req_perform_parallel does not cache results

[mkoohafkan]

It appears that req_perform_parallel() does not cache results.

library(httr2)

# create cache directory
td = file.path(tempdir(), "cache")
dir.create(td)

url <- paste0(
  "https://raw.githubusercontent.com/allisonhorst/palmerpenguins/",
  "master/inst/extdata/penguins.csv"
)

req <- request(url) |> req_cache(td, debug = TRUE)

# First request downloads the data
resps <- req_perform_parallel(list(req))

# check for files in cache after parallel request, directory is empty
list.files(td)
## character(0)

resps <- req_perform_sequential(list(req))

# check for files in cache after sequential request, directory is not empty
list.files(td)
## [1] "d5d1ddd7f99f55dbc920c63f942804c0.rds"

[hadley]

It looks like I'm calling cache_post_fetch() in the place I'd expect it to be, so this will need some exploration and new tests.

[mkoohafkan]

I think it might be this line:

httr2/R/multi-req.R

Line 203 in 824f142

resp <- cache_post_fetch(self$reqs, resp, path = self$paths)

Should it be self$req, rather than self$reqs? And I think also self$path, rather than self$paths.

[hadley]

@mkoohafkan yeah, that looks very suspicious!

[hadley]

Yeah, that appears to be the problem. If I fix it, I get this output:

library(httr2)

td <- withr::local_tempdir()

url <- paste0(
  "https://raw.githubusercontent.com/allisonhorst/palmerpenguins/",
  "master/inst/extdata/penguins.csv"
)

req <- request(url) |> req_cache(td, debug = TRUE)

resps <- req_perform_parallel(list(req))
#> Saving response to cache "d5d1ddd7f99f55dbc920c63f942804c0"
resps <- req_perform_parallel(list(req))
#> Pruning cache
#> Found url in cache "d5d1ddd7f99f55dbc920c63f942804c0"
#> Cached value is fresh; retrieving response from cache

^{Created on 2024-07-11 with reprex v2.1.0}