![]() Instead you could keep the history in a separate branch. Ii) There is only a little gain in preserving the history only to be able to use it for git bisect in the future. Your initial advice some 6 months ago was very good, and with it I made it to "work in my branch and contribute to master", but now, with the whole getting a little bit more complicated, I would very much welcome some more light. Then I totally forget about branch sk, push sk_new to github and ask all my pals to use from then on sk_new, so that merges and rebases do not get mixed? Git rebase -i -onto master CommitPreviousToTheCommitWhereMyBranchDivergedFromMasterTheLastTime ![]() Person X makes some modification based on my branch (lets assume my branch is called sk). I do not know, because I can not remember, if I just merged master into my sketcher_testing branch (which now sounds like a terrible idea), or something else happened.Ī. Is this ok? There is something else I have to consider?Ģ) How should I integrate other persons commits, so that this mess does not happen again. These I would prefer to maintain, for copyright and attribution reasons. I have commits from other authors that contributed to my branch. My question goes in the sense: if I squash changes in 200 files in a single commit and we then try to use git bisect to know where it failed, having a such a squashed commit might make it worse to detect the problem. Do each of this different features deserve a separate commit?. The status of each individual commit is "it compiles", but probably will not run stable if run (there will be bugs fixed in other commits). This I have no doubt (unless you tell me otherwise), I squash. I have several commits that are simply an evolution of the same "feature" within the ellipse implementation. This is more of "the sense of a commit" than the how to use git. Now, if I may, I would ask you for more advice.ġ) How should I squash the ellipse implementation In fact I have reproduced it in my branch. Now that you tell me what you have done (which I very very much appreciate), I can follow. ![]() If I would be younger I would say something like when I grow up I would like to be like you. I am thinking on rebasing the result of 2, by changing the order of the commits, so that all the commits corresponding to master go first, and all the ones of the ellipse go second.įrom your deepest git expertise, do you have some advice that could help me get out of this mess? However, I have once again a plurality of interleaved commits.ģ. In a desperate attempt to see if something would work, I did a temporal copy of an updated master (FreeCAD master), and did a: I have resolved one after the other, but end up missing some changes from master (I have compilation errors because some changes have not been integrated, i.e. So I did, in a copy of my sketcher_testing branch:
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |