>>> "KK" == Konstantin Khomoutov <kos...@bswap.ru> writes:
First of all thanks very much for your time and patience to explain git.
I appreciate it. Truth being told, most software is today developed in
git not mercurial[3], facebook being the only exception, I can think of.
So even with the hg-git plugin, knowing the basics of git seems a smart
thing, besides, I am afraid if facebook switches to git, mercurial will
fall in the abyss of obligation[1]
> I surely would not want all those 190 branches live in "my own" space in the
> repository - that'd just been too much of a cognitive burden which would
> require me to invent certain workarounds.
I see. I have to admit that the repositories I access and work with have
around say 3, 4 branches. The only exception is GNU Emacs, that I
usually access via mercurial and the hg-git converter, indeed this
repository has a bewildering amount of branches that are converted to
bookmarks.
Now I think I still quite some gaps in my understanding of git,
especially remote branches. Let me try to resume (and maybe draw some
unhelpful analogies with hg).
1. Scenario: I connect with one remote repository: A remote branch
allows me to obtain/control of a remote repository I am working
with. Well I have git remote -v (hg path), to know which remote
repository I am using. But, in hg, as you mentioned each commit
has more meta information stored, besides the branch name[4],
phases (a commit can be secret, so hg push will not push it,
draft, can be pushed, public is already pushed or pulled), so
remote branching is a way of obtaining information about the
branch on the remote.
Two questions
- how do I know that I have pushed already a local commit?
(Besides phases in hg, I can run hg out, to see what mercurial
would push).
- Say I I would have a couple of commits in say two
local branches. How can I push them both? Or must I push for
each branch? (in mercurial, the default is to push all branches,
the -b option allows you to push only certain branches, I have
the feeling that git is here, hm orthogonal.)
2. Scenario: I connect to various repositories (I have done that in
the past for pushing to a main and to a sandbox repository, but I
have never ever pulled from various repositories, Bob and Alice
in your example and have no idea, what would happen if Bob and
Alice have a main, feature and bug-fix branch in their respective
repositories and I would run «hg pull github» and «hg pull
bitbucket», I presume a big mess[2]. So in git that seems
possible
And I presume presume git branch -a would give me
* main
remotes/github/HEAD -> github/main
remotes/github/bug-fix
remotes/github/feature
remotes/github/main
remotes/bitbucket/HEAD -> bitbucket/main
remotes/bitbucket/bug-fix
remotes/bitbucket/feature
remotes/bitbucket/main
But,
git checkout main
Seems not a smart idea, it should be
git branch feature-github github/feature # create local branch off a remote one
git checkout feature-github # check local branch out
git branch -u github/feature # make current branch track origin/feature
And
git branch feature-bitbucket bitbucket/feature # create local branch off a
remote one
git checkout feature-bitbucket # check local branch out
git branch -u bitbucket/feature # make current branch track origin/feature
Correct?
regards
Footnotes:
[1] I am strongly reminded on the Xemacs/Emacs situation. I was a
convinced Xemacs user, till, well the dev team basically gave up.
[2] there is however an extension for remote branches, not sure it
still words
[3] although this is not relevant here, but git strongly reminds me on
Plain TeX, flexible, but rather complicated, while mercurial
reminds me on LaTeX, quite structured but most likely more rigid
than git.
[4] and these name branch are an advantage: you know precisely which
commit belongs to which branch, but also a disadvantage, deleting or
renaming a named branch is almost impossible
--
You received this message because you are subscribed to the Google Groups "Git
for human beings" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to git-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/git-users/8735ahmito.fsf%40mat.ucm.es.
smime.p7s
Description: S/MIME cryptographic signature
