FReq: `wstool merge` should perform init if .rosinstall doesn't already exist.

[catskul]

rosinstall allowed initialization with a .rosinstall file

wstool init does not, and wstool merge will not function without wstool init first which itself only creates an empty file.

wstool merge should create the .rosinstall file if it doesn't already exist.

[wjwwood]

wstool init can take a rosinstall file to start, i.e. wstool init <dst> [rosinstall file]. But wstool init cannot take input from stdin like wstool merge can, which has bothered me in the past.

I agree that something should be changed, and I think either having wstool init take from stdin or making wstool merge work even before there has been a call to wstool init might be good options.

[tkruse]

The code from merge can easily be reused for init. The code to be extracted into a common function is in multiproject_cli.py. +1 for that.

A reason wstool merge does not create a .rosinstall if none exists is to protect the user from accidentally invoking wstool merge in the wrong directory, which can be very confusing for users further on given the .rosinstall file is hidden.

wstool was designed following the ideal that ROS should be easy to use for people new to Linux, version-control, the shell and programming in general. This philosophy is intended to help growing a large community for ROS such that a new generation of maintainers can be fostered to pick up existing packages for maintenance.

Not everybody wants to follow that philosophy.

[wjwwood]

A reason wstool merge does not create a .rosinstall if none exists is to protect the user from accidentally invoking wstool merge in the wrong directory, which can be very confusing for users further on given the .rosinstall file is hidden.

That makes sense, I agree that it should probably not create an initial rosinstall file if one does not already exist in order to prevent users from accidentally creating a rosinstall file in the wrong place.

[dirk-thomas]

Closing since the repository is about to be archived: see #154.