Describe your Nim builds as tasks. Example nakefile.nim
:
import nake
const
ExeName = "my_app"
BleedingEdgeFeatures = "-d:testFeature1 -d:testFeature2"
task "debug-build", "Build with unproven features":
if shell(nimExe, "c", BleedingEdgeFeatures, "-d:debug", ExeName):
## zip up files or do something useful here?
task "run-tests":
# ...
Now you can run tasks from the nakefile.nim
:
$ nake debug-build
Use Nim's Nimble package manager to install the package and nake
binary:
$ nimble update
$ nimble install nake
If a new version of nake
is released, you can update to it by running those commands again. You can figure out the version you have installed by running nimble path nake
.
Use Nim's Nimble package manager to install locally the GitHub checkout:
$ git clone https://github.com/fowlmouth/nake
$ cd nake
$ nimble install
If you don't mind downloading the Git repository every time, you can also tell Nimble to install the latest development version directly from Git:
$ nimble update
$ nimble install -y nake@#head
Nake has its own nakefile, it will build nake as a binary. The nake
binary is just a shortcut for nim c -r nakefile $ARGS
:
$ cd ~/.nimble/libs/nake
$ nim c -r nakefile install
## boring ^
$ cd ~/my-project
$ nake debug-build
## wow look at the convenience (!!)
Once the nakefile is built, you can run it manually with ./nakefile
, but you can also run nake
again. If nake detects that the source file is newer than the binary, the nakefile will be rebuilt again, otherwise it just runs the nakefile binary directly. Running nake
in such case has an advantage of the nakefile being looked up in parent directories recirsively. You can always remove the nakefile
and the nimcache
directories if you need to force a rebuild.
Most nakefiles will involve running some commands in a shell. To verify what the shell invocations do you can pass the -c
or --careful
switch to a nake
binary and then it will ask you to confirm each command being run:
$ nake --careful install
Run? `nim c nake` [N/y]
Note that this parameter only applies to nake's convenience shell() and direShell() procs, malicious nakefile authors will likely implement their own shell spawning process.
If you run the nakefile without parameters or with the -t
or --tasks
switch it will report the available tasks. But if you run a nakefile with a specific task and this task doesn't exist, nake will report an error, list the available tasks and exit with a non zero status.
In your nakefiles you can define the defaultTask task. This is a task which will be executed if the user runs nake
without specifying a task. Example:
task defaultTask, "Compiles binary":
if binaryRequiresRebuilding():
doStuffToCompileProgram()
else:
echo "Binary is fresh, anything else?"
listTasks()
The documentation of nake
can be found online at http://fowlmouth.github.io/nake/, but you can run the docs task of the included nakefile.nim to generate the user API HTML documentation into a nake.html file. This documentation explains what symbols you can use other than the obvious task() template to define tasks. If you installed using Nimble, you first need to go to your local Nimble directory. UNIX example:
$ cd `nimble path nake`
$ nim c -r nake docs
$ open nake.html
The docs task will also generate HTML versions of all local RST files, which are indexed from the generated theindex.html.
The changes are listed on the releases page.