Pages: 1 2 3 4 5 6 7 8 9 10 11
I meant serious PRs, not crappy ones. This is very easy to catch at a glance. No need to go 'under-the-hood' to know which PR is crap and which isn't.
Also - I do not mean abandon code-testing; I mean to make user responsible for it (best option would be to make them add unittests to each PR). PR with tests = merge; PR without tests = refused. And write sth like Developers code of conduct, where you would make it clear whats required to be in PR. Relatively easy to do.
We already told you how difficult it is to set up unit tests for the 1.8 codebase, and how glad would we be if you wrote them. That's what Open Source is about, people contributing to a project with a team's supervision. Even the most serious PRs have to be tested for the aforementioned reasons.
Development coding standards are already there.
(2017-09-16, 10:05 AM)devs Wrote: [ -> ]I meant serious PRs, not crappy ones. This is very easy to catch at a glance. No need to go 'under-the-hood' to know which PR is crap and which isn't.
Also - I do not mean abandon code-testing; I mean to make user responsible for it (best option would be to make them add unittests to each PR). PR with tests = merge; PR without tests = refused. And write sth like Developers code of conduct, where you would make it clear whats required to be in PR. Relatively easy to do.
Unit tests test units of code, not complex data flows. Not user interface experiences. Also, we'd need unit tests in the entire codebase to be able to do that - which we don't have. Rewriting unit tests into every function within 1.8 would a waste of time, when testing PRs manually would be faster. We could abandon 2.0 and spend the next year writing unit tests - but then there would be no updates. How is this better?
(2017-09-16, 10:19 AM)Shade Wrote: [ -> ]We already told you how difficult it is to set up unit tests for the 1.8 codebase, and how glad would we be if you wrote them. That's what Open Source is about, people contributing to a project with a team's supervision. Even the most serious PRs have to be tested for the aforementioned reasons. Development coding standards are already there.
its not difficult at all......
Refer to
this
(2017-09-16, 10:27 AM)Tom K. Wrote: [ -> ] (2017-09-16, 10:05 AM)devs Wrote: [ -> ]I meant serious PRs, not crappy ones. This is very easy to catch at a glance. No need to go 'under-the-hood' to know which PR is crap and which isn't.
Also - I do not mean abandon code-testing; I mean to make user responsible for it (best option would be to make them add unittests to each PR). PR with tests = merge; PR without tests = refused. And write sth like Developers code of conduct, where you would make it clear whats required to be in PR. Relatively easy to do.
Unit tests test units of code, not complex data flows. Not user interface experiences. Also, we'd need unit tests in the entire codebase to be able to do that - which we don't have. Rewriting unit tests into every function within 1.8 would a waste of time, when testing PRs manually would be faster. We could abandon 2.0 and spend the next year writing unit tests - but then there would be no updates. How is this better?
Unittests are for code testing as the name suggests, right.....
Why you dont have unittests in 1.8's codebase? Because someone
didnt think of it, and didnt add this as a requirement for devs in Development Guidelines.
Rewriting of 1.8 would be waste oif time? Improving product is waste of time? Are you being serious?
Rewriting 1.8 would be a waste of time - because that's precisely what 2.0 is intended to be.
You keep making an argument, the rebuttal and answer is always the same. Not MyBB or my fault if you don't like the responses.
(2017-09-16, 10:46 AM)Ben Cousins Wrote: [ -> ]Rewriting 1.8 would be a waste of time - because that's precisely what 2.0 is intended to be.
You keep making an argument, the rebuttal and answer is always the same. Not MyBB or my fault if you don't like the responses.
If so, whats the point in wasting time in making product better by improving code?
Oh don't get me wrong, writing unit tests isn't hard. It's writing unit tests for a large and out of date codebase that is hard. Yes, nobody thought to write unit tests in 2004, at the time the project wasn't open source and I don't believe there were even development guidelines.
What you're essentially suggesting is we stop developing 2.0 (which is intended as a modern rewrite for 1.8 with unit tests) and rewrite a 15 year old codebase to include such tests. Without a considerable community effort to assist this (of which there has not been one) that would essentially mean no updates until the unit tests were put in place. This would likely mean no other development for maybe a year. Spending all that time on 1.8 just to make the testing of PRs simpler is ridiculous. You have to look at this from a cost-return point of view. Unit tests in 1.8 would be great, but not at the cost of actually fixing bugs. Therefore, they are not a priority. If you want to add them, go ahead, we would be very grateful.
2.0 is our rewrite of 1.8. It is the adding of unit tests. It is the cleanup of features that have been built upon and hacked and built upon. It is open source. It can be contributed to. The whole reason we opened MyBB up to be an open source project was to facilitate community contribution.
(2017-09-16, 11:11 AM)devs Wrote: [ -> ] (2017-09-16, 10:46 AM)Ben Cousins Wrote: [ -> ]Rewriting 1.8 would be a waste of time - because that's precisely what 2.0 is intended to be.
You keep making an argument, the rebuttal and answer is always the same. Not MyBB or my fault if you don't like the responses.
So if this is what 2.0 is meant to be, than when will it be released?
When it is completed? Which really depends on community and developer contribution...
(2017-09-16, 11:12 AM)Tom K. Wrote: [ -> ]Oh don't get me wrong, writing unit tests isn't hard. It's writing unit tests for a large and out of date codebase that is hard. Yes, nobody thought to write unit tests in 2004, at the time the project wasn't open source and I don't believe there were even development guidelines.
What you're essentially suggesting is we stop developing 2.0 (which is intended as a modern rewrite for 1.8 with unit tests) and rewrite a 15 year old codebase to include such tests. Without a considerable community effort to assist this (of which there has not been one) that would essentially mean no updates until the unit tests were put in place. This would likely mean no other development for maybe a year. Spending all that time on 1.8 just to make the testing of PRs simpler is ridiculous. You have to look at this from a cost-return point of view. Unit tests in 1.8 would be great, but not at the cost of actually fixing bugs. Therefore, they are not a priority. If you want to add them, go ahead, we would be very grateful.
2.0 is our rewrite of 1.8. It is the adding of unit tests. It is the cleanup of features that have been built upon and hacked and built upon. It is open source. It can be contributed to. The whole reason we opened MyBB up to be an open source project was to facilitate community contribution.
If thats what is needed
its not ridiculous. I would call it natural neccesity. MyBB has vast community, so there is no risk of the
entire community going away - of course, some will surely go away, but majority
will stay.
(2017-09-16, 11:14 AM)Tom K. Wrote: [ -> ] (2017-09-16, 11:11 AM)devs Wrote: [ -> ] (2017-09-16, 10:46 AM)Ben Cousins Wrote: [ -> ]Rewriting 1.8 would be a waste of time - because that's precisely what 2.0 is intended to be.
You keep making an argument, the rebuttal and answer is always the same. Not MyBB or my fault if you don't like the responses.
So if this is what 2.0 is meant to be, than when will it be released?
When it is completed? Which really depends on community and developer contribution...
What do you mean by this? There is no guideline?
No it IS ridiculous when we are already doing that with 2.0. Adding unit tests to 1.8 makes merging PRs easier but doesn't address the fact it's a 15 year old codebase that follows few modern programming conventions. Rewriting (ie: 2.0) adds unit tests AND updates the whole codebase. Which is better? You say people would stick around but if we said we're not going to fix anything for upwards of a year we'd get a lot of complaints. The latest update is a little behind where we expected and we got this thread.
Is there a timeline? Not really. How can we know? We can't know how long devs will have to contribute, we can't know if we're going to get tons of help from the community or none. I could give you a date if you like but it would be completely arbitrary.
Pages: 1 2 3 4 5 6 7 8 9 10 11