Timeline for Why are reproduction steps helpful for fixing software bugs?
Current License: CC BY-SA 4.0
47 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 12, 2021 at 15:33 | review | Close votes | |||
| Jul 17, 2021 at 3:09 | |||||
| Jul 12, 2021 at 15:10 | comment | added | haylem | Because bugs are not just about the software, but also have behavioral and environmental causes. For instance, if you didn't ask for steps to reproduce, you'd never be able to fix this: support.displaylink.com/knowledgebase/articles/… | |
| Jun 22, 2021 at 6:00 | history | tweeted | twitter.com/StackSoftEng/status/1407216782465769472 | ||
| Jun 18, 2021 at 20:25 | history | removed from network questions | ChrisF♦ | ||
| Jun 18, 2021 at 20:12 | comment | added | Michael Schumacher | @Chris I wonder how the support team can provide good service with that attitude. Shouldn't they be interested in getting the same kind information from whomever they are supporting, in order to reproduce issues before forwarding them? Or is this a team that forwards issue reports unverified? | |
| Jun 18, 2021 at 16:41 | comment | added | pacoverflow | @Kaz Agreed. At my company, it would be frowned upon for a developer to just close a bug as "not reproducible" without talking to the reporter. They would be seen as lazy. They are expected to work with the bug reporter to find out how the bug can be reproduced. | |
| Jun 18, 2021 at 15:34 | answer | added | Kaz | timeline score: 0 | |
| Jun 18, 2021 at 15:33 | comment | added | Michael | @IsaacCWay: Without reproduction steps (or an equally good description of the behavior) how do you even know there is a bug in the first place? There is a reason why most bug trackers ask for reproduction steps, expected behavior and actual behavior. | |
| Jun 18, 2021 at 15:08 | comment | added | Sabine | My first car had an issue that never got solved. My guess is that this was - at least to some extent - due to the fact that the mechanics never "reproduced" the issue that only started after driving for some 20km. OTOH, if I hadn't told them, they would never have known there was an issue... | |
| Jun 18, 2021 at 15:06 | comment | added | Kaz | If you can legitimately "mark it as unable to reproduce", it means the bug has repro steps. Marking a bug closed as "unable to reproduce" because it lacks repro steps is a passively-aggressively silly and inefficient way to teach the bug submitters to provide repro steps. You add a comment "how does this happen? Can you provide detailed steps to reproduce it, in the appropriate feld?" The leave the bug in the "UNCONFIRMED" or "NEW" triage stage until the details are given. | |
| Jun 18, 2021 at 12:49 | comment | added | ICW | It's literally impossible to debug something if you can't reproduce it. Stop thinking about analogies as they are often misleading (especially when comparing cars to software development), and think about the actual situation: How can I know if something is fixed if I have no way to measure whether or not my changes stopped the bug from occuring? | |
| Jun 18, 2021 at 12:29 | comment | added | Paul D. Waite | I feel like the analogy has distracted everyone from the actual, really good question being asked. | |
| Jun 18, 2021 at 12:02 | answer | added | Nathan | timeline score: 0 | |
| Jun 18, 2021 at 11:47 | history | protected | gnat | ||
| Jun 18, 2021 at 10:54 | comment | added | Thorbjørn Ravn Andersen | "We need to be certain that we see what you see, in order to be able to verify that the problem that goes away is your problem." | |
| Jun 18, 2021 at 10:30 | comment | added | MathematicalOrchid | When you go to the doctor, do you just say "My body isn't working correctly. Fix it!" and expect them to have the vaguest clue what exactly isn't working? No, of course not. ;-) | |
| Jun 18, 2021 at 10:03 | comment | added | Michael | Without reproduction steps, how else is the developer supposed to know what exactly is not working? If you bring your car to the garage and just say “it’s not working.” what are they supposed to do? | |
| Jun 18, 2021 at 9:15 | comment | added | Pieter B | I shouldn't have to spell out to the mechanic that my car makes odd noises the mechanic lubricates the squeaky door but not the rattling of the engine when going 70 on the highway........"what is an odd noise?" | |
| Jun 18, 2021 at 8:50 | answer | added | David | timeline score: 0 | |
| Jun 18, 2021 at 8:42 | comment | added | MPS | It helps knowing on the spot if the software was used in an unintended way. | |
| Jun 18, 2021 at 8:22 | comment | added | Qwerky |
if I take my car to the garage, the mechanic will look at the engine for me to diagnose and fix the problem, I don't need to know how my engine works, and I shouldn't have to spell out to the mechanic that my car makes odd noises - How does the mechanic know to start looking at the engine? What if she spends a couple of days investigating the electrical system, or starts stripping down the brake system? Would you be happy with that? Remember they're billing you by the hour. With no information as to what is wrong you might get a really big bill.
|
|
| Jun 18, 2021 at 7:53 | comment | added | Aulis Ronkainen | Reproduction steps are helpful because once fixed, it'll be easy to check if the bug is really fixed. Bugs can be fixed without the steps, but they surely do help. | |
| Jun 18, 2021 at 2:57 | comment | added | abathur | FWIW, the mechanic doesn't work for everything (especially, amusingly, as cars accumulate software). My partner's had intermittent error codes on her in-warranty car for ~9 months, and taken it in at least 3x to two different dealerships and ended up with "we can't find anything wrong; bring it in while it is doing it." It finally happened when she had time a few weeks ago. They still couldn't figure it out. They replaced the battery, reset the codes, and asked her to bring it back after 100 miles. | |
| Jun 18, 2021 at 0:49 | answer | added | Kain0_0 | timeline score: 0 | |
| Jun 18, 2021 at 0:29 | comment | added | Seth R | An auto mechanic will absolutely drive your car around to try and reproduce those odd noises. On cars, being physical objects, it is also a lot easier to see when something is broken. Software is invisible. What the software does when we use it is the only thing we have to go on. | |
| Jun 18, 2021 at 0:11 | answer | added | Alexander The 1st | timeline score: 0 | |
| Jun 17, 2021 at 23:08 | comment | added | Jeremy Friesner | If I can't reproduce the bug before I make any code changes, and I still can't reproduce the bug after I've made some code changes, how can I prove (to myself or to anyone else) that my code-changes actually fixed the bug? | |
| Jun 17, 2021 at 23:00 | answer | added | Michael Kay | timeline score: 0 | |
| Jun 17, 2021 at 22:22 | answer | added | bta | timeline score: 5 | |
| Jun 17, 2021 at 21:45 | answer | added | Jordan Rieger | timeline score: 5 | |
| Jun 17, 2021 at 21:43 | comment | added | jwdonahue | As a developer, I am inclined to close as not reproducible, any bug that does not specify repro steps. Whenever possible, however, I reach out to the support tech's boss and get that support tech retrained or fired. I usually have plenty of well defined problems to solve already, so I really don't have the time to play guessing games, or twenty questions. It's the support tech's job to elicit as much information about the bug as possible. They may not be responsible for verifying that the steps do in fact reproduce the bug, but I expect to see at least a plausible list of steps. | |
| Jun 17, 2021 at 20:55 | comment | added | Owen Reynolds | On the old NPR show "Car Talk" with Click&Clack they always talked about intermittent problems and spent plenty of time asking "did it only happen at night? In the rain? 1st thing in the morning?" Listening to that show taught you why careful reproduction steps were important | |
| Jun 17, 2021 at 20:38 | comment | added | Jordan Rieger | Good auto mechanics absolutely do ask for reproduction steps when diagnosing intermittent or obscure issues. | |
| Jun 17, 2021 at 20:20 | answer | added | Steve Summit | timeline score: 0 | |
| Jun 17, 2021 at 19:47 | history | became hot network question | |||
| Jun 17, 2021 at 14:18 | answer | added | gnasher729 | timeline score: 72 | |
| Jun 17, 2021 at 14:11 | vote | accept | Chris | ||
| Jun 17, 2021 at 13:50 | answer | added | Cap Barracudas | timeline score: 12 | |
| Jun 17, 2021 at 13:42 | answer | added | Doc Brown | timeline score: 6 | |
| Jun 17, 2021 at 13:42 | comment | added | JimmyJames | The mechanic analogy is just wrong. Here's a counter example: I have a car which has a battery that goes dead. I jump it, take it to the shop and they see no problem. That's annoying but imagine I dropped it off with no explanation. They wouldn't know if I had engine problems, issues with suspension or tires, or maybe I just wanted an oil change. I actually do drop off my car in their lot off hours and guess what, they don't even look at it until I call or drop by to tell them what I need. | |
| Jun 17, 2021 at 13:40 | answer | added | Karl Bielefeldt | timeline score: 85 | |
| Jun 17, 2021 at 12:37 | comment | added | gnat | if you are mostly focused on this then duplicate target question would fit better: Getting users to write decent and useful bug reports | |
| Jun 17, 2021 at 12:35 | review | Close votes | |||
| Jun 22, 2021 at 3:07 | |||||
| Jun 17, 2021 at 12:30 | history | edited | Chris | CC BY-SA 4.0 |
Added note on why other tickets do not answer the question
|
| Jun 17, 2021 at 12:20 | answer | added | Deduplicator | timeline score: 18 | |
| Jun 17, 2021 at 11:48 | review | First posts | |||
| Jun 17, 2021 at 13:24 | |||||
| Jun 17, 2021 at 11:47 | history | asked | Chris | CC BY-SA 4.0 |