Closed eitsupi closed 5 months ago
shouldn't these be named
LIB_R_POLARS_*
instead?
Good point.
The names are a very difficult issue, but I believe that libr_polars
is the name of the actual binary library that will be generated, and adding an underscore like LIB_R_POLARS
would make it confusing.
https://github.com/pola-rs/r-polars/blob/134ed5db04ae722a13203f0c4d8da5db6b5aeebc/src/Makevars.in#L6 https://github.com/pola-rs/r-polars/blob/134ed5db04ae722a13203f0c4d8da5db6b5aeebc/.github/workflows/release-lib.yaml#L26
I believe that
libr_polars
is the name of the actual binary library that will be generated
Would it be a breaking change to rename this binary lib_r_polars
? It's just weird to have libr
/LIBR
as prefix. Since renaming envvars is a breaking change, we might as well rename the binary at the same time.
Would it be a breaking change to rename this binary
lib_r_polars
? It's just weird to havelibr
/LIBR
as prefix.
This name is automatically generated from the crate name "r-polars". i.e. "lib" + "r-polars" -> "libr-polars" -> "libr_polars" This is the cargo's default behavior, so I don't think there's anything strange about it.
Sorry for being nitpicky on this, but I'd prefer envvars to be named R_POLARS_*
, e.g R_POLARS_FEATURES
. This doesn't match the binary name but from the user perspective it makes more sense IMO. The users never interact directly with the binary generated by cargo, so for them there's no consistency issue in having the binary named libr_polars
and the envvars named R_POLARS_*
.
Sorry for being nitpicky on this, but I'd prefer envvars to be named
R_POLARS_*
, e.gR_POLARS_FEATURES
. This doesn't match the binary name but from the user perspective it makes more sense IMO. The users never interact directly with the binary generated by cargo, so for them there's no consistency issue in having the binary namedlibr_polars
and the envvars namedR_POLARS_*
.
That's a good point to consider, and I was thinking the same way when I opened this PR.
However, we can separate environment variables related to Rust build from other environment variables (for example, environment variables that affect the behavior of r-polars may be added) by using "LIB" as a prefix. I thought maybe it wasn't a bad idea.
However, we can separate environment variables related to Rust build from other environment variables (for example, environment variables that affect the behavior of r-polars may be added) by using "LIB" as a prefix.
Right, that's a good distinction to make, thanks for explaining.
At some point, we should add a vignette listing all possible envvars, both for the Rust build and for the use in R (in addition to what you put in the "installation" vignette).
Thanks for your review! I would like @sorhawell to confirm this as well, so I will merge it afterwards.
Thanks for your review! I would like @sorhawell to confirm this as well, so I will merge it afterwards.
Hi @eitsupi and @etiennebacher .
You probably noticed the frequency and quality of PRs and reviews have gone down lately. I think so. I will take a pause from open source for half year or more I think. I want to try out some new directions in a life. It has been a great honor to work with you. You can replace me as maintainer and/or "first author" whenever it makes sense.
I will try to commit the Rstudio completion PR and give up on the vscode completion PR.
@sorhawell Thanks for the reply. Glad to hear you are doing well.
I will merge this and move on.
@sorhawell I hope this break will be good for you and you'll find things you enjoy doing. I'm always impressed by the amount of work you put to make this package working, thanks so much
+1 OSS work is very rewarding... but it can also be incredibly draining. @sorhawell I'm among the many grateful people for all the work that you've put into r-polars
. (Equally grateful for @eitsupi and @etiennebacher picking up the maintenance reins.) Enjoy the break and I hope that it brings you a sense of rest and rejuvenation.
Related to #80, #241, #300, and #445
LIBR_POLARS_FEATURES
instead ofRPOLARS_FULL_FEATURES="true"
. This allows users to freely select features.RPOLARS*
toLIBR_POLARS*
for consistency.These changes are brought over from prqlr.