# Other ideas
```
git rebase -i --exec <build command> <first sha you want to test>~
```
Or :
```
#!/usr/bin/env bash
# test_commits.sh
while read -r rev; do
git checkout "$rev"
if ! git submodule update && make clean && make; then
>&2 echo "Commit $rev failed"
exit 1
fi
done < <(git rev-list "$1")
```
Use :
```
./test_commits.sh origin/master..master
```
# Icon for environment
- what could go wrong : Nuke icon for very very bad ideas
# DONE : Custom search and list of environment
If we want to list all **wip** environment in the lecture we call :
If we don't want to list an environment, we do not use this. No need for an extra option in the rustyphoenixlecture
And then, there is a PContent which contains all **wip** of the lecture and their link to jump directly where they are defined in the lecture itself.
- This is not possible to create a part/chapter/section because the update of the file names was already done. So, this could only be about content.
So, for all environment we want to count (we need a map of the environment to be listed)
- we keep in which file the environment is
- we update the file with the section
- when we found the environment we want
- we add in the **PContent** which will be the list :
- the link to the location of the environment in the web site
- the link to the location of the environment in the lecture file
- the environment (to get an idea of its purpose)
- Finally, we have to update the PContentEnvList with the content we just created.
- to get them, without iterating again on all PVecContent, we can have a Vec< usize >
- vec for the index to pass through the PVecContent hierarchy if the PContentEnvList is in a
- we forbid the PContentEnvList to be in a nested PContent (such as PContentTitle)
- so on single usize is enough to get its position and update the PContentEnvList
- When we parse a PContentEnvList, we have to add the name of the environment in the PSearchEnvironment in the PLectureData (because is has to be muttable)
- and we add the PContentEnvList in the main PVecContent
```rust
pub struct PSearchEnvironment{
p_map_env_to_search: HashMap<String, Vec<usize>>,
}
```
# DONE : Cite label from other websites / other lectures
Each lecture provides a configuration file `toml` where all **labels**, **class**, **functions**, etc, are defined.
```toml
[lecture]
name = "SomeProjectName"
repo_url = "https://url/of/the/repository"
doc_url = "https://url/of/the/doc"
# Version of the lecture if it changed
version = "1.2.0"
# Labels which are defined in the lecture (name = "Text in lecture")
[label]
secIntroduction = "1.2.3"
# If we call @SomeProjectName::figTweety This gets => <a href="https://url/of/the/doc/redirection.html?label=figTweety"><b>Figure 26</b></a>
figTweety = "Figure 26"
secIntroduction = {id = "42", title = "Section"}
figTweety = {id = "542", title = "Figure"}
# Defined C++ classes by the lecture
# the corresponding label is class+ClassName (example: classPMatMandelbrot)
# So, each time the lecture parser see PMatMandelbrot, it replaces by a span with the proper class style and an url to the defined class in the dependency lecture
vec_cpp_class = ["PMatMandelbrot"]
vec_rust_class = ["PFileParserIter"]
vec_cpp_function = ["compute_mandelbrot"]
```
This could be a part of the distributed packaged with `pixi` in the **share/ProjectName/doc** directory.
Then in the main lecture, we have to add this dependency :
```toml
[lecture-dependencies]
performance_with_stencil = "https://pages.gitlab.com/url/to/this/lecture/"
```
So we can get automatically the lecture definition and put it in a **lecture_dependencies** directory.
# Idea to improve all
some *test*
- classic url : [text](some/url)
- video : 
- audio : 
- image : 
- svg : 
With highlighting :
- C++ source : 
- C source : 
- Rust source : 
- CMake source : 
- CMakeLists : 
- Toml config : 
- Yml config : 
- Dockerfile : 
- terminal output : 
- we could use the [] to get some extra parameters :
- file type  or 
- animation  => plays the typing of characters to show command lines
- first line is typed fully and then all lines comes one per one
- **first-line-char-delay-ms** : gives the time in milliseconds to draw one character
- **other-line-delay-ms** : gives the time in milliseconds to fully draw all the other lines
- the copy-paste button applies only on the first command row (maybe it has to be smart and keep **escaped newlines** for long commands
We could also auto detect special comment from embeded lecture inside these files. (some `//{`)
- We remember from which source we parse special comment :
- better feed back in case of error / typos in command
- allow the `edit button` even when the lecture is in source
- If there is no special comment we add :
- the source
- a link to get the file (in a `ressource` directory
- If there are some special comments we add :
- the alternated comments and source snipet (and source snipet should have a copy-paste button)
- then the full source
- then a link to get the full file (in a `ressource` directory
Config to do the archive and strip the embeded lecture comments :
```toml
[[archive]]
name = "Example.tar.gz"
directory = "Examples/"
ignore = ["*.pixi", "build"]
```
Best is to remember the time modification of all files to see if we have to rebuild the archive or not.