Re: [Pharo-dev] post-download preparation of Pharo 7 image

[Guillermo Polito]

On Thu, Aug 10, 2017 at 11:55 AM, Pavel Krivanek 
wrote:

> Hi,
>
> currently we are fighting with several issues that make contributions to
> Pharo 7 much more complicated than we would like. Even in case you have no
> troubles with long paths (on Windows) nor SSH keys, it is now not easy now
> to start to contribute or review pull requests.
>
> Let's say you already have:
> - a Github fork of pharo-project/pharo repository
> - a local clone of pharo-project/pharo repository on your disk
> - this repository has a 'remote' pointing to your fork that is named by
> your GitHub user name
>
> During bootstrap the Iceberg registers a repository from which the image
> was bootstrapped and binds Monticello working copies history to it. This
> repository is hidden by default, to see it in the repositories list you
> need to enable Iceberg setting named 'Include system repositories by
> default'. If you will do it, you will see that this repository is red. You
> need to manually set the real path to you local clone (Edit).
>
> Then you should set your fork as default push target (in Iceberg: select
> 'pharo' repository - 'Remotes' tab - context menu on your fork remote -
> Make default push remote).
>
> It is good to synchronize your fork with the origin so it will contain
> recent commits on the development branch. You can do it by doing PULL from
> the origin (pharo-project/pharo) and PUSH to your fork. The alternative is
> to create a pull request from the original repository to your fork in the
> GitHub web interface and then let merge it)
>
> The next optional step is to checkout the particular commit hash from
> which the image was bootstrapped. It is optional because the Iceberg now
> supports cherry picking so you can avoid of committing changes you do not
> want. But to start your branches from a state when the image working copy
> is in sync with the git working copy is of course cleaner.
>
> Currently the bootstrapped image has some packages that are marked as
> dirty and it will make mess if you will try to commit your changes. You can
> avoid it by reloading of these dirty packages. This will be not needed in
> future.
>
> All these steps you can automatized by the following code. You need to at
> least customize your repository clone path and GitHub user name. It is
> supposed to be run only once on freshly downloaded Pharo 7 image.
>
> | location userName pharoRepository fork |
>
> location := '/path/to/your/pharo/repository'.
> userName := 'github-user-name'.
>
> Iceberg showSystemRepositories: true.
> pharoRepository := IceRepository registry detect: [ :each | each name =
> 'No name' ].
> pharoRepository location: location asFileReference.
> fork := pharoRepository remotes detect: [ :remote | remote remoteName =
> userName ].
> pharoRepository pushRemote: fork.
> pharoRepository pullRemote: pharoRepository origin.
>


> pharoRepository checkoutBranch: 'development'.
>
pharoRepository backend pullFrom: pharoRepository origin.
>

Take into account this particular two lines can break your working copy if
you have uncommited changes.


> pharoRepository push.
> pharoRepository checkoutBranch: (SystemVersion current commitHash).
>



> [(pharoRepository savedPackages select: [:aPackage | aPackage isLoaded
> and: [ aPackage  isModified ]]) do: #reload ] on: MCMergeOrLoadWarning do:
> [ :warning | warning resume: true ]
>

And this is a patch because bootstrapped images have 3 dirty packages. We
should just fix the issue instead of patching it.

In any case, I vote to put this in the Pharo plugin in iceberg.

>
>
> As the result you should get a clean image based on the latest code that
> is ready to to create or review pull requests.
>
> Cheers,
> -- Pavel
>



-- 



Guille Polito


Research Engineer

French National Center for Scientific Research - *http://www.cnrs.fr*




*Web:* *http://guillep.github.io* 

*Phone: *+33 06 52 70 66 13

[Pharo-dev] post-download preparation of Pharo 7 image

[Pavel Krivanek]

Hi,

currently we are fighting with several issues that make contributions to
Pharo 7 much more complicated than we would like. Even in case you have no
troubles with long paths (on Windows) nor SSH keys, it is now not easy now
to start to contribute or review pull requests.

Let's say you already have:
- a Github fork of pharo-project/pharo repository
- a local clone of pharo-project/pharo repository on your disk
- this repository has a 'remote' pointing to your fork that is named by
your GitHub user name

During bootstrap the Iceberg registers a repository from which the image
was bootstrapped and binds Monticello working copies history to it. This
repository is hidden by default, to see it in the repositories list you
need to enable Iceberg setting named 'Include system repositories by
default'. If you will do it, you will see that this repository is red. You
need to manually set the real path to you local clone (Edit).

Then you should set your fork as default push target (in Iceberg: select
'pharo' repository - 'Remotes' tab - context menu on your fork remote -
Make default push remote).

It is good to synchronize your fork with the origin so it will contain
recent commits on the development branch. You can do it by doing PULL from
the origin (pharo-project/pharo) and PUSH to your fork. The alternative is
to create a pull request from the original repository to your fork in the
GitHub web interface and then let merge it)

The next optional step is to checkout the particular commit hash from which
the image was bootstrapped. It is optional because the Iceberg now supports
cherry picking so you can avoid of committing changes you do not want. But
to start your branches from a state when the image working copy is in sync
with the git working copy is of course cleaner.

Currently the bootstrapped image has some packages that are marked as dirty
and it will make mess if you will try to commit your changes. You can avoid
it by reloading of these dirty packages. This will be not needed in future.

All these steps you can automatized by the following code. You need to at
least customize your repository clone path and GitHub user name. It is
supposed to be run only once on freshly downloaded Pharo 7 image.

| location userName pharoRepository fork |

location := '/path/to/your/pharo/repository'.
userName := 'github-user-name'.

Iceberg showSystemRepositories: true.
pharoRepository := IceRepository registry detect: [ :each | each name = 'No
name' ].
pharoRepository location: location asFileReference.
fork := pharoRepository remotes detect: [ :remote | remote remoteName =
userName ].
pharoRepository pushRemote: fork.
pharoRepository pullRemote: pharoRepository origin.
pharoRepository checkoutBranch: 'development'.
pharoRepository backend pullFrom: pharoRepository origin.
pharoRepository push.
pharoRepository checkoutBranch: (SystemVersion current commitHash).
[(pharoRepository savedPackages select: [:aPackage | aPackage isLoaded and:
[ aPackage  isModified ]]) do: #reload ] on: MCMergeOrLoadWarning do: [
:warning | warning resume: true ]


As the result you should get a clean image based on the latest code that is
ready to to create or review pull requests.

Cheers,
-- Pavel