avoid cabal dependency hell by installing all your cabal dependencies at the same time
Please use the cabal sandbox feature available in cabal >= 1.18 There is some overlap with the cabal sandbox add-source feature and using cabal-meta. You will probably want to try just using cabal sandbox with add-source and use cabal-meta when you have problems with that workflow.
cabal-meta is a cabal wrapper that facilitates:
Lets explain the last point: If you run this command, you can easily get a failure:
cabal install foo && cabal install bar
Whereas if you run this command it should almost always work:
cabal install foo bar
cabal-meta facilitates the second command: installing everything at once. This is very useful when you want cabal to install packages not on hackage. You need not worry about how cabal interprets your version, instead you let cabal easily interpret a location.
When invoked, cabal-meta looks for a file
Each line of
sources.txt is either a hackage package, a directory, or a github repo (which is cloned into a vendor/ directory).
A directory is either a local cabal package or contains another
sources.txt to recurse into.
cabal-meta automatically uses cabal-src-install (if you have it installed) unless you are using cabal-dev (--dev option). cabal-src-install is used to add local packages to your cabal package database: please see cabal-src documentation. Please note that this is done after the package install. If there is a failure anywhere along the way, cabal-src-install will not be used.
cabal-meta to see help output. Normal usage:
cabal-meta --dev install
You can also supply arguments in an environment variable or a configuration file.
Controlling cabal-meta is done through
sources.txt contains a repo location and optionally a branch name and build flags.
Build flags start with a dash.
git://github.com/foo/bar ghc-7.6-compat -flag
To build a Yesod application using the latest code, create a sources.txt in the project directory with:
./ https://github.com/yesodweb/yesod https://github.com/yesodweb/shakespeare https://github.com/yesodweb/persistent https://github.com/yesodweb/wai
Now just run:
To build a Yesod web application using my already downloaded source from github, I have a sources.txt in my project consisting of
./ sphinx -fone-one-beta path/to/yesod/sources
./ refers to the current directory and thus your current project.
sphinx is a hackage package, and I have a build flag next to it that I don't have to worry about forgetting anymore. warning: a packge build flag will end up being applied to all packages
In the example above,
path/to/yesod/sources contains a
./hamlet ./persistent ./wai ./yesod
Each of these directories has a
sources.txt listing several dirctories containing cabal packages that will be installed.
Confused? It is just recursion, although we are interleaving IO :)