9 files changed, 107 insertions(+), 7 deletions(-)

M DESCRIPTION
M NAMESPACE
M NEWS.md
M R/apache-httpd.R
A R/exim.R
M R/nginx.R
M R/vershist-package.R
A man/exim_version_history.Rd
M man/vershist.Rd

@@ 1,8 1,8 @@
 Package: vershist
 Type: Package
 Title: Collect Version Histories For Vendor Products
-Version: 0.3.0
-Date: 2019-02-08
+Version: 0.4.0
+Date: 2019-06-08
 Authors@R: c(
     person("Bob", "Rudis", email = "bob@rud.is", role = c("aut", "cre"), 
            comment = c(ORCID = "0000-0001-5670-2640"))

@@ 4,6 4,7 @@ export(apache_httpd_version_history)
 export(apple_ios_version_history)
 export(complete_semver)
 export(etcd_version_history)
+export(exim_version_history)
 export(google_chrome_version_history)
 export(is_valid_semver)
 export(isc_bind_version_history)


@@ 26,6 27,7 @@ importFrom(dplyr,as_tibble)
 importFrom(dplyr,bind_cols)
 importFrom(dplyr,data_frame)
 importFrom(dplyr,distinct)
+importFrom(dplyr,everything)
 importFrom(dplyr,left_join)
 importFrom(dplyr,mutate)
 importFrom(dplyr,mutate_at)

@@ 1,3 1,7 @@
+0.4.0
+* Add Exim version history retriever
+* Improved Apache & nginx version retriever functions
+
 0.3.0
 * Significant improvements to Apache and nginx version retrievers
 * Swapped out as_tibble for as_data_frame

@@ 77,7 77,7 @@ apache_httpd_version_history <- function(refresh = FALSE) {
         major = 1L, minor = 3L, patch = 41L,
         prerelease = NA_character_, build = NA
       ) %>%
-      dplyr::arrange(rls_date) %>%
+      dplyr::arrange(rls_date, major, minor, patch) %>%
       dplyr::mutate(vers = factor(vers, levels=vers)) %>%
       dplyr::select(
         vers, rls_date, rls_year, major, minor, patch, prerelease, build

@@ 0,0 1,65 @@
+#' Retrieve Exim Version Release History
+#'
+#' Reads <https://marc.info/?l=exim-announce&r=1&w=2>
+#' to build a data frame of Exim version release numbers and dates with semantic version
+#' strings parsed and separate fields added. The data frame is also arranged in
+#' order from lowest version to latest version and the `vers` column is an
+#' ordered factor.
+#'
+#' @md
+#' @param refresh if `TRUE` and there `~/.vershist` cache dir exists, will
+#'        cause the version history database for apache to be rebuilt. Defaults
+#'        to `FALSE` and has no effect if `~/.vershist` cache dir does not exist.
+#' @export
+exim_version_history <- function(refresh = FALSE) {
+
+  tech <- "exim"
+  if (use_cache() && (!refresh) && is_cached(tech)) return(read_from_cache(tech))
+
+  pg <- xml2::read_html("https://marc.info/?l=exim-announce&r=1&w=2")
+
+  rvest::html_nodes(pg, xpath=".//a[contains(@href, 'exim-ann')]") %>%
+    rvest::html_attr("href") %>%
+    purrr::keep(grepl("&r=[[:digit:]]+&b=", .)) %>%
+    sprintf("https://marc.info/%s", .) %>%
+    purrr::map(httr::GET) -> exim_list
+
+  purrr::map(
+    exim_list,
+    ~httr::content(.x, encoding = "UTF-8") %>%
+      rvest::html_nodes("pre") %>%
+      rvest::html_text(trim = FALSE) %>%
+      stri_split_lines() %>%
+      unlist()
+  ) %>%
+    purrr::flatten_chr() %>%
+    purrr::keep(grepl("release", ., ignore.case = TRUE)) %>%
+    purrr::discard(grepl("elspy", .)) %>%
+    stri_replace_all_fixed("_", ".") -> rls
+
+  dplyr::tibble(
+    rls_date = stri_match_first_regex(rls, "([[:digit:]]{4}-[[:digit:]]{2}-[[:digit:]]{2})")[,2],
+    vers = stri_match_first_regex(tolower(rls), "\\[exim-announce\\][^[:digit:]]*([[:digit:]]+\\.[[:digit:]]+\\.[[:digit:]]+|[[:digit:]]+\\.[[:digit:]]+)")[,2]
+  ) %>%
+    dplyr::filter(!is.na(vers)) %>%
+    dplyr::filter(vers != "3.953") %>%
+    dplyr::mutate(dots = stri_count_fixed(vers, ".")) %>%
+    dplyr::mutate(vers = ifelse(dots == 1, sprintf("%s.0", vers), vers)) %>%
+    dplyr::mutate(
+      rls_date = as.Date(rls_date),
+      year = lubridate::year(rls_date)
+    ) %>%
+    dplyr::arrange(desc(rls_date)) %>%
+    dplyr::distinct(vers, .keep_all = TRUE) %>%
+    dplyr::arrange(rls_date) %>%
+    tidyr::separate(vers, c("major", "minor", "patch", "prerelease", "build"), sep="\\.", remove = FALSE, fill = "right") %>%
+    dplyr::select(vers, rls_date, rls_year = year, major, minor, patch, prerelease, build, -dots) %>%
+    dplyr::mutate_at(.vars=c("major", "minor", "patch", "build"), .funs=c(as.integer)) %>%
+    dplyr::arrange(rls_date, major, minor, patch) %>%
+    dplyr::mutate(vers = factor(vers, levels = vers)) -> out
+
+  if (use_cache() && (refresh || (!is_cached(tech)))) write_to_cache(out, tech)
+
+  out
+
+}

@@ 65,7 65,7 @@ nginx_version_history <- function(refresh = FALSE) {
           prerelease = NA, build = NA
         )
       ) %>%
-      dplyr::arrange(major, minor, patch) %>%
+      dplyr::arrange(rls_date, major, minor, patch) %>%
       dplyr::mutate(vers = factor(vers, levels=vers)) %>%
       dplyr::select(
         vers, rls_date, rls_year, major, minor, patch, prerelease, build

@@ 5,12 5,12 @@
 #'
 #' @md
 #' @name vershist
-#' @docType package
+#' @keywords internal
 #' @author Bob Rudis (bob@@rud.is)
 #' @import semver
 #' @importFrom purrr keep discard map map_df %>% safely set_names
 #' @importFrom dplyr mutate rename select as_tibble left_join bind_cols arrange
-#' @importFrom dplyr rename progress_estimated mutate_at distinct data_frame
+#' @importFrom dplyr rename progress_estimated mutate_at distinct data_frame everything
 #' @importFrom stringi stri_match_first_regex stri_detect_fixed stri_detect_regex
 #' @importFrom stringi stri_replace_all_regex stri_replace_first_fixed stri_trans_tolower
 #' @importFrom stringi stri_extract_first_regex stri_sub stri_replace_first_regex stri_split_lines


@@ 28,4 28,4 @@
 #' @importFrom git2r clone cred_ssh_key
 #' @useDynLib vershist
 #' @importFrom Rcpp sourceCpp
-NULL>
\ No newline at end of file
+"_PACKAGE"<
\ No newline at end of file

@@ 0,0 1,20 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/exim.R
+\name{exim_version_history}
+\alias{exim_version_history}
+\title{Retrieve Exim Version Release History}
+\usage{
+exim_version_history(refresh = FALSE)
+}
+\arguments{
+\item{refresh}{if \code{TRUE} and there \code{~/.vershist} cache dir exists, will
+cause the version history database for apache to be rebuilt. Defaults
+to \code{FALSE} and has no effect if \code{~/.vershist} cache dir does not exist.}
+}
+\description{
+Reads \url{https://marc.info/?l=exim-announce&r=1&w=2}
+to build a data frame of Exim version release numbers and dates with semantic version
+strings parsed and separate fields added. The data frame is also arranged in
+order from lowest version to latest version and the \code{vers} column is an
+ordered factor.
+}

@@ 9,6 9,15 @@
 Provides a set of functions to gather version histories of products
 (mainly software products) from their sources (generally websites).
 }
+\seealso{
+Useful links:
+\itemize{
+  \item \url{https://gitlab.com/hrbrmstr/vershist}
+  \item Report bugs at \url{https://gitlab.com/hrbrmstr/vershist/issues}
+}
+
+}
 \author{
 Bob Rudis (bob@rud.is)
 }
+\keyword{internal}