Hacker Newsnew | past | comments | ask | show | jobs | submit | InformalRelief's commentslogin

OP argues that the presence of the 'Pull Requests' tab is encouraging people to create pull requests, and therefore the project owner is somehow responsible for reacting to pull requests.

There's also a 'Fork' button, which is kind of prominent, so by OP's logic, people should feel encouraged to fork the project.

"But when I looked at the mailing list for the project, I saw a wasteland of good patches that were completely ignored, where the submitter would ping the list a couple times and then give up. Did it seem worth spending a week to disentangle our IP from the project in order to submit a set of patches that would, in all likelihood, get ignored? Of course not.

If you have commit access to a project that has this problem, please own the process for incoming pull requests." (from the article)

Ok, so if the mailing list is dead and your pull requests aren't the only ones being ignored, why not just fork the project and 'own the process' yourself? You could even try to contact people who tried to contribute and got ignored, and let them now that you will 'own the process for incoming pull requests'?

And if you don't have time or energy for that, why demand it from other people?


> Ok, so if the mailing list is dead and your pull requests aren't the only ones being ignored, why not just fork the project and 'own the process' yourself?

That's not necessarily straightforward even for experienced OSS contributors, but in the given context (newbies who have been encouraged to get into OSS by contributing a patch), it's really a non-starter.


Because Forks aren't easily discoverable (discussed elsewhere on this thread as well), the fork would be unlikely to gain critical mass, especially without an official sanction and prominent shoutout by the original repo. If the maintenance has truly lapsed, the owner may be too lazy to even tell people about the fork. This is backed up by the fact that they obviously haven't anointed a successor, as it were.


To that end, the author also suggested that it may serve the project (and subsequent forks) well if the project owner disabled PRs.


You can not disable PRs on github. The wiki yes, issues yes, but it's not possible to disable PRs, that simply isn't an option.

At best you can add a note/contribution guideline saying you don't accept PRs, and close them as soon as they're created, but it's not possible to prevent them.


Which would remove the most obvious place to look for commits for this new fork, or for this patch that you were just about to write.


Looks interesting. Is this useful if you work with relatively new code (no Cobol lurking in dark corners)?


In the book, Feathers defines legacy code as code without tests.

So it's not really about old languages or old code. It's pretty easy to write (or inherit) new, difficult to test code, with no tests, in whatever language you please.


I've been using the step-through-with-a-debugger approach, but what I'm missing are tools for visualisation. Drawing boxes gets way too complicated way too quickly. I'd like to have something like a UML diagram, but as far as I know UML can't really be used for that purpose.


During college I had a job in a coffee shop, and one day after a slow afternoon the bossman came in and later called me over to his computer, where he was reviewing his CCTV footage, and said "Look at how long it took you to make this cappucino, this is simply unacceptable!". I walked out.


Now imagine if your family relied on you having this job? How awful would that be.


Yup. And to be honest, the surveillance was not even the worst part, the guy was just generally a shitbag. But yeah, considering all the scandals and such involving corporations spying on employees, it's probably safe to say that it's mostly the most vulnerable people who have to put up with it. But IT freelancing is probably too much of a niche for there to be any public outrage about stuff like spying on contractors.


Much more interesting? Mathematically more interesting!


I like that it has a physical keyboard, what I can't stand is the resistive touchscreen.

Back in the day, I would have totally gotten an n900 if it weren't for the screen (I tried it, just didn't work for me.)


But tags. If I type in [android] in the stackoverflow search box, it gives me all questions tagged android. This doesn't work if I use google.


Yes, devices should always do exactly what you mean them to do.


I sometimes take a friend's phone from them and hand it back to them unlocked. Just for the look on their faces. Until they figure out that I just watched them unlock it.


Can somebody here ballpark how much this would cost if you rented servers instead of using a botnet?


The problem with renting is that your provider will probably shutdown you, if your scan rate is high. At least that is the problem with large dedicated server hosters.


Funnily, the post inmediately above you (posted a bit later) mentions:

"I had to spend USD15 to run 8 medium EC2 instances for 16 hours to only find 200++ million hosts"

https://news.ycombinator.com/item?id=5406233

and another study was done using EC2 instances:

https://news.ycombinator.com/item?id=5405483


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: