Open jpmcb opened 2 years ago
Also, as an aside, i've found myself in this situation upon installation on a new system where it seems the run: make
chunk was executed at all and there is no build/
directory, in the local plugin. This makes me think:
run:
correctly all the timeFor reference, here is my nvim-lua configuration
Describe the feature
At times, my nvim packages are (somehow) in a broken state and Packer is unable to recover them elegantly.
An example:
I use
telescope-fzf-native.nvim
which requires their C code be compiled into a localbuild/libfzf.so
file.Their recommended Packer configuration is:
which build the
.so
file into their build directory, ready to go.However, if (for some reason) that block was not executed correctly or that loaded module has entered a broken state, _I need a way for packer to completely rebuild that state
. In other words, I need a way to re-do the
runin that block's
use`.We can simulate this error state by
which removes that file, then I get the following error:
which make sense since that file no longer exists within the package. I don't see a way for Packer to recover from this state without doing some wonky stuff (like commenting out that configuration, running
:PackerClean
andPackerCompile
, and then un-commenting that block to force a re-download and re-run.)Proposal:
Something like would do something similar, but only for a single package.
:PackerRebuild
which clears that package state, removes it's compiled lua file, and re-builds and re-downloads everything would be great. Or to scope it down `PackerRebuildAgain, I do not see a way to get packer to re-do the
run:
command from above beyond changing my packer configuration file, re-compiling, and re-adding that line.