I’m wondering is it possible to do following “stupid” thing required by my current project and asking for your help in pointing me towards right solution.
The source code tree consist of let’s say 4000 of text files. These files a standard, the base, I need to keep them, but generally don’t want to add them to compilation process — because it will last forever. Only 500 of them are part of my custom solution and the team modifying them very often. These files need to be sent to recompilation, let’s say every hour.
We’ve used tool for creating combined file for all 500 files we need. It searches for given pattern in all files and creates resulting file — big text object where all source files a stored one by one. The same behavior as you have by running copy *.txt all.txt on Windows.
The question is is it possible to achieve same behavior with git grep only? Grep is fast and we’re using git already.
Maybe are there any other strategies to have store two overlapping sets of files in git and treat them differently without custom tools? Branches and tags can’t help as I see it…
The obvious answer was: to create own tool which will integrate seamlessly to the git, since git allows creation of its commands.
The example implementation could be found at https://github.com/shytikov/git-nav/