Closed robmaz closed 6 years ago
@robmaz - this PR is too big for being able to review. As I told you, I think that most of the small changes (like removal of binaries and other files), should be done step by step to be sure that there isn't any incompatibility problem. Like that, review will be easier.
I will submit a couple of PR showing the concept of splitting removal changes into sub-units that are easier to review. We can come back to this PR later, or you can split it in other ones...
I think we cannot really change the filesystem layout one file at a time, because this will take forever and will also be much more confusing than just replacing one layout with another. Of course the PR is big because there is all this binary crap in executables/ and Linux_executables/, which are just being renamed to libexec/MacOS/ and libexec/Linux/, respectively. The former bin/ becomes lib/distmap/ and the executables in the root folder now go into bin/. bin/distmap lines 9--11 refer to the new module location.
That is the extent of the PR and I think it is not a large PR, conceptually; only git unfortunately appears to see this as the deletion and addition of a million files.
Cheers Rupert
2017-11-30 16:01 GMT+01:00 Daniel Gómez-Sánchez notifications@github.com:
@robmaz https://github.com/robmaz - this PR is too big for being able to review. As I told you, I think that most of the small changes (like removal of binaries and other files), should be done step by step to be sure that there isn't any incompatibility problem. Like that, review will be easier.
I will submit a couple of PR showing the concept of splitting removal changes into sub-units that are easier to review. We can come back to this PR later, or you can split it in other ones...
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/robmaz/distmap/pull/18#issuecomment-348213519, or mute the thread https://github.com/notifications/unsubscribe-auth/Ad_FfPOeQ6YN4X4gWam0qFTq9KqFgcF7ks5s7sMvgaJpZM4QtK9G .
The problem are the binary files. But I suggest to either
The idea is (by commit/PR):
bin
folder to a lib
folder. this will be simple for git to recognize as a folder re-naming (there is no more binaries after #21)bin
. Generate a new folder and move a couple of scripts is really easy for git to realize.libexec
folder with the jar
files in its root (Linux and MacOS should have the same, and it is replicated in the current PR). This will be more changes for git, but there is less amount of files changed of directory.I think that it is much better to do 1 and 2 in two different PRs for easy review and no blocking the merge, and then 3 and 4 in a super big PR with lots of commits to easy review.
Let me know what do you think about that approach...
I was hoping to get your opinion on the proposed layout, which, I think, you can easily see if you look at the branch. I am not sure what the added value would be if I now recreated the branch in a slightly different way.
@robmaz - I can give you the feedback about the proposed layout: looks good but I have a couple of comments for some parts. The problem is that it is difficult to make the comments in this branch because the display is horrible. The added value for re-create the branch is:
As this is your repository and your project, it is up to you if you would like to do a more standard and principled software development. I can offer my help here: if you merge #21, I can make a PR with the re-name of bin to lib and another one of the scripts to bin. And then I can left you the libexec folder PR (that you can start without the merging of #21). Like that, we can work in parallel and changes are non-blocking others, making the final folder structure better and traceable...
The added value is: 1) improve visibility of the c
Ok, let's do it like that then. I appreciate your help, I just need a usable distmap soon, because I want to test it with the Wolbachia data.
Thanks @robmaz - and for an usable DistMap with trimming support, it will require a bit more work on issue #5 (I think that it is the blocking step).
plus a minor change in bin/distmap to find the modules again