In the existing make/build-all.sh script, the dependency information is intermingled with the code to download the dependencies.
It would be better to separate the information so that it is easier to use alternate or new versions.
In addition, it would also be good to unify usage for different scenarios:
1. (Default) Download dependencies from public repositories
2. Download dependencies from alternate repositories, such as on a local network, possibly behind a firewall
3. Use dependencies on local disk
A lesser goal is to provide a mechanism that can be used by an IDE when developing jtreg code.
It would be better to separate the information so that it is easier to use alternate or new versions.
In addition, it would also be good to unify usage for different scenarios:
1. (Default) Download dependencies from public repositories
2. Download dependencies from alternate repositories, such as on a local network, possibly behind a firewall
3. Use dependencies on local disk
A lesser goal is to provide a mechanism that can be used by an IDE when developing jtreg code.