Why would you not differentiate member variables?

Another post from the llvm mailing list, this time on the topic of member variable naming conventions.

I remember hacking away on some llvm code and having a terrible time with this. It can also lead into dangerous ambiguities when you use local variables that actually hide member variables. The whole naming convention they use is insane. I wonder if they use any static analysis tools?

Holacracy

Someone pointed me to this article on Holacracy which is an interesting approach to organisation. You effectively form 'circles' of people who are focused on an area and it is composed of diverse sets of people. You are on multiple circles at once and move around. Seems like an interesting idea but a little anarchic.

Internals of Firefox Threading

Interesting post on the internals of firefox's new multithreading 

Mega repo or microrepos? Llvm-dev take

There is an interesting discussion going on at the llvm-dev mailing list as to whether to split their massive svn repo into multiple smaller repositories with git.

http://lists.llvm.org/pipermail/llvm-dev/2016-July/103097.html

I have come across the problem so many time and the issue ultimately boils down to workflow: tools really don't support this. I want to build the modules individually and then sometimes for integration all together. If there is a dependency between repositories I sometimes need to express it.

Surely there is a better way.