Fix some suboptimal patterns identified by lintr

[Bisaloo]

One potential last thing is that find_compress() could use the endsWith() base R function:

rio/R/compression.R

Lines 1 to 35 in 588971f

	find_compress <- function(f) {
	if (grepl("\\.zip$", f)) {
	return(list(file = sub("\\.zip$", "", f), compress = "zip"))
	}
	if (grepl("\\.tar\\.gz$", f)) {
	return(list(file = sub("\\.tar\\.gz$", "", f), compress = "tar.gz"))
	}
	if (grepl(".tgz$", f)) {
	return(list(file = sub("\\.tgz$", "", f), compress = "tar.gz"))
	}
	if (grepl("\\.tar\\.bz2$", f)) {
	return(list(file = sub("\\.tar\\.bz2$", "", f), compress = "tar.bz2"))
	}
	if (grepl("\\.tbz2$", f)) {
	return(list(file = sub("\\.tbz2$", "", f), compress = "tar.bz2"))
	}
	if (grepl("\\.tar$", f)) {
	return(list(file = sub("\\.tar$", "", f), compress = "tar"))
	}
	if (grepl("\\.gzip$", f)) {
	## weird
	return(list(file = sub("\\.gzip$", "", f), compress = "gzip"))
	}
	if (grepl("\\.gz$", f)) {
	return(list(file = sub("\\.gz$", "", f), compress = "gzip"))
	}
	if (grepl("\\.bz2$", f)) {
	return(list(file = sub("\\.bz2$", "", f), compress = "bzip2"))
	}
	if (grepl("\\.bzip2$", f)) {
	## weird
	return(list(file = sub("\\.bzip2$", "", f), compress = "bzip2"))
	}
	return(list(file = f, compress = NA_character_))
	}

But from my (admittedly limited) testing, it seems that NA_character_ is a possible value for f, leading to the creation of a file named NA. This means endsWith() is not a good fit here.

But is it an explicit design choice to allow NA_character_ or an unexpected side effect?

[Bisaloo]

Tests failures are due to r-lib/testthat#1969.

[chainsawriot]

@Bisaloo Once again, thank you very much for the PR. Most of them LGTM. I still have to actually stress test the removal of the setwd(). Those code is not automatically checked and last time I checked, setwd() must still be there. #319

[chainsawriot]

For this, NA_character_ is surely an unexpected side effect and should be excluded. So, I think it's safe to use endsWith() here. I will open another issue related to this.

One potential last thing is that find_compress() could use the endsWith() base R function:

rio/R/compression.R

Lines 1 to 35 in 588971f

	find_compress <- function(f) {
	if (grepl("\\.zip$", f)) {
	return(list(file = sub("\\.zip$", "", f), compress = "zip"))
	}
	if (grepl("\\.tar\\.gz$", f)) {
	return(list(file = sub("\\.tar\\.gz$", "", f), compress = "tar.gz"))
	}
	if (grepl(".tgz$", f)) {
	return(list(file = sub("\\.tgz$", "", f), compress = "tar.gz"))
	}
	if (grepl("\\.tar\\.bz2$", f)) {
	return(list(file = sub("\\.tar\\.bz2$", "", f), compress = "tar.bz2"))
	}
	if (grepl("\\.tbz2$", f)) {
	return(list(file = sub("\\.tbz2$", "", f), compress = "tar.bz2"))
	}
	if (grepl("\\.tar$", f)) {
	return(list(file = sub("\\.tar$", "", f), compress = "tar"))
	}
	if (grepl("\\.gzip$", f)) {
	## weird
	return(list(file = sub("\\.gzip$", "", f), compress = "gzip"))
	}
	if (grepl("\\.gz$", f)) {
	return(list(file = sub("\\.gz$", "", f), compress = "gzip"))
	}
	if (grepl("\\.bz2$", f)) {
	return(list(file = sub("\\.bz2$", "", f), compress = "bzip2"))
	}
	if (grepl("\\.bzip2$", f)) {
	## weird
	return(list(file = sub("\\.bzip2$", "", f), compress = "bzip2"))
	}
	return(list(file = f, compress = NA_character_))
	}

But from my (admittedly limited) testing, it seems that NA_character_ is a possible value for f, leading to the creation of a file named NA. This means endsWith() is not a good fit here.

But is it an explicit design choice to allow NA_character_ or an unexpected side effect?

[chainsawriot]

@Bisaloo Once again, thank you very much for the PR. Most of them LGTM. I still have to actually stress test the removal of the setwd(). Those code is not automatically checked and last time I checked, setwd() must still be there. #319

OK, This one should be fine.

[chainsawriot]

@Bisaloo Please add yourself to Authors@R. Then I will merge this.