NEWS

---

RWsearch 5.2.0 (2024-10-12)

• Version packaged with R-4.4.1 and sent to CRAN (R-4.4.1)
• Major CHANGES since previous version RWsearch_5.1.4 (2024-03-03):
• Updated files zcrandb.rda, ztvdb.rda, zcheck_results.rds, zCRAN-archive.html with date 2024-09-26
• Documentation and vignettes updated accordingly with packages and values 2024-09-26
• p_man(): new function alias of p_htmlweb()
• p_text, p_text2tex, p_text2md, p_text2pdf: add "ChangeLog", "RqmdRmd" and "DOI" arguments
• p_deadline()
• h_Rpkgs(): new function to the HTML documentation of base and CRAN packages
• Minor changes in the manual to cope with the new zfiles of v5.1.9
• p_man(): new function alias of p_htmlweb()
• Remove s_Rman() which is identical to s_man()
• cranmirrors_down(): export cranmirrors_down rather than export cranmirrors in Roxygen / Rd file
• Update Vignette_1 with examples extracted from (z)crandb on 2024-09-26
• Update Vignette_3 with p_vers() and p_vers_deps() with dependencies used on 2024-09-26 Use the old graph (tfestimators, RWsearch, 2019) and the new graph (RWsearch 2024)
• Update Vignette_4 with examples extracted from (z)tvdb on 2024-09-26

RWsearch 5.1.9

• Transitory version to 5.2.0
• Updated files zcrandb.rda, ztvdb.rda, zcheck_results.rds, zCRAN-archive.html with date 2024-09-26
• New file zbinarydb.rda with date 2024-09-26
• Previous files with date 2021-06-04 were updated in v4.9.1 and used until v5.1.8

RWsearch 5.1.8

• p_down, p_down0: add "RqmdRmd" argument
• p_text, p_text2tex, p_text2md, p_text2pdf: add "ChangeLog", "RqmdRmd" and "DOI" arguments
• p_texth(): replace "..." by "^..." in items
• p_texth(): add "ChangeLog", RqmdRmd" and "doi.org..." to arguments and items
• p_texth(): comment the lines "ReadMe" and "github.com"
• p_texth(): select or exclude "R$|Rmd$|qmd$" files from vignettes
• p_downh(): split "pdf$|html$" and "R$|Rmd$|qmd$" files from vignettes
• p_downh(): Change the download order: README, ChangeLog, NEWS before vignettes

RWsearch 5.1.7

• p_deadline()

RWsearch 5.1.6

• version packaged for R-4.4.0

RWsearch 5.1.5

• h_Rpkgs(): new function to the HTML documentation of base and CRAN packages
• Version packaged for R-4.4.0

RWsearch 5.1.4 (2024-03-03)

• Version packaged with R-4.3.2 and sent to CRAN (R-4.3.3)

RWsearch 5.1.3

• Version packaged for R-4.3.2
• replace \item{}{} by \item in s_crandb_tvdb to pass new checks
• roxygen2: "_PACKAGE" replace @docType package
• file a_intro.R split in Wsearch-package.R and a_intro.R

RWsearch 5.1.2

• Version packaged for R-4.2.2

RWsearch 5.1.1

• s_man() and s_Rman(): new functions to earch in the search pages of the R-project
• h_Rman(): new function to the search page of R-project/CRAN
• h_Rblog(): new function for the new R blog address (cited by Thomas Kalibera)
• h_R.R: replace the text "in browser" by "in the browser"
• p_inst.R: p_inst(dependencies = NA)

RWsearch 5.1.0

• Version packaged for R-4.2.0

RWsearch 5.0.5 (2022-02-22)

• adjust the examples to reduce the idle time, probably p_down()

RWsearch 5.0.3

• As suggested by Uwe Ligges, if (interactive()) added to all examples that open html or pdf files:
• it prevents the pdf viewer and the html browser to open
• The 'Additional issue' raised with v_4.9.8 (currently on CRAN) is solved:
• it was a too long example that created an overflow in the console

RWsearch 5.0.2

• This is version 5.0.0 and 5.0.1 renamed
• It was sent to CRAN but not published as
• check --rin-donttest revealed some files blocked by the pdf viewer or the html browser

RWsearch 5.0.1

• Version 5.0.1 was sent to CRAN and blocked as v5.0.0 was in the pending section
• Same package than v.5.0.0 but with a new version number

RWsearch 5.0.0

• Move to 5.0.0 and sent to CRAN but not PUBLISHED
• Changes from version 4.9.3 up to version 5.0.0 are about code cleaning and reducing examples
• this version 5.0.0 was stored in the pending section and finally not published

RWsearch 4.9.20

• Temp-xxx-xxxx file/directory in R CMD check Debian is linked to
• Firefox having not enough time to display the 3 pages produced by
• p_display0357(lst)
• p_display.R: p_display5(lst) ; Sys.sleep(3)
• tvdb.R: p_display7(lst, ..., verbose = FALSE) ; Sys.sleep(4)
• s_sos.R: Sys.sleep(3)
• a_intro.R: Sys.sleep(3)

RWsearch 4.9.17

• checkdb_down(): remove the return() function
• trydownloadurl(), tryopenfile(): replace the argument msg by msgTF
• p_vers.R: add to the examples two lines on binarydb_down(). Remove the old comparison code
• p_down.R: subset lst to lst2 in the examples to reduce the number of packages to load
• p_table2pdf.R reduce the number of examples
• p_html.R: quote the examples p_pdf(), p_vig() to reduce the number of sheets in Firefox
• p_graph.R: reduce the number of examples
• p_display.R: create a common dir
• tvdb.R: reduce the number of examples
• refresh all files to date 2022-01-07

RWsearch 4.9.16

• Remove glitches in the previous 4.9.15 version: ... and (
• h_direct.R: rewrite all function with fmsgTF() and furl()
• h_engine.R: rewrite all function with fmsgTF() and furl()
• e_check(): replace utils::browseURL() by trybrowseURL()
• trybrowseURL(): trybrowseURL(): replace the argument msg by the arguments msgF and msgT
• p_check(): reuse utils::browseURL() in one case
• tvdb.R: funIN replace '%in2%'

RWsearch 4.9.14

• p_html(): rewrite the function with tryCatch()
• p_pdf(): send dir2 to writepkg2pdf()
• writepkg2pdf(): function rewritten with a new argument dir2 to return the file.path
• n_graphF(), n_graphS(): add saveNetwork() et tempfile() because Debian requires it
• p_graph.R: adjust the examples
• p_down.R: adjust the example lapply(..., x[1:2])
• p_check.R: remove the example p_graphF()
• p_archive(): add tryconurl() to detect wrong url(s)
• p_archive_lst(): tempfile(fileext = ".html")
• trybrowseURL(),fmsgTF(), furl(): new functions, mainly for bookmarks
• h_R.R: rewrite all function with fmsgTF() and furl()

RWsearch 4.9.13

• tvdb_pkgs(): remove the last internal variable
• s_tvdb(): remove the last internal variable
• p_text(): tryopenfile() is OK. Remove the former code < 4.9.9
• p_texth(): trydownloadurl() is OK. Remove commented close_libcurl() (end of the loop at line 237)
• table_pdf(), table_pdfh(): the moved openpdf is OK. Remove the former code < 4.9.11
• p_table2pdf.R: add file.path(tempdir(), "ptablepdf")
• p_page(): add tryconurl() to detect wrong url(s)
• p_html2(): rewrite the function with tryCatch(), tryopenfile() and 00Index rather than 00index
• p_man(): add tryconurl() to detect wrong url(s)
• p_pdfweb(): add tryconurl() to detect wrong url(s)

RWsearch 4.9.12

• crandb.R: add file.path(tempdir(), "crandbdown")
• cranmirrors.R: add file.path(tempdir(), "cranmirrors")
• p_display.R: add file.path(tempdir(), "pdisplay")
• s_crandb.R: add file.path(tempdir(), "scrandb")
• s_crandb_tvdb.R: add file.path(tempdir(), "scrandbtvdb")

RWsearch 4.9.11

• table_pdfh, table_pdf(): argument openpdf is moved from table_pdfh() to table_pdf()

RWsearch 4.9.10

• p_texth(): tempfile(fileext = ".html") replaces tempfile() and downloaded files are viewed as html
• a_intro.R: add a subdirectory to tempfiles
• .htmltable(): tryopenfile() replaces utils::browseURL()
• h_zbib(): link to ZoteroBib

RWsearch 4.9.9

• tryconfile() and tryopenfile() created after tryconurl() and trydownloadurl()
• p_text() [discarded]: file.edit(fileEncoding = "") replaces (fileEncoding = "UTF-8") and opens the file rather than the tempfile(). It uses a side effect of file.edit() (see the code of file.edit())
• p_text(): file.edit() replaced by tryopenfile() everywhere as much as possible
• tryopenfile() replaces openpdffile()
• a_intro.R: p_text(lst, pager = TRUE) replaces (..., editor = TRUE)
• a_intro.R and s_sos.R: tail(data.frame(res)) to reduce the output

RWsearch 4.9.8 (2021-12-03)

• a_intro.R + h_sos(): reinstall examples
• a_intro.R + h_so(): reinstall examples

RWsearch 4.9.7

• close_libcurl(): rewrite the function
• tryconurl(): extend the condition and close the connection visible (before this version) in showConnections()
• trydownloadurl(): use the new tryconurl() and add a message for non-available url(s)
• cranmirrors_down(): exit gently when url cannot be downloaded (BR request)
• crandb_down(): issue a message rather than a warning() when url/packages.rds cannot be downloaded
• checkdb_down(): change the message when url/check_results.rds cannot be downloaded
• archivedb_down(): issue a message when url/packages.rds cannot be downloaded
• tvdb_down(): issue a message when url/packages.rds cannot be downloaded
• tvdb_down(), tvdb_load(): return the tvdb invisibly
• p_man(): replace finzi.psych.upenn.edu by search.r-project.org/CRAN/refmans
• a_intro.R: replace finzi.psych.upenn.edu by search.r-project.org
• a_intro.R + h_sos(): remove examples
• a_intro.R + h_so(): remove examples

RWsearch 4.9.6

• p_vers(): permute columns difbc and compare. -3 rather than 3.
• binarydb_load(): replace get(load()) by load()

RWsearch 4.9.5

• p_vers(): new gcc column lists packages whose source code is compiled

RWsearch 4.9.4

• binarydb.R: new file for functions
• binarydb_down()
• binarydb_load()
• p_vers() uses binarydb if loaded in .GloablEnv

RWsearch 4.9.3 (2021-06-05)

• Some minor changes in the vignettes and the documentation
• Seems OK

RWsearch 4.9.2

• p_archive(), p_archive_lst(): url0 inside the function. Message rewritten,
• p_archive(), p_archive_lst(): dontest extended to all examples
• checkdb_load() rewritten: unique(), table(), if_stop()
• Task views rewritten and renumbered
• VignetteIndexEntry reordered

RWsearch 4.9.1

• Updated files zcrandb.rda, ztvdb.rda, zcheck_results.rds, zCRAN-archive.html to today
• Minor changes in the examples, notably the dates in crandb_fromto()

RWsearch 4.9.0

• cranmirrors_down(): replace ftp by https in the url to the .csv file (BR request)
• h_engine.R and vignette RWsearch-4: remove h_1bis.com (website is defunct)

RWsearch 4.8.9

• Correct url (add final cran[...]/views/) in RWsearch-3-TaskViews.Rmd thanks to the new package urlchecker
• New vignette RWsearch-5.
• h_deepl(), h_meteoblue(), h_windy(), h_ianaTLD(), h_ianaWHOIS()

RWsearch 4.8.8

• Update the README with the lines about p_archive, p_deps, p_check features
• Update the 4 vignettes
• Replace tab by four dots in all files
• p_depsrec()

RWsearch 4.8.7

• cnscinfun(): rewritten
• cnsc(): call cnscinfun()
• DESCRIPTION: Description text updated
• DESCRIPTION: Authors_R
• DESCRIPTION: Maintainer email changed from fattailsr to rpackages

RWsearch 4.8.6

• p_inst(): line if (length(LP) != 1L) rewritten
• p_unload_all()
• DESCRIPTION: rewrite the text with comparaison of versions and the safe package unload
• p_vers(): data.frame() replaces data.frame(version = NULL, crandb = NULL, compare = NULL)
• p_vers(): order(dfr$ndeps, dfr$tdeps, -dfr$compare) replaces order(dfr$ndeps, -dfr$compare)
• p_inst(): line rewritten LP <- LP[seq_len(max(1, length(LP) - 1))]

RWsearch 4.8.5

• p_vers(): isNamespaceLoaded() in the first column. version and crandb moved to 2 and 3
• p_vers(subset = "compare < 4"): subset the data.frame. "compare < 0" is usual
• p_deps_deps(): list of recursive dependencies for every first level dependencies
• p_deps_sorted() is removed. p_vers_deps() provides a better information with the loaded namespace
• p_vers_deps(): (BIG novelty) help proceed to desinstallation and re)installation in a safe way!

RWsearch 4.8.4

Remove return() in the following files/functions

• archivedb.R: Lines 139, 1136, 145, 174
• cnsc.R: Lines 76, 108, 141
• crandb.R: Line 80, 192, 206, 236
• funmaintext.R: Lines 51, 72, 95, 113, 122
• p_graph.R: Line 231
• p_html.R: Line 184
• p_table2pdf.R: Line 250
• s_crandb.R: Lines 121, 142, 179, 211
• s_crandb_tvdb.R: Line 113
• s_tvdb.R: Line 40
• tvdb.R: Lines 106, 129, 144
• DESCRIPTION: remove pkgnet from Suggests

RWsearch 4.8.3

• p_vers(): packageDescription(pkg, fields = "Version") replaces packageVersion(pkg) to match crandb values
• p_vers(): Add a column with the versions in crandb (1.4-7 rather than 1.4.7)
• p_vers(): Column compare returns -2, -1, 0, 1, 2, 3
• p_vers(): list of packages and dependencies is allowed; and returns a list
• p_vers(ndeps = TRUE): Calculate the number of recursive dependencies
• p_inst(lib = NULL): try hard to select the most appropriate lib and stop in case of conflict
• p_deps_count(): duplicated pkgs are ignored => unique(pkgs)
• p_deps_ndeps()
• p_deps_sorted()
• p_deps.R: write zzz (or equivalent) at the end of the functions rather than return(zzz)
• p_deps(): if (is.null(pkgs)) return(NULL), if (is.na(pkgs)) return(NA)
• p_deps(verbose = TRUE)

RWsearch 4.8.2

• p_inst()
• p_down(untar), p_down0(untar), p_downarch(untar), targz_down(untar = FALSE)
• p_down0(RWsearch, targz = TRUE, untar = TRUE) is easier to use than
• targz_down("*.tar.gz", untar = TRUE, url = ".../contrib") downloads from the /contrib directory
• p_vers(): remove the warning on the uninstalled packages.
• h_yahoofin(): words moved to fbr(address, words)

RWsearch 4.8.1

• p_graph.R: add to exclpkgs the line \code{exclpkgs = c("ggplot2", list.files(.Library))}.
• a_intro.R and s_crandb.R: add the example s_crandb("^f", select = "P")
• p_texth() and p_downh(): exclude github.com/[..]/readme.md from txtrme and README (bug)
• p_vers(): add the line pkgs <- unique(pkgs) to display the rownames (bug)

RWsearch 4.8.0 (2020-04-20)

• v-4.7.8 with a new version number

RWsearch 4.7.8

• DESCRIPTION -> 'explore' replaces the mis-spelled word 'exlore'
• p_down() in examples: dir <- file.path(tempdir(), ...) ; list.files() to pass CRAN check --run-donttest
• f_pdf(..., dir = tempdir()) is not feasible as going back to the current directory is impossible
• f_pdf() in examples: if (FALSE) used to by-pass CRAN check (and donttest{} removed)
• zcrandb.rda, ztvdb.rda, zCRAN-archive.html refreshed to CRAN at 2020-04-19
• new aabb/zcheck_results.rds -> a small checkdb in .GlobalEnv
• zcrandb.rda, zcheck_results.rds -> UTF-8 characters changed to ASCII
• checkdb.R: example added
• tvdb.R: examples rewritten. The large crandb file is not downloaded anymore
• s_tvdb.R in examples: line tvdb_dfr() added
• s_sos.R: example simplified
• s_hs.R: example removed as it digs to deep in the computer and takes too much time
• s_scrandb_tvb.R: example simplified
• p_text2pdf.R in examples: remove the slow p_text(rmd) example (and copy it in W165*.R)
• p_text2pdf.R, p_table2pdf.R in examples: dir <- file.path(tempdir(), "ptext/ppdf")
• p_html.R in examples: p_html(RWsearch) replaces p_html(pacman)
• p_graph.R: examples rewritten with (RWsearch, visNetwork) et (actuar, fitdistrplus)
• p_down.R in examples: system.time(p_down(lst, ...)). 32 files in 27 seconds
• p_display.R in examples: p_display(RWsearch, ..., brew, dir = tempdir())
• p_deps.R: examples rewritten with (RWsearch, visNetwork) et (actuar, fitdistrplus)
• p_check.R in examples: p_checkdeps_lst(igraph,...) and RWsearch replaces pacman
• p_archive.R in examples: brew replaces pacman
• f_args.R in examples: brew complements sos
• e_check.R in examples: christophe.dutang replaces francois.michonneau
• crandb.R: last example rearranged
• archivedb.R: last example rewritten with 2 histograms

RWsearch 4.7.7

• h_yahoofin() -> add '/' in the address (bug)
• p_vers()
• Shift to R 4.0.0 beta for R CMD build/check/INSTALL and use --run-donttest
• a_intro.R: archivedb_load(system.file("aabb", ...)) -> "aabb" replaces "zarchive" (bug)
• h_Rnews(), h_Rversions() rewritten with separate repos and *.html (and no more Zurich repo)
• p_down.R, p_archive.R: parameter 'before' -> canonical form CRAN.R-project rather than cloud.r-project
• h_R.R in examples: h_crantv(repos = "https://cloud.r-project.org") replaces h_cranbydate() (forgot repos = bug)
• p_check.R in examples: crandb_load(..., zcrandb) replaces crandb_load(..., crandb) (forgot minifile = bug)
• p_down.R in examples: targz_down() replaces targz_downarch() which does not exist (bug)

RWsearch 4.7.6

• p_inun(), p_inun_crandb() simplified -> Intermediate lst and return() removed
• p_graphF(), p_graphS() accept list of packages and open html pages as per the list

RWsearch 4.7.5

• p_check_lst() -> gsub() pour Tinstall -> T_install, Tcheck -> T_check, Ttotal -> T_total
• New file checkdb.R
• checkdb_down(), checkdb_load()
• p_check_lst() now requires checkdb preloaded if number_of(pkgs) > npkgs. Much faster
• p_checkdeps_lst() exclpkgs are now reported in p_check_lst()

RWsearch 4.7.4

• crandb_down(method = "libcurl"). Before was default
• p_archive() moved to new file p_archive.R
• p_archive_lst()
• l_targz()
• p_downarch() moved from file archivedb.R to file p_down.R
• p_downarch(option = "before")
• Before, only the version cited at the last line was downloaded (not necessary the last version = bug)
• targz_down()

RWsearch 4.7.2

• crandb_down(method = "libcurl"). Before was default
• DESCRIPTION -> rewrite the Description section
• file n_graph.R renamed p_graph.R
• p_network() moved from file p_deps.R to file p_graph.R
• p_graphF()
• p_graphS()
• Corresponding documentation rewritten

RWsearch 4.7.1

• Suppress the verbose option in p_deps(), p_depsrev(), p_deps_inpkgs(), p_deps_unpkgs(), p_deps_inun()
• new file p_check.R
• p_check() and p_check_lst() moved from file p_html.R to file p_check.R
• Messages in p_check() (unknown date, does not exist) added to p_check_lst()
• p_checkdeps()
• p_checkdeps_lst()
• p_deps_count()

RWsearch 4.7.0 (2020-02-15)

• v-4.6.9 with a new version number

RWsearch 4.6.9

• p_network(recursive = TRUE) new argument.
• recursive = FALSE reduces the size of the network when reverse = TRUE
• h_etz()
• h_cranchecks()
• p_check(substr(subdate, 1, 4), "/") -> add the year directory to archived packages (CRAN novelty)
• p_check_lst() -> print as a list, not in the browser
• cloud.r-project.org replaces cran.univ-paris1.fr which has been discontinued

RWsearch 4.6.8

• hexagonal logo stored in man/figures/logo.png and used in
• README.md with logo and video about RWsearch at useR! 2019 in Toulouse
• h_doi()
• h_Rnews()
• h_Rversions()
• New version number to fix the date and the version. vignettes updated

RWsearch 4.6.6

• h_biocstats()
• h_crancheckwindows()
• tryconurl() -> hidden function
• p_check() displays the check results of regular packages AND packages moved to CRAN archive
• archivedb_rempkgs() -> Documentation updated with a reference to p_check(). Example added
• archivedb_down() and trydownloadurl() updated with tryconurl()
• p_downh() -> grep("ReadMe",..., invert = TRUE)
• p_texth() -> grep("ReadMe",..., invert = TRUE)

RWsearch 4.6.5

• p_texth() rewritten. It now extracts readme, news, vignette from index.html
• p_downh() rewritten. It now extracts readme, news, vignette from index.html
• trydownloadurl() now returns invisible 0 and invisible 1 as per utils::download.file() (before was NULL)
• trydownloadurl() is vectorized
• p_deps_inun(), p_inun() -> text rewritten
• DESCRIPTION -> suggests xfun

RWsearch 4.6.4

• p_deps_unpkgs(), p_deps_inpkgs(), p_deps_inun() return the dependencies installed
• or not installed in the computer, available or not available in current CRAN (crandb).
• p_down(binary = FALSE, type = "binary") downloads Mac OSX and Windows binaries
• New example in cnscinfun()

RWsearch 4.6.3

• cranmirrors_down()
• h_collins()
• h_googletranslate()
• h_interglot()
• h_promt()
• h_systran()
• h_cpan()
• h_ctan()

RWsearch 4.6.2 (2019-08-19)

• datalist rewritten as per KH recommendation (zcrandb: crandb ztvdb: tvdb)

RWsearch 4.6.1

• h_cranstatus()
• h_orcid()
• h_yahoofin()
• archivedb_list() explicitely returns a (long) list. Before, was invisible(lst).
• \donttest{p_downarch()} (in example)
• datalist: crandb, tvdb rather than zcrandb, ztvdb (request from kh_@_CRAN)

RWsearch 4.6.0

• p_incrandb()
• p_inun_crandb() replaces p_inun_cran(). Names in the output list are adjusted accordingly:
• in_incrandb (in_incrandb), in_notincrandb (in_notincrandb), un_incrandb (un_incrandb), un_notincrandb (n_notincrandb)
• p_texth(), table_pdfh() -> file(open = "w+") replaces "wt". "w+" is the default in the documentation
• close_libcurl() -> nc[num] rather than num (bug 'connection does not exist')
• funreadme(), funnews(), funvignettes() -> move close_libcurl() in p_texth()

RWsearch 4.5.9

• p_network() -> exclpkgs deals with intermediate packages (nodes somewhere in the middle of the graph)
• for both direct and reverse dependencies. Before, NA appeared in the table and n_graphF crashed
• Today, ILB organized a conference for Nicole El Karoui 75 years anniversary
• and Women in Science: https://www.louisbachelier.org/evenement/wise/

RWsearch 4.5.8

• one single directory (before 3): inst/aabb includes assets, template.brew, zCRAN-archive.html
• e_check() -> names(sapply()) rather than lapply() returns nice urls.
• h_gepuro() -> 67153+ packages listed. See also
• https://github.com/gepuro/rpkg/blob/master/index.html
• https://github.com/gepuro/addin4githubinstall
• p_deps() -> pkgs not in crandb -> warning() rather than stop()
• p_deps() -> lst[pkgs] to keep the order in case of NULL (when pkg is not in crandb)
• p_network() -> pkgs[!(pkgs %in% exclpkgs)] everywhere
• use package "goodpractice" -> missing end line in archivedb.R -> add 4 lines

RWsearch 4.5.7

• DESCRIPTION -> Imports networkD3
• p_network()
• n_graphF()
• n_graphS()
• close_libcurl()
• close_libcurl() replaces closeAllConnections() in p_down(), p_text2pdf() -> Much better
• if (closecon) -> suppressed
• p_text(pkgs) -> much better: all Readme, news, vignettes found, but tempfile is open
• p_text() -> editor() et pager() moved from p_texth() to p_text()

RWsearch 4.5.6

• p_deps(), p_depsrev() -> c(list.files(.libPaths()), crandb[, "Package"])

RWsearch 4.5.5

• RStudio substitutes getOption("editor") with its own function, at least in Windows
• p_texth() -> edit(name = NULL) replaces file.edit() to avoid this subtitution
• p_text(), p_table2pdf() -> make.names(nom) replaces next line make.names(filename2)
• Example: filename 2019-0502-file.txt stays unchanged (before, it became X20190502file.txt)
• p_display(), p_down(), table_pdf(), p_text() -> gsub(".", "_") replaces gsub(".", "")
• Example: pkg/filename/directory "R.oo" is changed to "R_oo" rather than "Roo"

RWsearch 4.5.4

• trydownloadurl(), p_downh(in vignettes) -> download code is simplified. TC replaces con
• new files archivedb.R, zCRAN-archive.html and new functions
• archivedb_down(), archivedb_load(), archivedb_npkgs()
• archivedb_pkgs(), archivedb_rempkgs(), archivedb_list()
• p_downarch() + mention in the documentation related to p_down()
• DESCRIPTION -> Imports XML + mention of CRAN archive
• a_intro.R -> @importFrom XML readHTMLTable

RWsearch 4.5.3

• p_depsrev() -> p_deps(reverse = TRUE)
• h_arxiv()
• h_arxivpdf()
• h_ssrn()
• h_ssrnauth()
• crandb_comp() -> odfile more recent than oldfile -> warning rather than stop

RWsearch 4.5.2

• crandb_down(bydate, rev) -> new arguments bydate and rev.
• h_evene()
• h_yacy()
• p_deps() -> a wrapper of tools::package_dependencies()

RWsearch 4.5.1

• crandb_down() -> save(compress = "xz", compression_level = 6). Gain over gzip is 45 %
• zcrandb.rda -> accents on names are removed. 7 UTF-8 glyphs are transformed to pure ASCII
• s_tvdb()
• h_ecosia() -> Ecosia rather than DuckDuckGo

RWsearch 4.5 (2019-03-27)

• page h_ttp.R split in 3 pages as in the vignette: h_direct.R, h_R.R, h_ttp.R
• typo in h_ttp() example
• h_academie()
• spurious UTF-8 characters are removed
• File dates are set to a commonvalue
• Submitted to CRAN

RWsearch 4.4.3

• p_text(README = TRUE, NEWS = TRUE, vignettes = TRUE) reset to TRUE (when set to FALSE ?)
• funreadme(), funnews(), funvignettes() -> tryCatch(condition = ) replaces try(silent = )
• funreadme(), funnews(), funvignettes() + tryCatch() -> options("warn" = 0) (before suppression)
• funreadme(), funnews(), funvignettes() -> (method = libcurl)
• p_text() -> capabilities("libcurl")
• p_text(closecon), p_text2md(closecon), p_text2tex(closecon), p_text2pdf(closecon)
• closecon => Close all unused-ghost connections (README, NEWS) but also the used ones (unfortunately)
• DESCRIPTION -> latexpdf replaces latexpdf (>= 0.1.5) since a warning is issued if older version

RWsearch 4.4

• Back to tvdb_pkgs(gR, Genetics) in tvdb examples
• NEWS RWsearch replaces Cherche
• Submitted to CRAN but fails (spurious UTF-8 characters)

RWsearch 4.3.6

• .htmltable() -> normalizePath to print the html file address
• p_display(verbose), p_display5(verbose), p_display7(verbose), .htmltable(verbose)
• p_display() -> loop replaces mapply
• tvdb_pkgs(Genetics) replaces tvdb_pkgs(gR, Genetics) in tvdb examples

RWsearch 4.3.5

• All examples in the documentation include dir = file.path(tempdir(), "subdir")
• p_table2pdf(cleantex), table_pdfh(cleantex)
• table_pdfh() -> gsub(".tex", ".pdf", filename, fixed = TRUE)
• p_text2pdf() -> gsub(".tex", ".pdf", filename, fixed = TRUE)
• crandb_down(dir = tempdir(), repos = "https://cran.univ-paris1.fr") in examples crandb.R and tvdb.R
• s_sos(try(sos::finFn())
• Examples in zcrandb.rda and ztvdb.rda
• dir = tempdir() in p_display() p_display5() and p_display7() examples

RWsearch 4.3.4

• 4.3.0 CRAN censors request an argument (dir = ...) for all functions that write on disk
• All *.R files renamed
• Commented examples are suppressed
• DESCRIPTION -> Language: en-GB used by package 'spelling'
• 4.3.1 (dir = ".") in crandb_down(), tvdb_down(), p_down(), p_down0()
• 4.3.2 (dir = ".") in p_pdf(), table_pdf(), p_table2pdf()
• p_html -> writepkg2pdf(), openpdffile()
• 4.3.3 (dir = ".") in p_text2pdf(), p_text2md(), p_text2pdf(). New p_text2tex()
• logical replaces boolean
• 4.3.4 p_text2pdf(setwd(dir2)) for pdf generation

RWsearch 4.2.1

• 4.2.1 sent to CRAN. Three dices but not accepted.
• CRAN says too many \dontrun
• 4.2.3 \dontrun and \donttest optimal. --run-donttest is successfull
• 4.2.4 \donttest{} everywhere. --run-donttest crashes. Sent to CRAN
• 4.2.4 CRAN censors request an argument (dir = ...) for all functions that write on disk

RWsearch 4.1.11

• Windows + Debian
• 4.1.8 Debian Many corrections in the examples to cope with the check under Debian
• 4.1.9 Back on Windows => OK
• 4.2 sent to CRAN. NOTE on examples 10-60" => \donttest{} is recommanded
• 4.1.10 Windows \donttest{} rather than \dontrun{} to get meaningful example()
• p_text2pdf() 43" but --run-donttest crashes (vim).
• 4.1.11 Windows \donttest{} rather than \dontrun{} to get meaningful example()
• no example > 5" but --run-donttest crashes (vim).
• 4.1.12 Debian
• 4.1.14 Debian. \dontrun{} everywhere.

RWsearch 4.1.4

• New computer with Debian. OK in user mode without Texlive (not yet installed)
• Explains well that xdg-open/firefox is not allowed in super-user mode (which is used by CRAN to check packages)
• explains well all previous failures from v4.0 (except p_down())
• \dontrun{} rather than ==don't run in the examples should effectively by-pass the calls to the browser

RWsearch 4.1.3

• s_sos() example desactivated
• CRAN has returned 2 NOTES on Debian and 3 NOTES on Windows.

RWsearch 4.1.2

• ERROR on Debian in s_sos()
• /usr/bin/xdg-open: 870: /usr/bin/xdg-open: konqueror: not found
• Error in file(file, if (append) "a" else "w") :
• all connections are in use
• Calls: Anonymous ... print.findFn -> Anonymous -> brew.cached -> sink -> file
• Execution halted
• It seems that CRAN Debian computer has no defined browser
• and the numerous tries to find one crash it.

RWsearch 4.1.1

• trydownloadurl() => open = "rb" hard-coded inside the function
• gsub(".", "", make.names(word), fixed = TRUE) added in functions:
• table_pdf(), p_text(list=)
• p_text(), p_display() => stop ("pkgs is a list with no names.") as in p_down()
• p_display() => names(pkgs) <- gsub(".", "", make.names(names(pkgs)), fixed = TRUE)

RWsearch 4.1

• ERROR on Debian in table_pdf()

RWsearch 4.0.7

• p_down() entirely rewritten. Uses home-made trydownloadurl()
• trydownloadurl(open = "r", method = "libcurl")
• (open = "") creates real connections for non-existing links and is very bad!
• (open = "r") for most files
• (open = "rb") for "rb" for .gz, .bz2, .xz, .tgz, .zip, .rda, .rds, .RData
• (method = "default") creates real connections for non-existing links. libcurl is ok

RWsearch 4.0.6

• ERROR on Debian in p_down()
• p_down() => gsub(".", "", make.names(mot), fixed = TRUE)
• example in intro: s_crandb_list(thermodynamic, "chemical reaction", "gas")
• example in p_down(): s_crandb_list(searchr, probability)

RWsearch 4.0.5

• ERROR on Debian in p_down()
• p_down() => create directories with no space (and point is replaced by _)
• gsub(".", "_", gsub(" ", "", mot), fixed = TRUE)
• and only if it does not exist already
• if (!dir.exists(path)) dir.create(path)
• f_pdf() => shorter example
• s_crandb_tvdb() => p_display7(lst[c("inTV", "notinTV")]) in example
• protect many http examples to reduce call to the web at CRAN check

RWsearch 4.0

• ERROR on Debian -> p_down() generates too many connections
• all search engines generate xdg-open errors. Here h_so()
• /usr/bin/xdg-open: 870: /usr/bin/xdg-open: w3m: not found
• and similar lines for: konqueror, chromium, chromium-browser, google-chrome, firefox,
• x-www-browser, www-browser, firefox, links2, iceweasel, elinks, lynx, w3m, seamonkey
• xdg-open: no method available for opening 'https://stackoverflow.com/search?q=R+deep+neural+network'
• /usr/bin/xdg-open: 870: /usr/bin/xdg-open: mozilla: not found
• /usr/bin/xdg-open: 870: /usr/bin/xdg-open: epiphany: not found
• and similar lines for: konqueror, chromium, chromium-browser, google-chrome, firefox,
• x-www-browser, www-browser, firefox, links2, iceweasel, elinks, lynx, w3m, seamonkey

RWsearch 3.8.1

• h_nabble, h_rbloggers, h_rdoc, h_rdoctv, h_rdrr, h_rseek
• moved from h_engine to h_ttp to put all R related links together
• DESCRIPTION => Date corrected

RWsearch 3.8

• Check before release

RWsearch 3.7

• Vignette 4
• fme(), fbr() => "words" replaces "mots"
• h_framasoft() + h_framasoft0()
• h_lexilogos() replaces h_lexi()

RWsearch 3.6

• Vignettes 1, 2, 3

RWsearch 3.5

• Welcome to RWsearch!
• RWsearch is the new name of this package. Meaning is: Search in R and in the Web.
• The former name was 'Cherche' (Search in French), never publicly released
• Previous NEWS related to 'Cherche' are deleted to avoid any conflict with the version numbers
• Some examples are simplified
• h_cranbydate(), h_cranbyname()