Positive test case how?

Question

Software testing

I'm doing an internship in the field of testing, rummaged through a lot of sites, asked around people, but the more information, the more I get confused.
There is a site todolist.ru for which you need to create test cases. Using equivalence classes and boundary values ​​at least once when compiling test cases and when finding bugs, issue a bug report. There is no more information (no technical specifications and criteria, just "test completely")
The words positive equivalent and boundary value led me to a dead end.
How to create at least one test case, please show an example.
As for equivalence and boundary value, please explain too, I read that this is for the input data, but sometimes not only.
For example, the fact that on each page of the site, there is a button to return to the main page, is this equivalence? The registration form is there without prompts, I myself found out that the login is 1-50 characters, and the password is 1-128, is it worth specifying boundary values, 51 characters in the login, when compiling a positive test case?
And what does this test case look like?
Do I need to indicate the actual result when compiling a test case? If it converges / does not converge with the expected? Or just indicate the OR, and all discrepancies in the bug report.
And the last thing, if I found a bug without a test case, I just crawled around the site and found a grammatical error, should I create a special test case for this error? Or you can immediately in the bug report?
Thank you for your attention and understanding)

Answer 1

Everything is mixed into porridge.
I think you need to thoroughly read on the theory of testing and figure out where and what.
Boundary values ​​and equivalence classes, as well as examples of their application, are googled for "Test Design Practices". Well, or in any book on software testing.
In short: this is about the fact that there is no point in sorting out various parameter values ​​that lead to the same result and you can collapse them to a minimum of checks.
Eg. if your valid date of birth is 1910 - 2000 -> it makes no sense for you to iterate over all values ​​from -MAX_INT to + MAX_INT, you just need to check >1910, 1910, 1911 - 1999, 2000 and > 2000.
Because in essence you have 3 classes of equivalent cases: the user enters a date of birth too early, the user enters a date of birth within the allowable range, and the user enters a date of birth too late.
Plus, a check of the boundaries of these intervals is added to this - i.e. specific cases at the boundary of each class.
This allows you to determine that the class boundaries pass exactly along the values ​​you need (1910 - 2000), and not earlier / later.
Regarding the classification of tests for positive and negative.
Everything is simple here. Positive cases test the performance of scenarios with the expected behavior of the user.
Eg. entered a valid email and password, they match those of the previously registered user -> authorization was successful.
I clicked the "Reset password" button -> an email with a password reset was sent to the mail.
Etc.
Negative scenarios cover situations where the user deliberately or accidentally does something wrong - from validating values, to duplicating requests, all sorts of injections and more.
Eg. case "Transfer money from the card -> Press back in the browser -> Transfer again" will be negative -> it is necessary to check that the previous state of the page (in which the money has not yet been withdrawn from the card) will not allow the user to withdraw more than he has.