cran2deb: extend license handling, specify repository, translate R->Debian package...

License handling code now deals with version numbers and also the more
common malformed license fields found in the wild.  License version
information is discarded completely at the moment (though retained in
debian/copyright).

debian/copyright also includes a copyright notice formed from the Author
field of DESCRIPTION.

Debian package names must be lower case; the Debian source package
of an R package is the lowercase form of the R package name, whilst the
binary package is r-<lower case repo>-<lower case R name>.

git-svn-id: svn://svn.r-forge.r-project.org/svnroot/cran2deb@14 edb9625f-4e0d-4859-8d74-9fd3b1da38cb

diff --git a/pkg/trunk/build_some b/pkg/trunk/build_some
new file mode 100755 (executable)
index 0000000..b72d466
--- /dev/null
+++ b/pkg/trunk/build_some
@@ -0,0 +1,16 @@
+#!/usr/bin/rc
+
+mkdir -p log
+./cranpkgs >all_pkgs
+for (pkg in `{cat all_pkgs}) {
+    if (~ $pkg *.* */*) {
+        echo bad name $pkg >log/ERROR
+    } else {
+        echo $pkg...
+        fail=0
+        ./cran2deb $pkg >log/$pkg >[2=1] || fail=1
+        if (~ $fail 0) {
+            rm -f log/$pkg
+        }
+    }
+}

diff --git a/pkg/trunk/cran2deb b/pkg/trunk/cran2deb
index 54008a5934d323f50913c849f257da058c634a44..9e67627e79e3ba7956c9476a3cb184d929874ff3 100755 (executable)
--- a/pkg/trunk/cran2deb
+++ b/pkg/trunk/cran2deb
@@ -7,7 +7,7 @@ version.new <- function(rver,debian_revision=1, debian_epoch=0) {
 
     # ``Writing R extensions'' says that the version consists of at least two
     # non-negative integers, separated by . or -
-    if (!length(grep('^([0-9]*[1-9][.-])+[0-9]*[1-9]$',rver))) {
+    if (!length(grep('^([0-9]+[.-])+[0-9]+$',rver))) {
         stop(paste('Not a valid R package version',rver))
     }
 
@@ -76,10 +76,15 @@ setup <- function() {
     return (tmp)
 }
 
-prepare.pkg <- function(dir, pkgname) {
+cleanup <- function(dir) {
+    unlink(dir,recursive=T)
+    invisible()
+}
+
+prepare.pkg <- function(dir, pkgname,repo='cran',repoURL='http://cran.uk.r-project.org/') {
     # based loosely on library/utils/R/packages2.R::install.packages
     # should do nothing Debian specific
-    archive <- download.packages(pkgname, dir, type="source")[1,2]
+    archive <- download.packages(pkgname, dir, repos=repoURL, type="source")[1,2]
     wd <- getwd()
     setwd(dir)
     if (length(grep('\\.zip$',archive))) {
@@ -101,17 +106,27 @@ prepare.pkg <- function(dir, pkgname) {
                   ,gsub(.standard_regexps()$valid_package_version, ""
                   ,archive))
     pkg$description = read.dcf(paste(pkg$path,'DESCRIPTION',sep='/'))
-    pkg$repo = 'cran'
+    pkg$repo = repo
+    pkg$repoURL = repoURL
     pkg$version = pkg$description[1,'Version']
     return(pkg)
 }
 
-debian_ok_licenses=c('GPL-2','GPL-3','LGPL-2','LGPL-2.1','LGPL-3'
-                    ,'AGPL-3','Artistic-1.0','Artistic-2.0','Unlimited')
+debian_ok_licenses=c('GPL','LGPL','AGPL','Artistic','Unlimited')
+# these are commonly found 'in the wild'. the R packages should be fixed...
+debian_wild_licenses=c('gpl version 2 or newer'
+                      ,'gpl version 2 or later'
+                      ,'gnu gpl'
+                      ,'gpl v2'
+                      ,'gpl (version 2 or later)'
+                      ,'gpl version 2'
+                      ,'gpl 2 or later'
+                      ,'gpl2')
 
 chomp <- function(x) {
-    return(sub('^[:space:]+','',sub('[:space:]+$','',x)))
+    return(sub('^[[:space:]]+','',sub('[[:space:]]+$','',x)))
 }
+
 prepare.new.debian <- function(pkg) {
     maintainer = 'cran2deb buildbot <cran2deb@example.org>'
 
@@ -126,7 +141,11 @@ prepare.new.debian <- function(pkg) {
     pkg$archive = debarchive
     # not sure can rename directory from R(!)
     pkg$debversion = version.new(pkg$version)
-    pkg$debname = paste('r',pkg$repo,pkg$name,sep='-')
+    if (!length(grep('^[A-Za-z0-9][A-Za-z0-9+.-]+$',pkg$name))) {
+        stop(paste('Cannot convert package name into a Debian name',pkg$name))
+    }
+    pkg$debname = paste('r',tolower(pkg$repo),tolower(pkg$name),sep='-')
+    pkg$srcname = tolower(pkg$name)
     debdir <- paste(pkg$path,'debian',sep='/')
     debfile <- function(x) { paste(debdir,x,sep='/') }
     unlink(debdir,recursive=T)
@@ -134,7 +153,7 @@ prepare.new.debian <- function(pkg) {
 
     # TODO: ``Writing R extensions'' mentions that a package may also have
     # {NEWS,ChangeLog} files.
-    cat(paste(paste(pkg$name,' (',pkg$debversion,') unstable; urgency=low',sep='')
+    cat(paste(paste(pkg$srcname,' (',pkg$debversion,') unstable; urgency=low',sep='')
              ,'' ,'  * Initial release.',''
              ,paste(' --',maintainer,'',format(Sys.time(),'%a, %d %b %Y %H:%M:%S %z'))
              ,'',sep='\n'),file=debfile('changelog'))
@@ -149,17 +168,27 @@ prepare.new.debian <- function(pkg) {
         stop('package has no License: field in description!')
     }
     accept=NULL
-    for (license in strsplit(chomp(pkg$description[1,'License']),' *\\| *')[[1]]) {
-        if (license %in% debian_ok_licenses) {
+    for (license in strsplit(chomp(pkg$description[1,'License'])
+                            ,'[[:space:]]*\\|[[:space:]]*')[[1]]) {
+
+        # don't care about versions of licenses
+        license0 = chomp(sub('\\([<=>!]+[[:space:]]*[0-9.]+\\)',''
+                        ,sub('-[0-9.]+','',license)))
+        if (license0 %in% debian_ok_licenses) {
+            accept=license
+            break
+        }
+        license0 = sub('\\.$','',tolower(license0))
+        if (license0 %in% debian_wild_licenses) {
+            message('Accepted wild license. FIX THE PACKAGE!')
             accept=license
             break
         }
         # TODO: put debian_ok_licenses in DB
         # TODO: file {LICENSE,LICENCE} (+ maybe COPYING?)
-        # TODO: name (op v)
     }
     if (is.null(accept)) {
-        stop('No acceptable license.')
+        stop(paste('No acceptable license:',pkg$description[1,'License']))
     } else {
         message(paste('Auto-accepted license',accept))
     }
@@ -170,15 +199,19 @@ prepare.new.debian <- function(pkg) {
     }
     writeLines(strwrap(
         paste('This Debian package of the GNU R package',pkg$name
-             ,'was generated using cran2deb by'
+             ,'was generated automatically using cran2deb by'
              ,paste(maintainer,'.',sep='')
-             ,'The author(s) of the original GNU R package were:'
+             ,''
+             ,'The original GNU R package is Copyright (C) '
+             # TODO: copyright start date, true copyright date
+             ,format(Sys.time(),'%Y')
              ,pkg$description[1,'Author']
+             ,'and possibly others.'
              ,''
              ,'The original GNU R package is maintained by'
              ,pkg$description[1,'Maintainer'],'and was obtained from:'
              ,''
-             ,paste('    ',contrib.url(getOption("repos")))
+             ,pkg$repoURL
              ,''
              ,''
              ,'The GNU R package DESCRIPTION offers a'
@@ -207,7 +240,7 @@ prepare.new.debian <- function(pkg) {
         shlibdep = '${shlibs:Depends}'
     }
     control = data.frame()
-    control[1,'Source'] = pkg$name
+    control[1,'Source'] = pkg$srcname
     control[1,'Section'] = 'math'
     control[1,'Priority'] = 'optional'
     control[1,'Maintainer'] = maintainer
@@ -223,8 +256,13 @@ prepare.new.debian <- function(pkg) {
         control[2,'Architecture'] = 'any'
     }
     control[2,'Depends'] = paste('r-base-core', shlibdep, sep=', ')
-    control[2,'Description'] = paste('GNU R package "'
-                                    ,pkg$description[1,'Title'],'"\n'
+    descr = 'GNU R package "'
+    if ('Title' %in% colnames(pkg$description)) {
+        descr = paste(descr,pkg$description[1,'Title'],sep='')
+    } else {
+        descr = paste(descr,pkg$name,sep='')
+    }
+    control[2,'Description'] = paste(descr,'"\n'
                                     ,pkg$description[1,'Description']
                                     ,sep='')
     # Debian policy says 72 char width; indent minimally
@@ -245,18 +283,23 @@ build.debian <- function(pkg) {
     setwd(wd)
 }
 
-cleanup <- function(dir) {
-    unlink(dir, recursive=T)
-
-    invisible()
-}
-
 
-test <- function(name) {
+go <- function(name) {
     dir <- setup()
+    pkg <- try((function() {
         pkg <- prepare.new.debian(prepare.pkg(dir,name))
         build.debian(pkg)
-#    cleanup(dir)
+        return(pkg)
+    })())
+    cleanup(dir)
+    if (inherits(pkg,'try-error')) {
+        stop(call.=F)
+    }
     return(pkg)
 }
 
+if (exists('argv')) { # check for littler
+    for (arg in argv) {
+        go(arg)
+    }
+}