18.060 --> 00:08:18.560 +I mean, you normally develop with 1 version + +00:08:20.280 --> 00:08:20.540 +and then you think everything is okay. + +00:08:21.720 --> 00:08:21.820 +But then when you're tested with the + +00:08:23.460 --> 00:08:23.960 +different versions, you find out that there + +00:08:26.080 --> 00:08:26.580 +are some changes and there are things you + +00:08:30.140 --> 00:08:30.400 +might not sort of keep track of what's + +00:08:34.340 --> 00:08:34.640 +happening also. So that's a way to get + +00:08:38.559 --> 00:08:38.940 +noticed that the core developers of Emacs + +00:08:41.120 --> 00:08:41.480 +have changed something that you sort of based + +00:08:44.380 --> 00:08:44.840 +your code on. Now I got another question + +00:08:47.900 --> 00:08:48.160 +here. Did you have to change hyperbole code + +00:08:50.580 --> 00:08:50.760 +and design to be more readily testable as you + +00:08:52.160 --> 00:08:52.660 +were increasing your test coverage? + +00:08:55.520 --> 00:08:56.020 +Well, we haven't done that to a lot, + +00:09:00.160 --> 00:09:00.320 +to a big degree, although I believe that that + +00:09:03.760 --> 00:09:04.260 +is an important thing for sort of the future + +00:09:06.020 --> 00:09:06.500 +to do that because some of the hyperbolic + +00:09:08.520 --> 00:09:08.720 +functions are very complicated and long and + +00:09:10.640 --> 00:09:11.140 +that makes testing them rather difficult. + +00:09:14.660 --> 00:09:14.900 +So, at a few places we have sort of broken up + +00:09:17.260 --> 00:09:17.720 +functions in smaller pieces so it'd be easier + +00:09:20.280 --> 00:09:20.660 +to do like unit tests of the different parts + +00:09:27.740 --> 00:09:27.980 +of it. But there's a lot of more work that + +00:09:28.680 --> 00:09:29.180 +has to be done there. + +00:09:33.820 --> 00:09:34.020 +[Speaker 2]: 1 of the nice things is you know the great + +00:09:36.760 --> 00:09:36.820 +environment in Lisp where we're able to do a + +00:09:40.520 --> 00:09:40.900 +lot of interactive bottom-up testing before + +00:09:42.840 --> 00:09:43.280 +we even get to lighting tech pieces. + +00:09:48.740 --> 00:09:49.140 +So it does tend to be more higher level bugs, + +00:09:51.140 --> 00:09:51.640 +I think, that get caught in cross-functional + +00:09:55.940 --> 00:09:56.100 +interaction. We had 1 recently that was an + +00:09:58.100 --> 00:09:58.600 +Emacs version change. It had been a function + +00:10:01.100 --> 00:10:01.600 +that had existed for a long time. + +00:10:03.340 --> 00:10:03.840 +It had an and rest in it, + +00:10:05.740 --> 00:10:06.240 +in its argument list, so it would assemble + +00:10:08.600 --> 00:10:09.100 +the list of arguments from individual + +00:10:10.320 --> 00:10:10.820 +arguments that you would give it, + +00:10:13.100 --> 00:10:13.600 +and they decided in a recent version, + +00:10:15.200 --> 00:10:15.700 +I think with Stefan's input, + +00:10:19.400 --> 00:10:19.840 +to change that to a list and allow the prior + +00:10:22.740 --> 00:10:22.900 +behavior, but it would issue a warning if you + +00:10:23.620 --> 00:10:24.060 +use the prior behavior. + +00:10:25.560 --> 00:10:25.840 +So all of a sudden, the way you were supposed + +00:10:27.180 --> 00:10:27.680 +to do it became semi-invalid. + +00:10:30.440 --> 00:10:30.940 +And so we started getting the warning, + +00:10:32.760 --> 00:10:33.040 +and we've tried to eliminate all those + +00:10:35.600 --> 00:10:36.060 +warnings in recent hyperbole developments. + +00:10:37.120 --> 00:10:37.620 +So we're like, what do we do? + +00:10:39.020 --> 00:10:39.440 +You know, because we wanted to be backward + +00:10:42.140 --> 00:10:42.640 +compatible to where you couldn't use a list. + +00:10:44.620 --> 00:10:45.120 +It required you to use individual arguments. + +00:10:48.380 --> 00:10:48.560 +And now it's sort of requiring you to do + +00:10:51.660 --> 00:10:51.820 +that. And all of that was caused by the + +00:10:52.940 --> 00:10:53.440 +automatic testing on it. + +00:11:08.680 --> 00:11:08.860 +So you said, Max, you were going to tell us + +00:11:12.740 --> 00:11:13.220 +what you learned. So what are the major + +00:11:15.368 --> 00:11:15.396 +things that you learned in doing all of this + +00:11:15.680 --> 00:11:16.180 +work? All of this work? + +00:11:26.520 --> 00:11:26.740 +[Speaker 0]: Well, I tried to cover some of it in the + +00:11:29.380 --> 00:11:29.800 +presentation, but as I was going along, + +00:11:33.420 --> 00:11:33.920 +the presentation became like twice as long as + +00:11:36.180 --> 00:11:36.680 +fitted into the time we had so I had to cut + +00:11:42.380 --> 00:11:42.880 +it out. But I think some of the core things + +00:11:44.340 --> 00:11:44.840 +still is in the presentation. + +00:11:49.560 --> 00:11:50.060 +From a personal perspective, + +00:11:52.440 --> 00:11:52.940 +And this might not be hard to realize, + +00:11:56.960 --> 00:11:57.460 +but forcing yourself to test functions, + +00:12:02.900 --> 00:12:03.060 +test code really forces you to understand the + +00:12:05.080 --> 00:12:05.280 +code a little bit better in a way that sort + +00:12:07.300 --> 00:12:07.400 +of makes it easier than just to read the + +00:12:11.460 --> 00:12:11.960 +code. I don't know how it is for the rest + +00:12:13.780 --> 00:12:13.980 +listening to this, but for me it works so + +00:12:16.580 --> 00:12:17.080 +that if I just read the code then I don't + +00:12:20.140 --> 00:12:20.320 +sort of become as sharp as I should be but if + +00:12:22.500 --> 00:12:22.640 +I try to write the test case for it then I + +00:12:24.680 --> 00:12:24.880 +really need to understand better of all the + +00:12:27.660 --> 00:12:28.160 +edge cases and all the sort of states and etc + +00:12:30.060 --> 00:12:30.320 +that is involved and I think that's That's + +00:12:33.080 --> 00:12:33.200 +what's sort of 1 of the learning things I + +00:12:34.960 --> 00:12:35.280 +wanted to communicate as well that I don't + +00:12:38.940 --> 00:12:39.080 +think I covered in detail in the + +00:12:41.480 --> 00:12:41.980 +presentation. Maybe all this, + +00:12:48.060 --> 00:12:48.340 +but try it. 1 other sort of more from the fun + +00:12:50.740 --> 00:12:51.000 +side is that I really think it's fun to write + +00:12:55.080 --> 00:12:55.440 +the test. So if you haven't tests in your + +00:12:58.020 --> 00:12:58.520 +package, you should start doing that because + +00:13:05.740 --> 00:13:06.080 +it is fun. It might feel like some extra + +00:13:08.080 --> 00:13:08.580 +work, but it really pays off in the long run, + +00:13:10.320 --> 00:13:10.760 +especially if you have it in like a pipeline + +00:13:12.520 --> 00:13:12.980 +and where you can run it regularly when you + +00:13:13.940 --> 00:13:14.380 +do new commits, et cetera. + +00:13:16.560 --> 00:13:17.060 +So, I mean, that's maybe obvious from, + +00:13:19.160 --> 00:13:19.440 +if you look from the commercial side or your + +00:13:21.080 --> 00:13:21.340 +work side to do it like that. + +00:13:22.260 --> 00:13:22.660 +But even for your hobby project, + +00:13:26.260 --> 00:13:26.760 +it can be very sort of pay off really well. + +00:13:32.900 --> 00:13:33.160 +[Speaker 2]: It's worked really well when we're adding new + +00:13:35.020 --> 00:13:35.180 +functionality or we're changing some of the + +00:13:36.560 --> 00:13:37.060 +plumbing in the system. + +00:13:40.400 --> 00:13:40.580 +You know, you go and you do some surgery and + +00:13:41.320 --> 00:13:41.820 +then you run the tests. + +00:13:45.400 --> 00:13:45.900 +And sometimes 6 to 10 tests will fail. + +00:13:48.260 --> 00:13:48.420 +And you find there, you know, + +00:13:50.460 --> 00:13:50.660 +it tends to be they're all interconnected and + +00:13:52.920 --> 00:13:53.320 +it leads you back to the single source. + +00:13:56.660 --> 00:13:56.980 +You fix that and you know it could be an edge + +00:14:00.560 --> 00:14:00.760 +case and off by 1 or Sometimes it's an + +00:14:03.520 --> 00:14:03.800 +assumption about the way something is used + +00:14:05.980 --> 00:14:06.480 +and it's not actually always true. + +00:14:09.520 --> 00:14:10.020 +And so, Matt's just really good at + +00:14:13.540 --> 00:14:14.040 +identifying some of those scenarios and + +00:14:17.480 --> 00:14:17.980 +keeping us honest, I guess I would say. + +00:14:22.900 --> 00:14:23.400 +So I love, I run it as much as I before, + +00:14:26.400 --> 00:14:26.900 +you know, even before I commit something. + +00:14:29.960 --> 00:14:30.060 +So I get to see, you know, + +00:14:30.940 --> 00:14:31.440 +if anything has progressed. + +00:14:39.480 --> 00:14:39.920 +So yeah, I really recommend this process to + +00:14:42.120 --> 00:14:42.620 +people. I haven't seen it done. + +00:14:45.720 --> 00:14:46.080 +I don't think that, I don't know any other + +00:14:47.800 --> 00:14:48.300 +package that has done it to this level. + +00:14:51.560 --> 00:14:51.820 +And it's been working really great for us. + +00:14:55.440 --> 00:14:55.640 +And I think, well, we'll see too when we + +00:14:56.780 --> 00:14:57.280 +release to the general public. + +00:15:04.380 --> 00:15:04.540 +[Speaker 0]: But Bob, also, maybe the test part of + +00:15:06.400 --> 00:15:06.560 +different packages is not the first thing you + +00:15:08.900 --> 00:15:09.100 +look at. So I know there are packages that + +00:15:10.960 --> 00:15:11.380 +have testing, a lot of testing, + +00:15:13.860 --> 00:15:14.160 +but how much, much testing they have or not, + +00:15:16.060 --> 00:15:16.220 +I don't know. It's not what you normally look + +00:15:17.900 --> 00:15:18.400 +into when you look at someone's else code. + +00:15:20.600 --> 00:15:20.820 +You look maybe on the functionality side but + +00:15:22.760 --> 00:15:23.000 +not on how they've done the sort of the + +00:15:26.540 --> 00:15:26.760 +quality side. So there could be other + +00:15:28.780 --> 00:15:29.280 +packages out there that are well equipped. + +00:15:31.800 --> 00:15:32.300 +[Speaker 2]: I hope so. I hope so. + +00:15:39.860 --> 00:15:40.180 +[Speaker 0]: What's the craziest bug you found when + +00:15:44.700 --> 00:15:45.200 +writing these tests? Well, + +00:15:50.760 --> 00:15:50.940 +What springs to my mind just now is that we + +00:15:52.760 --> 00:15:52.960 +were doing some tests or I would do some + +00:15:55.920 --> 00:15:56.420 +tests for when you narrow, + +00:15:57.940 --> 00:15:58.440 +what do you say that? When you, + +00:16:04.500 --> 00:16:05.000 +in outlining, when you sort of compress + +00:16:06.480 --> 00:16:06.980 +things in an outline, so you just, + +00:16:08.540 --> 00:16:09.040 +sorry Bob, maybe you have it, + +00:16:12.100 --> 00:16:12.600 +[Speaker 2]: When you hide text. + +00:16:12.740 --> 00:16:13.240 +[Speaker 0]: What I'm looking for? Yeah, + +00:16:15.580 --> 00:16:15.920 +when you hide. So I was doing some cursor + +00:16:17.780 --> 00:16:17.980 +movement over that. And I always assume that + +00:16:22.540 --> 00:16:22.900 +if you do like a prefix argument to like a + +00:16:23.800 --> 00:16:24.240 +simple cursor movement, + +00:16:26.420 --> 00:16:26.920 +like control F moving 1 character position, + +00:16:28.340 --> 00:16:28.840 +and you would give it the, + +00:16:36.580 --> 00:16:37.080 +and then the prefix, like you want to move + +00:16:39.140 --> 00:16:39.640 +like 2 or 3 positions, + +00:16:43.040 --> 00:16:43.140 +you would do like control U 3 and then + +00:16:44.240 --> 00:16:44.740 +control F and you move 3. + +00:16:46.560 --> 00:16:46.960 +I always assumed that that would be exactly + +00:16:49.240 --> 00:16:49.440 +the same as if you just hit the key control F + +00:16:50.740 --> 00:16:51.240 +3 times, but it's not. + +00:16:53.160 --> 00:16:53.560 +So it's not the bug, it's a feature, + +00:16:54.620 --> 00:16:55.080 +but that was the craziest thing. + +00:16:58.180 --> 00:16:58.360 +I spent the night trying to figure out why + +00:17:00.720 --> 00:17:01.000 +our code was wrong, but It turns out that's + +00:17:03.560 --> 00:17:04.060 +how Emacs behaves. Try it out yourself. + +00:17:07.920 --> 00:17:08.300 +Try to move over the 3 dots at the end of + +00:17:09.140 --> 00:17:09.640 +that and see what happens. + +00:17:14.060 --> 00:17:14.240 +Do it with cursor hitting the key or using a + +00:17:16.260 --> 00:17:16.680 +prefix argument and you see it behaves + +00:17:18.720 --> 00:17:19.220 +differently. That was the craziest thing. + +00:17:21.960 --> 00:17:22.339 +I think there was some other crazy thing or + +00:17:24.280 --> 00:17:24.480 +deep learning also, but I can't come up with + +00:17:26.599 --> 00:17:26.760 +it at the moment. So maybe I can write it in + +00:17:27.900 --> 00:17:28.400 +the Q&A later. + +00:17:31.200 --> 00:17:31.440 +[Speaker 1]: I think we're out of time on the stream, + +00:17:33.360 --> 00:17:33.600 +but people are welcome to join Mats and Bob + +00:17:35.280 --> 00:17:35.640 +here on BigBlueButton to further discuss + +00:17:36.480 --> 00:17:36.980 +this. Thank you both. + +00:17:38.674 --> 00:17:38.792 +[Speaker 0]: Okay, thank you. Thanks, + +00:17:46.100 --> 00:17:46.600 +Makaay. Thank you. I don't know, + +00:17:48.740 --> 00:17:49.240 +Is it only me and Bob here? + +00:17:50.680 --> 00:17:51.180 +So Bob, do you want to say something? + +00:17:57.440 --> 00:17:57.940 +[Speaker 2]: Well, I think it's been a great day. + +00:18:00.720 --> 00:18:01.220 +And I'm glad we did this. + +00:18:02.280 --> 00:18:02.780 +It takes a lot of energy. + +00:18:15.140 --> 00:18:15.640 +I'm just really excited about the progress + +00:18:20.580 --> 00:18:20.740 +that this, and we're actually doing a lot of + +00:18:23.940 --> 00:18:24.160 +QA at work and my professional software work + +00:18:28.500 --> 00:18:28.840 +and looking at you know how we can do more + +00:18:32.980 --> 00:18:33.480 +test driven development and so everybody's + +00:18:35.980 --> 00:18:36.200 +talking about this you know we've got AI over + +00:18:37.540 --> 00:18:38.040 +here that can generate test cases. + +00:18:40.200 --> 00:18:40.700 +But, you know, strangely enough, + +00:18:43.100 --> 00:18:43.380 +with the rapidity of development and web + +00:18:46.720 --> 00:18:47.220 +applications, I think the level of testing + +00:18:50.140 --> 00:18:50.280 +has gone down in recent years compared to + +00:18:51.500 --> 00:18:51.780 +where it used to be, right? + +00:18:53.040 --> 00:18:53.540 +Because the pace has gone up. + +00:18:57.340 --> 00:18:57.840 +And so I think it's starting to turn again + +00:18:58.740 --> 00:18:59.240 +where people are saying, + +00:19:01.940 --> 00:19:02.440 +we can't just release crap into the + +00:19:08.120 --> 00:19:08.620 +Webisphere and we have to better ourselves. + +00:19:13.620 --> 00:19:13.820 +And with all these advanced tool sets that + +00:19:16.100 --> 00:19:16.600 +you have, that you can do CICD testing, + +00:19:19.860 --> 00:19:20.180 +you know, I just, I just see it coming + +00:19:21.900 --> 00:19:22.100 +around, you know, as people develop new + +00:19:24.000 --> 00:19:24.160 +things. So That's kind of exciting to me + +00:19:26.980 --> 00:19:27.480 +because I came from a manufacturing culture + +00:19:30.300 --> 00:19:30.780 +originally where we, our company actually + +00:19:33.800 --> 00:19:34.300 +started a lot of the manufacturing quality + +00:19:37.420 --> 00:19:37.920 +efforts that you saw in Japan and elsewhere + +00:19:40.600 --> 00:19:40.740 +in America for a long time and that was you + +00:19:42.040 --> 00:19:42.540 +know entirely through testing. + +00:19:46.640 --> 00:19:47.020 +We used to just build incredible test cases + +00:19:49.120 --> 00:19:49.320 +because we were combining software with + +00:19:51.100 --> 00:19:51.380 +hardware. And if, you know, + +00:19:53.460 --> 00:19:53.600 +the hardware doesn't work and you ship a + +00:19:55.080 --> 00:19:55.520 +million units, you're, + +00:19:57.340 --> 00:19:57.840 +you're in trouble. So, + +00:20:00.260 --> 00:20:00.760 +that was just something we had to do. + +00:20:04.280 --> 00:20:04.780 +And so it's nice to start to see that curve + +00:20:07.020 --> 00:20:07.520 +come around. And I think, + +00:20:10.380 --> 00:20:10.880 +you know, Matt Vance is very modest, + +00:20:16.680 --> 00:20:16.920 +but I think he's really the 1 that started us + +00:20:20.400 --> 00:20:20.580 +down this path and really made it into a + +00:20:24.620 --> 00:20:24.840 +reality. So everybody else just gets to + +00:20:25.760 --> 00:20:26.260 +benefit from that work. + +00:20:27.540 --> 00:20:28.040 +So thanks. + +00:20:32.760 --> 00:20:33.260 +[Speaker 1]: That's awesome. + +00:20:39.960 --> 00:20:40.460 +[Speaker 0]: Thanks. Okay. Yeah. So if there's nothing + +00:20:43.200 --> 00:20:43.520 +more here, then maybe we should just close + +00:20:45.440 --> 00:20:45.940 +this and I go over to write in the etherpad + +00:20:47.960 --> 00:20:48.460 +the replies we had. + +00:20:51.900 --> 00:20:52.120 +[Speaker 1]: Right, yeah, I think, let's see, + +00:20:53.520 --> 00:20:53.760 +I see 1 other person here, + +00:20:55.080 --> 00:20:55.580 +I believe Ihor just joined us. + +00:20:58.780 --> 00:20:59.060 +Yeah. Yeah, so if you do want to discuss with + +00:21:00.220 --> 00:21:00.480 +Mats and Bob, you're welcome to, + +00:21:02.200 --> 00:21:02.700 +otherwise, yeah, we can close the room now. + +00:21:05.800 --> 00:21:06.020 +[Speaker 3]: Well, I think I missed most of the talk + +00:21:06.900 --> 00:21:07.400 +because I had power outage, + +00:21:12.180 --> 00:21:12.440 +but the part I heard was about the mock + +00:21:16.860 --> 00:21:17.220 +library. And you mentioned that you don't + +00:21:20.200 --> 00:21:20.700 +like CL-let, but instead you use mock. + +00:21:29.700 --> 00:21:29.800 +[Speaker 0]: Yeah, I was more saying that you have to do a + +00:21:31.560 --> 00:21:32.040 +lot more work when you use the CL letdef. + +00:21:34.540 --> 00:21:34.780 +It's for more ambitious and maybe more + +00:21:37.000 --> 00:21:37.500 +complicated cases where you want to really + +00:21:38.840 --> 00:21:39.340 +make a new implementation, + +00:21:41.940 --> 00:21:42.440 +test implementation. If you use the mock, + +00:21:44.380 --> 00:21:44.880 +you get a lot of things out of the box, + +00:21:47.440 --> 00:21:47.940 +verifying that you actually, + +00:21:50.820 --> 00:21:51.040 +like the mock was actually called for + +00:21:53.320 --> 00:21:53.820 +instance, whereas if you do with the CLLatf, + +00:21:56.520 --> 00:21:56.780 +you would have to take correct track of that + +00:22:02.020 --> 00:22:02.520 +yourself. And so, so a lot of more work. + +00:22:03.760 --> 00:22:04.260 +Oh yeah. + +00:22:07.940 --> 00:22:08.200 +[Speaker 3]: I'm saying that most of the time CLLess is + +00:22:09.720 --> 00:22:10.220 +used for simple cases actually. + +00:22:12.320 --> 00:22:12.820 +Because, just for example, + +00:22:15.100 --> 00:22:15.600 +the function always returns the same. + +00:22:17.980 --> 00:22:18.420 +And it tends to be simple lambda that ignores + +00:22:19.040 --> 00:22:19.540 +all the input arguments. + +00:22:23.000 --> 00:22:23.480 +So that's really trivial most of the time but + +00:22:25.520 --> 00:22:25.920 +I actually thought the opposite that mock is + +00:22:27.640 --> 00:22:28.140 +supposed to be used for non-trivial cases. + +00:22:32.280 --> 00:22:32.520 +[Speaker 0]: Sorry, what was the question? + +00:22:35.280 --> 00:22:35.780 +Mock was supposed to be used for non-trivial. + +00:22:47.680 --> 00:22:48.180 +Yeah I mean I don't know how to explain this. + +00:22:50.140 --> 00:22:50.640 +I mean, CLF can be used for non-trivial + +00:22:54.400 --> 00:22:54.840 +definitely. You can define then any behavior + +00:22:56.180 --> 00:22:56.680 +you want. You can write your own function, + +00:22:58.440 --> 00:22:58.660 +but you need to keep track of whether that + +00:22:59.620 --> 00:23:00.100 +function is called or not, + +00:23:06.260 --> 00:23:06.380 +for instance. So you have to make note of + +00:23:08.440 --> 00:23:08.940 +that the function was called so you can fire + +00:23:12.440 --> 00:23:12.800 +sort of an error in case your function wasn't + +00:23:16.960 --> 00:23:17.440 +called because that would be 1 error case. + +00:23:20.660 --> 00:23:20.860 +[Speaker 3]: So you mean the mock fires an error if the + +00:23:22.580 --> 00:23:23.080 +mocked function was actually not called? + +00:23:30.060 --> 00:23:30.560 +[Speaker 0]: Yes, it does. Yes. So if your assumptions, + +00:23:33.900 --> 00:23:34.120 +you sort of document with the mock also your + +00:23:37.080 --> 00:23:37.220 +assumptions how your code is going to be + +00:23:40.020 --> 00:23:40.380 +called. And if those are wrong, + +00:23:41.120 --> 00:23:41.540 +you will get an error. + +00:23:43.680 --> 00:23:44.060 +So you would, so if the implementation would + +00:23:44.840 --> 00:23:45.100 +maybe change, for instance, + +00:23:46.640 --> 00:23:47.140 +and not call the thing you're mocking, + +00:23:50.460 --> 00:23:50.960 +then you will notice that. + +00:23:53.100 --> 00:23:53.560 +But if you see a letdef, + +00:23:54.840 --> 00:23:55.040 +then you will have to keep track of that + +00:23:57.560 --> 00:23:58.060 +yourself. Okay, I see. + +00:23:58.260 --> 00:23:58.760 +I see. + +00:24:01.240 --> 00:24:01.740 +[Speaker 3]: And you know, our mode also uses a lot of + +00:24:09.340 --> 00:24:09.620 +test. In our mode, we have a lot of tests + +00:24:13.940 --> 00:24:14.440 +[Speaker 0]: Ah, okay. Yeah. Yeah. I'm sure I have. + +00:24:15.900 --> 00:24:16.400 +[Speaker 3]: also. We rely on CLLatF for, + +00:24:19.220 --> 00:24:19.720 +we don't use third-party libraries at all. + +00:24:22.140 --> 00:24:22.640 +[Speaker 0]: Oh, you use CLLatF, okay. + +00:24:26.680 --> 00:24:27.180 +Yeah. Yeah. Yeah. At First I found it very + +00:24:29.480 --> 00:24:29.700 +powerful to use that, but then I sort of, + +00:24:32.120 --> 00:24:32.320 +I learned more about how we can use the + +00:24:34.340 --> 00:24:34.840 +mocking library for what I needed. + +00:24:36.900 --> 00:24:37.400 +And I prefer that at the moment. + +00:24:40.560 --> 00:24:41.060 +[Speaker 3]: I see, that is interesting. + +00:24:42.500 --> 00:24:42.700 +Because I had seen it, + +00:24:45.440 --> 00:24:45.600 +but I didn't consider that it's gonna be + +00:24:46.800 --> 00:24:47.300 +useful even in simple cases. + +00:24:52.640 --> 00:24:53.140 +[Speaker 0]: It has its limitations. + +00:24:58.260 --> 00:24:58.760 +So it's like life, how you turn depends. + +00:25:03.740 --> 00:25:04.020 +But maybe I should look more into the org + +00:25:05.880 --> 00:25:06.100 +mode and the test case to learn more about + +00:25:07.480 --> 00:25:07.980 +that. So thanks for pointing that out. + +00:25:14.620 --> 00:25:15.120 +[Speaker 3]: We are trying to cover as much as we can. + +00:25:17.520 --> 00:25:17.740 +It's almost impossible for org. + +00:25:20.500 --> 00:25:21.000 +But yeah, we keep adding more tests. + +00:25:22.780 --> 00:25:23.280 +[Speaker 0]: That's great. + +00:25:52.720 --> 00:25:53.200 +Someone's typing. I don't know. + +00:25:54.340 --> 00:25:54.840 +Any more questions? No? + +00:26:01.060 --> 00:26:01.560 +Okay, then I'll go back and try to document + +00:26:05.200 --> 00:26:05.360 +this in the etherpad. Thank you everybody for + +00:26:08.860 --> 00:26:09.160 +[Speaker 1]: Thank you guys. Great work. + +00:26:09.400 --> 00:26:09.900 +[Speaker 0]: joining. Great. Thank you. + +00:26:11.100 --> 00:26:11.600 +Take care. Bye-bye. + +00:26:15.060 --> 00:26:15.560 +[Speaker 1]: Take care. Bye. Silence. |