On Sat, Apr 15, 2017 at 02:24:41PM +0200, Michal wrote: > Hi, > Maybe my email wasn't nice enough, but breaking compilation
You were the first one to experience the build breakage, it worked for most of us, but you didn't even give the smallest hints about the error(s) you met, making it much harder to fix the problem. Fortunately others were much more constructive and reported it the normal way. > and simplest > config with server using "source" got me very angry. I didn't send any > reproducer, because even simple > "server name 1.1.1.1:80 source 1.2.3.4 track other" > wasn't parsing. It's much easier when the proper info is sent, you can imagine as many test combinations as you want, you'll always find one that does not work. I tested several configs myself before merging Fred's patches and all of them worked so I was confident enough. In this project we try very hard not to break existing setups. Some configs have never been modified since version 1.1 emitted 16 years ago and still work fine. We even implemented emulation for some older settings that are not natively implemented anymore, so we know pretty well that breaking setups is not nice. But this is a development branch and during development it's expected that things break from time to time, this is why nobody is supposed to run this in production, and everyone is welcome to report issues. > I'm posting this patch as open source committer, private person with private > e-mail address and let's be honest - @haproxy.com guy didn't check > simplest and very important thing as compilation on different platforms. > Even -dev branch shouldn't break such thing. Please stop saying that bullshit. Tests were made and I tested myself. It's just that you had the first config revealing a regression and you would have been welcome to report it as anyone else usually does here. If you're too narrow-minded to stand a development cycle, please go pollute another project, we don't need people who complain against those who do the work. And feel free to call me a clueless incompetent bastard if that makes you feel better. Let me tell you something : I couldn't have brought this project where it is alone without all the contributions we received, so I do have a lot of respect for contributors and their work and I will not let anyone insult those who devote time to improve the project. And that's true for your contributions as well. If there's something you don't like, propose improvements but do not tell people they're doing a bad work. I hope this is clear. > Please don't mess allegrogroup here, I'm posting those patches from my > private e-mail. This is the "From" you put in the commit author. As you see it's never nice to see your work criticized when you thought you did a nice job and it has an impact on others, and constructive bug reports are generally nicer for everyone than broad accusations. Now let's turn constructive again and try to find how to deal with this weight propagation issue. Willy
