On 02/25/2015 09:53 AM, Dan Smith wrote: ...
I think "drivers" is too generic a term. These are all radios. If we have another class of drivers, they should go in a different directory, named appropriately.
They are not all radios. There are storage formats in there, repeaters, and APRS units. They're all drivers, and "drivers" is my preference for the module name.
Okay, I did not realize this either. I am still learning. If you are looking down the road, will there be value in creating separate radio/, storage/, repeater/, arps/ directories (under chirp/, of course)? It would be nice to plan ahead so we don't ever have to do this kind of shuffle again.
Personally, I like this idea, as it would help newbies (like me) understand that there are different classes of drivers. Moreover, it becomes easier to make sure that changes to a certain class are effected across all relevant drivers (i.e. just be sure to touch everything in the appropriate directory).
Deleting and adding is how git tracks files, and mercurial uses it as the storage engine.
Why do you keep saying this? Unless something has changed recently in either tool, this definitely not the case. Mercurial uses revlog, which is similar in concept, but significantly different in practice.
Apparently my memory is not what it used to be. Mea culpa.
Use 'hg log --follow chirpc', and you'll see the history is all there.
When I "hg import" your patch, --follow does not show the history. When I "hg rename" it, it does. As such, I didn't apply your patch, did the rename myself, and committed it with your authorship, which is why the repository shows the correct history.
I see. I did not realize you did that you did that. I wonder whether it would have worked had I used -C, as I did for this patch.
If you actually used mercurial to try to apply your patch to a copy of the tree from before I applied the proper fix, you'll see the history is not maintained.
I will do a 'hg import' of my revised series on a clone of my hg repo and see if I can make it work. If not, I'll suck it up and try to send them from mercurial. Meanwhile, thanks for working with me to try and make sending from git a possibility.