This repository provides tools for managing the MOOSE Framework demo application: Babbler. The main
branch of the Babbler repository is available at idaholab/babbler
.
Each child directory in babbler_devel/
, whose name takes the format: stepXX_*
, is treated like a commit. The contents of these directories reflect the exact state of a branch at their respective commits. The order in which commits are applied to idaholab/babbler
follows, chronologically, the order of the XX
indices used in the aforementioned directory naming convention. Thus, it is critical that this convention is followed.
Start by cloning the repository in whichever parent folder idaholab/moose
has been cloned to, preferably, ~/projects/
:
cd ~/projects/
git clone [email protected]:crswong888/babbler_devel.git
The Babbler application (or perhaps, some other application used for the purposes of a MOOSE tutorial) can be initialized while simultaneously establishing the initial commit. For example, we might name the initial commit's directory "step01_moose_app" and run the following commands:
cd ~/projects/
moose/scripts/stork.sh Babbler
mv babbler/ step01_moose_app/
Many of the tasks performed by shell scripts available from babbler_devel/scripts/
make use of the local Git configuration of crswong888/babbler_devel
, so, in general, it is wise to keep track of all of the files:
git add step01*
Also, be sure to keep the local clone of crswong888/babbler_devel
up-to-date by occasionally running git commit
.
Before starting to make changes to the Babbler application, a directory to represent a new commit should be created. The scripts/new_commit.sh
makes a copy of the specified <source>
directory to a specified <destination>
directory in such a way that makes sense for a Git-based version control system:
./scripts/new_commit.sh <source> <destination> --add
As an example, if the name for the second idaholab/babbler
commit were "step02_input_file," then it can be created as follows:
./scripts/new_commit.sh step01_moose_app/ step02_input_file/ --add
Then, changes to Babbler, which will reflect the second commit, can be made in step02_input_file/
. Once all of the desired changes have been made in the commit directory, ensure that C++ code is styled properly, changes have been tracked, and that everything works properly:
git clang-format -f
git add step02*
./scripts/test_commit.sh step02_input_file/
The test_commit.sh
script will compile the Babbler application as it exists in the specified directory (e.g., step02_input_file/
), run TestHarness, and execute every input file in the directory, while ensuring that all of these tasks exit properly. Only input files whose name are suffixed with "_error
" will be skipped.
The commit message used for a given commit directory is specified in babbler.log
. The message must follow the name of the associated directory enclosed in square brackets and is terminated by the empty bracket, []
, syntax. As an example, the commit messages associated with the step01_moose_app/
and step02_input_file/
directories can be set as follows:
[step01_moose_app]
Initial files
[]
[step02_input_file]
created an input file to solve diffusion problem
[]
Commit messages may be multiple lines and lines starting with "#
" will be ignored, just like a standard Git commit message. However, leading/trailing white space will not be read. Any commit directory that does not have a corresponding message will not be committed to idaholab/babbler
.
A devel
branch of the idaholab/babbler
repository can be automatically generated and reflect the history provided by each of the commit directories, and their associated messages provided in babbler.log
, by running the scripts/push_babbler_devel.sh
script. It pushes to devel
via the babbler
submodule contained in crswong888/babbler_devel
. In addition to pushing commits, it will also compile and test the application from within the submodule after each commit is applied. If any process fails, the script will conduct an abort procedure.
Only changes tracked by, or committed to, the local copy of crswong888/babbler_devel
will be applied to the idaholab/babbler
commits. So be sure to add tracking before running the script:
git clang-format -f
git add step*
./scripts/push_babbler_devel.sh
After successfully pushing to the devel
branch of the remote repository, confirm that it is ready to be merged into the main
branch. If everything looks good on devel
, then run the following script:
./scripts/merge_babbler_devel.sh
This will delete the devel
branch once it has been merged into main
. It will also compile and test the submodule at the tip of the branch.
The most sophisticated tool available from this repository is the scripts/rebase_commits.sh
file. Its purpose is to merge changes from files in a specified commit directory to similar files in each directory which lies ahead of it on the branch's history. For example, consider a case where changes have been made to some files in step02_input_file/
. For any commit directories labeled in accordance with the established naming conventions, and with XX > 02
, a proper merge with the step02*
version of one of its files will be made.
After edits have been made, the rebase can be applied:
./scripts/rebase_commits.sh step02_input_file/
The script works by cross-linking a complex set of diff
outputs between files. These processes are, currently, not well documented. For now, all that anyone needs to know is that it should work just like a standard Git rebase.
Another tool invokes the rebase_commits.sh
script in such a way that it rebases all commit directories on top of the stork application located at ../moose/stork
. This is useful for keeping Babbler looking and functioning like a brand new MOOSE application and it can be ran like so:
./scripts/rebase_stork.sh
This repository is under heavy development. Currently, its primary purpose is to keep copies of the stepXX_*
directories up-to-date in the MOOSE repository, while providing convenient means for doing so. Eventually, all of the scripting used here may be moved to idaholab/moose
, or, at least, this is the proposed idea.
To copy a commit directory, e.g., step02_input_file/
to ../moose/tutorials/tutorial01_app_development/
, run the shell script:
./scripts/copy_to_moose.sh step02_input_file/ --add
Alternatively, a sequence of commit directories whose step indices are within a range bounded by the numbers <lower>
and <upper>
may be copied to MOOSE with a single command:
./script/multicopy_to_moose.sh <lower> <upper>
For example, both step01_moose_app/
and step02_input_file/
may be copied like so:
./scripts/multicopy_to_moose.sh 1 2 --add --force
Once all of the commit directories have been copied over, it is possible to test all versions of the Babbler application within the MOOSE directory by running the following:
./scripts/test_moose.sh