Nix workflow is clunky #5
Labels
Comments
|
For solving this, I've realized that we need a I looked at the implementation and it didn't seem at all difficult but looks like other people are already on it: Maybe we can include it as a patch for now. |
steve-chavez
added a commit
to steve-chavez/plmustache
that referenced
this issue
Feb 4, 2025
Solves PostgREST#5. This is done by avoiding building the postgres extension inside the nix derivation. Instead we build it with `make` on the current directory and install it on postgres using the upcoming postgres `extension_control_path` GUC. This is taken from the patch on https://www.postgresql.org/message-id/flat/E7C7BFFB-8857-48D4-A71F-88B359FADCFD%40justatheory.com and backpatched to postgres 12, 13, 14, 15 and 16, Thanks to this, it's now possible to do watch commands: ``` nxpg-watch nxpg-16 nxpg-build (rebuilds when changing sources) nxpg-watch nxpg-13 nxpg-tmp nxpg-test (runs tests when changing sources) ``` Future improvements: * Makes possible to enable code coverage, since the `.gcno` files will live outside the nix store. * It's wasteful to rebuild the patched pgs everytime, later on this `nxpg` could live in another repo and benefit from Nix caching. Additionally: * removes clean_generated in favor of builtin EXTRA_CLEAN * Don't hardcode sources in Makefile * Update github action
steve-chavez
added a commit
to steve-chavez/plmustache
that referenced
this issue
Feb 4, 2025
Solves PostgREST#5. This is done by avoiding building the postgres extension inside the nix derivation. Instead we build it with `make` on the current directory and install it on postgres using the upcoming postgres `extension_control_path` GUC. This is taken from the patch on https://www.postgresql.org/message-id/flat/E7C7BFFB-8857-48D4-A71F-88B359FADCFD%40justatheory.com and backpatched to postgres 12, 13, 14, 15 and 16, Thanks to this, it's now possible to do watch commands: ``` nxpg-watch nxpg-16 nxpg-build (rebuilds when changing sources) nxpg-watch nxpg-13 nxpg-tmp nxpg-test (runs tests when changing sources) ``` Future improvements: * Makes possible to enable code coverage, since the `.gcno` files will live outside the nix store. * It's wasteful to rebuild the patched pgs everytime, later on this `nxpg` could live in another repo and benefit from Nix caching. Additionally: * removes clean_generated in favor of builtin EXTRA_CLEAN * Don't hardcode sources in Makefile * Update github action
steve-chavez
added a commit
that referenced
this issue
Feb 4, 2025
Solves #5. This is done by avoiding building the postgres extension inside the nix derivation. Instead we build it with `make` on the current directory and install it on postgres using the upcoming postgres `extension_control_path` GUC. This is taken from the patch on https://www.postgresql.org/message-id/flat/E7C7BFFB-8857-48D4-A71F-88B359FADCFD%40justatheory.com and backpatched to postgres 12, 13, 14, 15 and 16, Thanks to this, it's now possible to do watch commands: ``` nxpg-watch nxpg-16 nxpg-build (rebuilds when changing sources) nxpg-watch nxpg-13 nxpg-tmp nxpg-test (runs tests when changing sources) ``` Future improvements: * Makes possible to enable code coverage, since the `.gcno` files will live outside the nix store. * It's wasteful to rebuild the patched pgs everytime, later on this `nxpg` could live in another repo and benefit from Nix caching. Additionally: * removes clean_generated in favor of builtin EXTRA_CLEAN * Don't hardcode sources in Makefile * Update github action
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Problem
To compile, right now I do:
And to recompile:
Proposal
Similarly to the postgrest repo, a:
Would be ideal. Even just a
plmustache-buildcommand where exiting manually is not needed would be an improvement.The text was updated successfully, but these errors were encountered: