tukusejssirs
commented
1 year ago This issue is related to #185.
We can already do something like this:
repo_root='some/path/to/a/folder'
# git clone https://github.com/Ladder99/ladder99 "$repo_root" # -b v0.10.1
git clone git@github.com:Ladder99/ladder99.git "$repo_root" # -b v0.10.1
cd "$repo_root"
PATH="$PATH:$repo_root/shell" L99_HOME="$repo_root" l99 use example
PATH="$PATH:$repo_root/shell" L99_HOME="$repo_root" l99 start
PATH="$PATH:$repo_root/shell" L99_HOME="$repo_root" l99 status
ob: I prefer not letting apps updating my .bashrc file (as you do in shell/install/cli), therefore I appended my .bashrc with return. I think it could be better to output (in shell/install/cli log) what should be done by the user manually. YMMV.
ob: Well, I think that when we run a program/app without installing it, we should use either absolute or relative path to l99 script. However, when we install a program/app (i.e. copy it to the host file system e.g. into /usr/bin with other data elsewhere, like C libraries should go /usr/lib), then l99 should be found in PATH and thus accessible without path specification. ob: Moreover, when I run some/path/l99 directly without the env vars defined, it errors out that I have not set it up, albeit it is set up that way. IMO it should run as expected, whether it is run directly (with path specified) or without it and all env vars should be optional (used only when we want to override something).
bb: Ah, that makes sense - okay we should let user run './l99' by itself, or 'shell/install/cli' to their system and run as 'l99'. And should have default values for the envars.