Re: [git-users] change encoding and text replace

[Huu Da Tran]

On Wednesday, February 5, 2014 9:59:30 AM UTC-5, Philipp Kraus wrote:

 It is more a social problem and I don't want to fix encoding errors 
 manually.
 Mostly user does not configurate their editors correct and I get after a 
 commit
 encoding errors within the files, so I would like to convert it 
 automatically by git,
 in a best view OS independed.


For social issues, just reject the commit? 

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: how to check if my commit is propagated to later branchs?

[Huu Da Tran]

On Friday, December 27, 2013 9:03:31 AM UTC-5, Jagadeesh N. Malakannavar 
wrote:

 I have made some commits on rel-1. After some months I branched out rel-2. 
 How do I make sure that commits I made on rel-1 is in rel-2?
 Just want to list out all missing commits on rel-2.


Just look at the git log:

git checkout rel-1
git log
git checkout rel-2
git log

You should see the commit there.

Hope this helps.

HD. 

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: Merging from pull

[Huu Da Tran]

On Thursday, December 5, 2013 7:08:14 AM UTC-5, Phillip Tutt wrote:

 What I would expect is the code now on the desktop would be the same as 
 what was on the laptop...however, I found that some kind of merge was 
 happening and conflicts, but shouldn't it just be a fast forward?

Note, I did not make any changes on the desktop from the time I pushed to 
 the remote to the time I pulled back down, but even if I did, I would 
 expect the merge to work, but I don't have much faith in it at the moment 
 due to my limited experience.


Try to display your git log  for both your Desktop and Laptop and compare 
them.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: Merging from pull

[Huu Da Tran]

Please compare the log from laptop and desktop.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: Merging from pull

[Huu Da Tran]

On Friday, December 6, 2013 10:48:29 PM UTC-5, Phillip Tutt wrote:

 For sure.  This is where I was fumbling to try to get it all working:

 1. I set up a repository on the desktop, with the Git init here context 
 menu on the folder 
 2. I add code and commited:
   git add . --all
   git commit -m  some comment
 3. On thing to note is I don't have a master branch...'master' didn't make 
 much sense to me, I have a release branch and feature branches.  At the 
 moment I am working in the feature branch 'admintemplate' with the view to 
 merge it into the release branch when it is ready.
 4. I installed gitblit on my thecus NAS and set up a repository using the 
 gitblit 
 5. I created the remote:
 git remote add thecus url provided by gitblit
 6. I pushed the Desktop code to the Remote:
 git push thecus admintemplate
 7. I created a git repository on the Laptop using Git init here


This should have been a

git clone url provided by gitblit
git checkout admintemplate

Because this will make sure you have the same base commit to start before 
the pull (for next time). You then simply skip the next step (git add 
remote).
 

 8. I then added the remote to the Laptop (as in step 5. above)
 9. I think I just did a pull to get the code from the remote to the Laptop:
 git pull thecus admintemplate
 10. Modified code on the Laptop, pushed back to the remote.
 11. Performed a pull from the remote to the Desktop...and that's when I 
 got the merge conflicts:
 git pull thecus admintemplate

 Where I feel unsure is about tracking and cloning...I have read the basic 
 docs but my understanding is not deep enough to know exactly what is going 
 on under the hood.





 I set up the remote with the following:
 git remote add thecus 


Hope this helps.

HD. 

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

Re: [git-users] Re: git clean help

[Huu Da Tran]

Make sure directory is *really* empty.

ls -al Spoon-Knife

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: GIT project maintainer as a developer

[Huu Da Tran]

On Monday, November 18, 2013 2:29:41 PM UTC-5, Greg Freeman wrote:

 I have an issue with my workflow that I'm trying to resolve. I am a 
 developer as well as the project maintainer. We are using the forking 
 workflow so everyone has their own repo then I pull all their changes to my 
 copy to deal with the merge, then push these changes into the master.


All responses by Philip  should be taken into consideration.

When speaking of fork, that is something that was invented for service 
sites like Github or bitbucket where someone would make a clone online and 
that clone is linked to the original for easily make Pull Requests. Are you 
using such services. If so, there is another way so you would not need to 
have different repo names.

Hope this helps,
H.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: GIT project maintainer as a developer

[Huu Da Tran]

On Wednesday, November 20, 2013 10:42:28 AM UTC-5, Greg Freeman wrote:

 Huu you are correct. I left out something that is quite important. I am 
 using bitbucket/github depending on the project. This fork is through their 
 site. As far as I understand it, the benefit, besides easy pull requests, 
 is also so you can push your local changes to a server-side repo without 
 them being integrated back into the main code base. Then, if your hard 
 drive or something goes bad, your local changes that have been pushed to 
 your fork in the cloud are backed up. Am I correct in this understanding 
 also?


Yes, but the backup is also possible using any remote server to host that 
file. 
 

To answer your question, yes I am using a hosting site, and I'm interested 
 in the other way to avoid different repo names.


In that case, this is what we are doing. We have a corporate account which 
is the main repo (this is where the project should be created) and each 
developer then fork into their personal account. This could also have 
been possible to achieve without such service, but I'll not explain since 
it does not apply in your case.

Let's assume you have those accounts set up. As maintainer of a project, 
your would have creator/admin privileges (based on bitbucket).

Each dev would do the following:

git clone g...@bitbucket.org:your_account/project_name.git project_name
cd project_name
git remote add main g...@bitbucket.org:company_account/project_name.git

Then, when you code in your changes, you create feature branches according 
to your workflow (I'll assume develop branch).

git checkout -b your_initial/feature-name main/develop

This will create a tracking branch from the company develop branch.

you make your changes, commit them to your local repo... then you push to 
your bitbucket repo:

git push origin your_initial/feature-name

Now, you go on the web interface, and you create a Pull Request to the 
company account, with creating a new branch... by choosing compare and 
choose your account and the your_initial/feature-name on the left side, 
and then the company account and your_initial/feature-name (new 
branch...)   (not sure exactly, but i think it's the last one in the list).

Then, you put on the maintainer hat, and select the company account's repo, 
approve/merge the pull request. This create a new branch. We use this 
approach so developers can continue working on the feature if not 
completed, but as maintainer, we get a sense of changes. Next pull request 
will not be a new branch.  When everything is good and tested, you can 
merge it back to the develop branch by simply choosing compare, select 
your_initial/feature-name and develop, then click merge.

Let me know if anything is not clear, I'll try to formulate it better.

HD.


-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: GIT project maintainer as a developer

[Huu Da Tran]

On Wednesday, November 20, 2013 11:36:30 AM UTC-5, Huu Da Tran wrote:

 Then, you put on the maintainer hat, and select the company account's 
 repo, approve/merge the pull request. This create a new branch. We use this 
 approach so developers can continue working on the feature if not 
 completed, but as maintainer, we get a sense of changes. Next pull request 
 will not be a new branch.  When everything is good and tested, you can 
 merge it back to the develop branch by simply choosing compare, select 
 your_initial/feature-name and develop, then click merge.


I forgot to mention the following steps... Then you announce the merge or 
on a daily basis, each developer need to merge back those changes.

on the local machine:

git pull main develop

I like to pull straight from the remote so the merge commit message 
indicates the remote repo and not a local one. But they could also do

git fetch main
git merge main/develop

You will see changes from the develop branch... and also other new branches 
or updated branches following pull requests.

To go further, I usually delete branches from the fork to not confuse when 
doing pull requests. Since developer should not commit to master or 
develop, just delete them from the fork.

HD.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] A few GIT questions.

[Huu Da Tran]

Git is a tool to keep track of evolution of a file.

What you need is more of a content management.

Git can be used to track versions of all the files, but you probably want to 
have a website or other to actually tag (git tags goes on a specific state of 
all files in the repo, not to confuse with keyword tagging) and comment on 
files.

Hope this helps.

To go further, some questions to ask yourself is how are you doing it now, what 
are the problems you are trying to solve?

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] can I make two local repos be identical, always, except for current branch?

[Huu Da Tran]

You want to have all remote branches to be always sync'ed but not your local 
branches? So basically, you want git fetch to update both clones?

I am not sure it makes sense speaking of current branch in two different repos 
and keeping non-current branches sync'ed.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Easy question about merging

[Huu Da Tran]

If you really want to stay on the safe side, this would be the easiest and 
safest for you.

git checkout branchX
git checkout -b branchX-merge-master
git merge master

fix any conflicts, do your tests. Commit conflict changes. When you are happy, 
repeat

git merge master

Until everything is up to date. This is just if any commits are added to master 
in the meantime. When you are very happy

git checkout master
git merge branchX-merge-master

And voilà.

Hope everything is clear. I am typing on a phone. If you need more details or 
explanations, i'll get on a computer and give a more detail step by step.

HD.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

Re: [git-users] error during git pull

[Huu Da Tran]

Instead of doing:
git pull remotename remotebranch

Try:
git fetch remotename remotebranch
git checkout remotename/remotebranch
git log conflictingfile

You will still need to rename the file prior to do this.

Hope this helps.

HD.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

Re: [git-users] Branch level authentication

[Huu Da Tran]

You could also look into solutions like bitbucket by creating repos and 
assigning different access permissions.

Or on a unix server, achieve the same with different repos and unix user access 
levels.

All this is defined at process level.

Hope that helps,

HD.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Independent modules

[Huu Da Tran]

I find this confusing on the dev level.

This can be developed in a different structure and then use an installation 
procedure to layout the desired final structure.

This way, you can easily have them in separate repos and even have some way to 
create dependencies, if any.

Take a system as an example. You have many binaries and libs. They are 
developed independently, but reside at the end in a shared structure.

Sorry this does not answer your precise question but hope it still helps.

HD.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] git tree subfolder workflow

[Huu Da Tran]

Is the size of all directories a concern at all?

You clone it once and just leave it there and do normal git.

In other words, can you explain the problems you are trying to avoid?

I worked with big repos and cloning is slow, but then branching, commit and 
push is fast, so maybe you are trying to do something different?

Hope this helps,

HD.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: git tree subfolder workflow

[Huu Da Tran]

I do not think it's possible to only clone a part of a repo. That being said, I 
have not experimented with subtree nor submodule. Have you look into that yet?

Is there any way for you to access to repo from a remote host, but not mounted 
nfs? Seems like a flaw in network design.

One hour is not the long if you let run over night (my experience took over 
35min without nfs). Is that nfs plus your network time from the server? If so, 
maybe you can get on that server and clone locally on the server so it is no 
more nfs. Then from your local machine, clone your cloned?

Unfortunately, I can only offer potential workaround. I have no experience to 
say that when pushing back to the nfs repo would be any faster if you could 
clone only a directory.

Hope that helps a little.

HD.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: Independent modules

[Huu Da Tran]

On Saturday, August 17, 2013 2:50:50 PM UTC-4, pbr...@gmail.com wrote:

 Aforementioned structure is forced by external runtime environment. No 
 possibilites for changes.
 Projects are not compiled to libs etc.


I am sorry if my reply was not clearly expressed. It is not necessary to 
compile anything for it to be installed. Your runtime environment doest not 
require to also be the development environment.

Imagine the following:

dev_env
repo1
images/*
scripts/*
libraries/*
repo2
images/*
   ...

so repos are independent. Then you would have a simple install script 
(bash/unix to be adapted or windows batch):

for d in dev_env/*; do
mkdir AppDir/Images/$d
cp $d/images/* AppDir/Images/$d/
mkdir AppDir/Scripts/$d
cp $d/scripts/* AppDir/Scripts/$d/
etc...

Hope this is more clear.

HD.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: pushing a commit on a local branch to a different remote branch

[Huu Da Tran]

On Wednesday, July 31, 2013 6:53:58 PM UTC-4, python.b...@gmail.com wrote:

 I synced to a repo tree which is branch jb... I have a local commit on 
 this branch which I am trying to push to a remote branch jb_mr2 and 
 running into following error..is there a git push command which I can use 
 to push this local commit to jb_mr2


Is there a specific reason why you do not fetch that remote jb_mr2 branch 
(if it exists), check that out, cherry pick the commit and push the jb_mr2 
branch back?

git fetch remote jb_mr2
git checkout -b jb_mr2 remote/jb_mr2
git cherry-pick ...
git push remote jb_mr2 


 qca-bld-03{73} git push ssh://
 company.com:29418/platform/vendor/com-proprietary/ship/ftm72bc75e409e50dcad29bd790b4b6478dc6668f12:refs/for/jb_mr2
 Counting objects: 9, done.
 Delta compression using up to 32 threads.
 Compressing objects: 100% (6/6), done.
 Writing objects: 100% (6/6), 7.01 KiB, done.
 Total 6 (delta 3), reused 0 (delta 0)
 remote: Resolving deltas: 100% (3/3)
 remote: Processing changes: refs: 1, done
 To ssh://company.com:29418/platform/vendor/com-proprietary/ship/ftm
  ! [remote rejected] 72bc75e409e50dcad29bd790b4b6478dc6668f12 - 
 refs/for/jb_mr2 (branch jb_mr2 not found)
 error: failed to push some refs to 'ssh://
 company.com:29418/platform/vendor/com-proprietary/ship/ftm'


-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: Workflow from personal to local to remote?

[Huu Da Tran]

On Wednesday, July 31, 2013 11:22:10 AM UTC-4, leam hall wrote:

 I'm trying to design a workflow for our team and could use some advice. My 
 git-fu is young. 

 We have  a remote repo that will store the authoritative code. That will 
 be pulled to a server local repo where the team has access. Each person 
 will clone the local repo into their home directory, edit files, and then 
 push to the local repo. Somehow we will trigger code review and acks, and 
 then someone will push from the local repo to the authoritative repo. 

 What I need help understanding is:

 1. How to prevent code being pushed from local repo to authoritative until 
 it is reviewed? Is this a process issue or does git have a way to enforce 
 it?


Git itself is only using working conventions. So it's a process external to 
git to actually prevent people from doing (bad?) things.

The easiest setup would be to have:  /local/git/repo-to-review.git   with 
permissions to all dev to write to it. Then have another 
/local/git/authoritative.git which only has write permissions for that 
man-who-will-get-fired-for-merging-bad-code user.

Otherwise, you could use things like bitbucket. Have a repo with one 
read-access to all devs. They would fork into their bitbucket account, then 
clone their bitbucket to their local machine. Make changes to their local 
machine, push code to their bitbucket. Then make a pull request to the 
authoritative bitbucket. The code review process is started and merge if 
satisfied.

 

 2. For the few of us who can push from local to authoritative, how do I 
 configure my .gitconfig to normally push to the local repo? My code should 
 be reviewed, just like everyone else's.


If you prefer not relying on the tools like bitbucket but rather have 
convention-accepted process, I would suggest you have the origin remote 
being your dev bitbucket, and have another remote name for the 
authoritative repo. On the example above:

git clone /local/git/repo-to-review.git my-project
cd my-project
git remote add production /local/git/authoritative.git

then it's just that normal work of git checkout and git push and pull...
 

Thoughts? Pointers to documentation is fine, I just don't know what to look 
 for.

 Leam


Hope this was helpful.

HD.

-- 
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.
For more options, visit https://groups.google.com/groups/opt_out.

[git-users] Re: Free Git Book

[Huu Da Tran]

There is a mention of a ebook now being free of charge. Admitting total 
laziness, it would be great to have a link to the amazon store now free 
ebook... unless I misunderstood.

Thanks.

-- 

[git-users] Re: File diff shows up as single line

[Huu Da Tran]

On Friday, December 21, 2012 6:07:09 PM UTC-5, Aaron Woehler wrote:
 I have two files in my repository that started showing up as single lines. 
 How do I fix this? The files show up fine in Eclipse and other editors.

Mixing unix and windows files?

 
 
 Here is what git-gui shows me.
 
 
 
 @@ -1 +1 @@
 
 -cffunction name=table_data_inputcfargument name=label 
 required=yescfargument name=bindto required=yes .
 
 \ No newline at end of file
 
 +cffunction name=table_data_inputcfargument name=label 
 required=yescfargument name=bindto required=yes .
 
 \ No newline at end of file
 
 
 Environment:
 Windows 7 
 Git-gui 0.17.gitgui
 git 1.8.0.msysgit.0

-- 

[git-users] exclude a subdirectory in sparse-checkout

[Huu Da Tran]

Hi all,

I'm trying to exclude a new subdirectory that was added by someone else.

in my $GITDIR/info/sparse-checkout file, I have:

/web/

Now, I want to exclude the /web/tests/ subdirectory. So I changed my file 
to:

/web/
!/web/tests/

but this does not work. After calling

git read-tree -m -u HEAD

the subdirectory /web/tests is still present. If I change to:

/web/
!tests

the /web/tests subdirectory does disappear, but this may exclude possible 
'tests' directory under any other subdirectories that I may want to keep.

Is it possible to do what I want?

Thanks.

H.

-- 

Re: [git-users] exclude a subdirectory in sparse-checkout

[Huu Da Tran]

On Monday, November 26, 2012 12:42:55 PM UTC-5, William Mizuta wrote:

 Try to use git rm -r /web/tests to remove the directory from your git 
 project.


I want to keep it in the repo, but not in the sparse-checkout. Sorry if 
that was not clear in my original question.

Thanks.
H.

--