Update URL parsing regexes and tests (#13411)

* Update URL parsing regexes and tests

* Get rid of no longer used README

* Merge py-udunits and py-cf-units

* netcdf -> netcdf-c

* setup_environment -> setup_*_environment

* Fix doc tests

* Few last minute fixes

* Simplify prefix removal copypasta

4 files changed, 192 insertions, 65 deletions

diff --git a/lib/spack/docs/getting_started.rst b/lib/spack/docs/getting_started.rst
index b7ba386bb2..319d7f3c88 100644
--- a/lib/spack/docs/getting_started.rst
+++ b/lib/spack/docs/getting_started.rst
@@ -97,7 +97,7 @@ Check Installation
 With Spack installed, you should be able to run some basic Spack
 commands.  For example:
 
-.. command-output:: spack spec netcdf
+.. command-output:: spack spec netcdf-c
 
 
 ^^^^^^^^^^^^^^^^^^^^^^^^^^
diff --git a/lib/spack/spack/cmd/url.py b/lib/spack/spack/cmd/url.py
index a24fcc575c..3101f28f08 100644
--- a/lib/spack/spack/cmd/url.py
+++ b/lib/spack/spack/cmd/url.py
@@ -434,23 +434,10 @@ def name_parsed_correctly(pkg, name):
     Returns:
         bool: True if the name was correctly parsed, else False
     """
-    pkg_name = pkg.name
+    pkg_name = remove_prefix(pkg.name)
 
     name = simplify_name(name)
 
-    # After determining a name, `spack create` determines a build system.
-    # Some build systems prepend a special string to the front of the name.
-    # Since this can't be guessed from the URL, it would be unfair to say
-    # that these names are incorrectly parsed, so we remove them.
-    if pkg_name.startswith('r-'):
-        pkg_name = pkg_name[2:]
-    elif pkg_name.startswith('py-'):
-        pkg_name = pkg_name[3:]
-    elif pkg_name.startswith('perl-'):
-        pkg_name = pkg_name[5:]
-    elif pkg_name.startswith('octave-'):
-        pkg_name = pkg_name[7:]
-
     return name == pkg_name
 
 
@@ -475,8 +462,32 @@ def version_parsed_correctly(pkg, version):
     return False
 
 
+def remove_prefix(pkg_name):
+    """Remove build system prefix ('py-', 'perl-', etc.) from a package name.
+
+    After determining a name, `spack create` determines a build system.
+    Some build systems prepend a special string to the front of the name.
+    Since this can't be guessed from the URL, it would be unfair to say
+    that these names are incorrectly parsed, so we remove them.
+
+    Args:
+        pkg_name (str): the name of the package
+
+    Returns:
+        str: the name of the package with any build system prefix removed
+    """
+    prefixes = [
+        'r-', 'py-', 'tcl-', 'lua-', 'perl-', 'ruby-', 'llvm-',
+        'intel-', 'votca-', 'octave-', 'gtkorvo-'
+    ]
+
+    prefix = next((p for p in prefixes if pkg_name.startswith(p)), '')
+
+    return pkg_name[len(prefix):]
+
+
 def remove_separators(version):
-    """Removes separator characters ('.', '_', and '-') from a version.
+    """Remove separator characters ('.', '_', and '-') from a version.
 
     A version like 1.2.3 may be displayed as 1_2_3 in the URL.
     Make sure 1.2.3, 1-2-3, 1_2_3, and 123 are considered equal.
diff --git a/lib/spack/spack/test/url_parse.py b/lib/spack/spack/test/url_parse.py
index c557f78405..7c292938d2 100644
--- a/lib/spack/spack/test/url_parse.py
+++ b/lib/spack/spack/test/url_parse.py
@@ -60,6 +60,8 @@ from spack.version import Version
     ('cppad-20170114.gpl', 'cppad-20170114'),
     # Arch
     ('pcraster-4.1.0_x86-64', 'pcraster-4.1.0'),
+    ('dislin-11.0.linux.i586_64', 'dislin-11.0'),
+    ('PAGIT.V1.01.64bit', 'PAGIT.V1.01'),
     # OS - linux
     ('astyle_2.04_linux', 'astyle_2.04'),
     # OS - unix
@@ -85,20 +87,31 @@ from spack.version import Version
     # Combinations of multiple patterns - darwin
     ('ghc-7.0.4-x86_64-apple-darwin', 'ghc-7.0.4'),
     ('ghc-7.0.4-i386-apple-darwin', 'ghc-7.0.4'),
+    # Combinations of multiple patterns - centos
+    ('sratoolkit.2.8.2-1-centos_linux64', 'sratoolkit.2.8.2-1'),
     # Combinations of multiple patterns - arch
     ('VizGlow_v2.2alpha17-R21November2016-Linux-x86_64-Install',
      'VizGlow_v2.2alpha17-R21November2016'),
     ('jdk-8u92-linux-x64', 'jdk-8u92'),
     ('cuda_6.5.14_linux_64.run', 'cuda_6.5.14'),
+    ('Mathematica_12.0.0_LINUX.sh', 'Mathematica_12.0.0'),
+    ('trf407b.linux64', 'trf407b'),
     # Combinations of multiple patterns - with
     ('mafft-7.221-with-extensions-src', 'mafft-7.221'),
     ('spark-2.0.0-bin-without-hadoop', 'spark-2.0.0'),
+    ('conduit-v0.3.0-src-with-blt', 'conduit-v0.3.0'),
+    # Combinations of multiple patterns - rock
+    ('bitlib-23-2.src.rock', 'bitlib-23-2'),
     # Combinations of multiple patterns - public
     ('dakota-6.3-public.src', 'dakota-6.3'),
     # Combinations of multiple patterns - universal
     ('synergy-1.3.6p2-MacOSX-Universal', 'synergy-1.3.6p2'),
     # Combinations of multiple patterns - dynamic
     ('snptest_v2.5.2_linux_x86_64_dynamic', 'snptest_v2.5.2'),
+    # Combinations of multiple patterns - other
+    ('alglib-3.11.0.cpp.gpl', 'alglib-3.11.0'),
+    ('hpcviewer-2019.08-linux.gtk.x86_64', 'hpcviewer-2019.08'),
+    ('apache-mxnet-src-1.3.0-incubating', 'apache-mxnet-src-1.3.0'),
 ])
 def test_url_strip_version_suffixes(url, expected):
     stripped = strip_version_suffixes(url)
@@ -109,24 +122,40 @@ def test_url_strip_version_suffixes(url, expected):
     # No suffix
     ('rgb-1.0.6', '1.0.6', 'rgb'),
     ('nauty26r7', '26r7', 'nauty'),
+    ('PAGIT.V1.01', '1.01', 'PAGIT'),
+    ('AmpliconNoiseV1.29', '1.29', 'AmpliconNoise'),
     # Download type - install
     ('converge_install_2.3.16', '2.3.16', 'converge'),
     # Download type - src
     ('jpegsrc.v9b', '9b', 'jpeg'),
+    ('blatSrc35', '35', 'blat'),
+    # Download type - open
+    ('RepeatMasker-open-4-0-7', '4-0-7', 'RepeatMasker'),
     # Download type - archive
     ('coinhsl-archive-2014.01.17', '2014.01.17', 'coinhsl'),
     # Download type - std
     ('ghostscript-fonts-std-8.11', '8.11', 'ghostscript-fonts'),
+    # Download type - bin
+    ('GapCloser-bin-v1.12-r6', '1.12-r6', 'GapCloser'),
+    # Download type - software
+    ('orthomclSoftware-v2.0.9', '2.0.9', 'orthomcl'),
     # Download version - release
     ('cbench_release_1.3.0.tar.gz', '1.3.0', 'cbench'),
     # Download version - snapshot
     ('gts-snapshot-121130', '121130', 'gts'),
     # Download version - distrib
     ('zoltan_distrib_v3.83', '3.83', 'zoltan'),
+    # Download version - latest
+    ('Platypus-latest', 'N/A', 'Platypus'),
+    # Download version - complex
+    ('qt-everywhere-opensource-src-5.7.0', '5.7.0', 'qt'),
+    # Arch
+    ('VESTA-x86_64', '3.4.6', 'VESTA'),
     # VCS - bazaar
     ('libvterm-0+bzr681', '681', 'libvterm'),
     # License - gpl
-    ('PyQt-x11-gpl-4.11.3', '4.11.3', 'PyQt-x11')
+    ('PyQt-x11-gpl-4.11.3', '4.11.3', 'PyQt'),
+    ('PyQt4_gpl_x11-4.12.3', '4.12.3', 'PyQt4'),
 ])
 def test_url_strip_name_suffixes(url, version, expected):
     stripped = strip_name_suffixes(url, version)
@@ -182,6 +211,7 @@ def test_url_parse_offset(name, noffset, ver, voffset, path):
 @pytest.mark.parametrize('name,version,url', [
     # Common Repositories - github downloads
 
+    # name/archive/ver.ver
     ('nco', '4.6.2', 'https://github.com/nco/nco/archive/4.6.2.tar.gz'),
     # name/archive/vver.ver
     ('vim', '8.0.0134', 'https://github.com/vim/vim/archive/v8.0.0134.tar.gz'),
@@ -257,6 +287,15 @@ def test_url_parse_offset(name, noffset, ver, voffset, path):
 
     # Common Tarball Formats
 
+    # 1st Pass: Simplest case
+    # Assume name contains no digits and version contains no letters
+
+    # name-ver.ver
+    ('libpng', '1.6.37', 'http://download.sourceforge.net/libpng/libpng-1.6.37.tar.gz'),
+
+    # 2nd Pass: Version only
+    # Assume version contains no letters
+
     # ver.ver
     ('eigen', '3.2.7', 'https://bitbucket.org/eigen/eigen/get/3.2.7.tar.bz2'),
     # ver.ver-ver
@@ -266,10 +305,17 @@ def test_url_parse_offset(name, noffset, ver, voffset, path):
     # vver_ver
     ('luafilesystem', '1_6_3', 'https://github.com/keplerproject/luafilesystem/archive/v1_6_3.tar.gz'),
 
-    # No separators
+    # 3rd Pass: No separator characters are used
+    # Assume name contains no digits
+
+    # namever
     ('turbolinux', '702', 'file://{0}/turbolinux702.tar.gz'.format(os.getcwd())),
     ('nauty', '26r7', 'http://pallini.di.uniroma1.it/nauty26r7.tar.gz'),
-    # Dashes only
+
+    # 4th Pass: A single separator character is used
+    # Assume name contains no digits
+
+    # name-name-ver-ver
     ('Trilinos', '12-10-1',
      'https://github.com/trilinos/Trilinos/archive/trilinos-release-12-10-1.tar.gz'),
     ('panda', '2016-03-07',
@@ -278,7 +324,7 @@ def test_url_parse_offset(name, noffset, ver, voffset, path):
      'http://gts.sourceforge.net/tarballs/gts-snapshot-121130.tar.gz'),
     ('cdd', '061a',
      'http://www.cs.mcgill.ca/~fukuda/download/cdd/cdd-061a.tar.gz'),
-    # Only underscores
+    # name_name_ver_ver
     ('tinyxml', '2_6_2',
      'https://sourceforge.net/projects/tinyxml/files/tinyxml/2.6.2/tinyxml_2_6_2.tar.gz'),
     ('boost', '1_55_0',
@@ -287,9 +333,6 @@ def test_url_parse_offset(name, noffset, ver, voffset, path):
      'https://github.com/dhmunro/yorick/archive/y_2_2_04.tar.gz'),
     ('tbb', '44_20160413',
      'https://www.threadingbuildingblocks.org/sites/default/files/software_releases/source/tbb44_20160413oss_src.tgz'),
-
-    # Only dots
-
     # name.name.ver.ver
     ('prank', '150803', 'http://wasabiapp.org/download/prank/prank.source.150803.tgz'),
     ('jpeg', '9b', 'http://www.ijg.org/files/jpegsrc.v9b.tar.gz'),
@@ -302,61 +345,51 @@ def test_url_parse_offset(name, noffset, ver, voffset, path):
     ('geant', '4.10.01.p03', 'http://geant4.cern.ch/support/source/geant4.10.01.p03.tar.gz'),
     ('tcl', '8.6.5', 'http://prdownloads.sourceforge.net/tcl/tcl8.6.5-src.tar.gz'),
 
-    # Dash and dots
+    # 5th Pass: Two separator characters are used
+    # Name may contain digits, version may contain letters
 
     # name-name-ver.ver
-    # digit in name
     ('m4', '1.4.17', 'https://ftp.gnu.org/gnu/m4/m4-1.4.17.tar.gz'),
-    # letter in version
     ('gmp', '6.0.0a', 'https://gmplib.org/download/gmp/gmp-6.0.0a.tar.bz2'),
-    # version starts with 'v'
     ('LaunchMON', '1.0.2',
      'https://github.com/LLNL/LaunchMON/releases/download/v1.0.2/launchmon-v1.0.2.tar.gz'),
     # name-ver-ver.ver
     ('libedit', '20150325-3.1', 'http://thrysoee.dk/editline/libedit-20150325-3.1.tar.gz'),
-
-    # Dash and unserscores
-
     # name-name-ver_ver
     ('icu4c', '57_1', 'http://download.icu-project.org/files/icu4c/57.1/icu4c-57_1-src.tgz'),
-
-    # Underscores and dots
-
     # name_name_ver.ver
     ('superlu_dist', '4.1', 'http://crd-legacy.lbl.gov/~xiaoye/SuperLU/superlu_dist_4.1.tar.gz'),
     ('pexsi', '0.9.0', 'https://math.berkeley.edu/~linlin/pexsi/download/pexsi_v0.9.0.tar.gz'),
     # name_name.ver.ver
     ('fer', '696', 'ftp://ftp.pmel.noaa.gov/ferret/pub/source/fer_source.v696.tar.gz'),
-
-    # Dash dot dah dot
-
+    # name_name_ver-ver
+    ('Bridger', '2014-12-01',
+     'https://downloads.sourceforge.net/project/rnaseqassembly/Bridger_r2014-12-01.tar.gz'),
     # name-name-ver.ver-ver.ver
     ('sowing', '1.1.23-p1', 'http://ftp.mcs.anl.gov/pub/petsc/externalpackages/sowing-1.1.23-p1.tar.gz'),
     ('bib2xhtml', '3.0-15-gf506', 'http://www.spinellis.gr/sw/textproc/bib2xhtml/bib2xhtml-v3.0-15-gf506.tar.gz'),
     # namever.ver-ver.ver
     ('go', '1.4-bootstrap-20161024', 'https://storage.googleapis.com/golang/go1.4-bootstrap-20161024.tar.gz'),
 
-    # Underscore dash dot
+    # 6th Pass: All three separator characters are used
+    # Name may contain digits, version may contain letters
 
     # name_name-ver.ver
     ('the_silver_searcher', '0.32.0', 'http://geoff.greer.fm/ag/releases/the_silver_searcher-0.32.0.tar.gz'),
     ('sphinx_rtd_theme', '0.1.10a0',
      'https://pypi.python.org/packages/source/s/sphinx_rtd_theme/sphinx_rtd_theme-0.1.10a0.tar.gz'),
-
-    # Dot underscore dot dash dot
-
     # name.name_ver.ver-ver.ver
     ('TH.data', '1.0-8', 'https://cran.r-project.org/src/contrib/TH.data_1.0-8.tar.gz'),
     ('XML', '3.98-1.4', 'https://cran.r-project.org/src/contrib/XML_3.98-1.4.tar.gz'),
-
-    # Dash dot underscore dot
-
     # name-name-ver.ver_ver.ver
     ('pypar', '2.1.5_108',
      'https://storage.googleapis.com/google-code-archive-downloads/v2/code.google.com/pypar/pypar-2.1.5_108.tgz'),
     # name-namever.ver_ver.ver
     ('STAR-CCM+', '11.06.010_02',
      'file://{0}/STAR-CCM+11.06.010_02_linux-x86_64.tar.gz'.format(os.getcwd())),
+    # name-name_name-ver.ver
+    ('PerlIO-utf8_strict', '0.002',
+     'http://search.cpan.org/CPAN/authors/id/L/LE/LEONT/PerlIO-utf8_strict-0.002.tar.gz'),
 
     # Various extensions
     # .tar.gz
@@ -399,18 +432,61 @@ def test_url_parse_offset(name, noffset, ver, voffset, path):
     # .txz
     ('kim-api', '2.1.0', 'https://s3.openkim.org/kim-api/kim-api-2.1.0.txz'),
 
-    # Weird URLS
+    # 8th Pass: Query strings
 
-    # github.com/repo/name/releases/download/name-vver/name
-    ('nextflow', '0.20.1', 'https://github.com/nextflow-io/nextflow/releases/download/v0.20.1/nextflow'),
     # suffix queries
     ('swiftsim', '0.3.0', 'http://gitlab.cosma.dur.ac.uk/swift/swiftsim/repository/archive.tar.gz?ref=v0.3.0'),
-    ('swiftsim', '0.3.0', 'https://gitlab.cosma.dur.ac.uk/api/v4/projects/swift%2Fswiftsim/repository/archive.tar.gz?sha=v0.3.0'),
+    ('swiftsim', '0.3.0',
+     'https://gitlab.cosma.dur.ac.uk/api/v4/projects/swift%2Fswiftsim/repository/archive.tar.gz?sha=v0.3.0'),
     ('sionlib', '1.7.1', 'http://apps.fz-juelich.de/jsc/sionlib/download.php?version=1.7.1'),
+    ('jube2', '2.2.2', 'https://apps.fz-juelich.de/jsc/jube/jube2/download.php?version=2.2.2'),
+    ('archive', '1.0.0', 'https://code.ornl.gov/eck/papyrus/repository/archive.tar.bz2?ref=v1.0.0'),
+    ('VecGeom', '0.3.rc',
+     'https://gitlab.cern.ch/api/v4/projects/VecGeom%2FVecGeom/repository/archive.tar.gz?sha=v0.3.rc'),
+    ('parsplice', '1.1',
+     'https://gitlab.com/api/v4/projects/exaalt%2Fparsplice/repository/archive.tar.gz?sha=v1.1'),
+    ('busco', '2.0.1', 'https://gitlab.com/api/v4/projects/ezlab%2Fbusco/repository/archive.tar.gz?sha=2.0.1'),
+    ('libaec', '1.0.2',
+     'https://gitlab.dkrz.de/api/v4/projects/k202009%2Flibaec/repository/archive.tar.gz?sha=v1.0.2'),
+    ('icet', '2.1.1',
+     'https://gitlab.kitware.com/api/v4/projects/icet%2Ficet/repository/archive.tar.bz2?sha=IceT-2.1.1'),
+    ('vtk-m', '1.3.0',
+     'https://gitlab.kitware.com/api/v4/projects/vtk%2Fvtk-m/repository/archive.tar.gz?sha=v1.3.0'),
+    ('GATK', '3.8-1-0-gf15c1c3ef',
+     'https://software.broadinstitute.org/gatk/download/auth?package=GATK-archive&version=3.8-1-0-gf15c1c3ef'),
     # stem queries
     ('slepc', '3.6.2', 'http://slepc.upv.es/download/download.php?filename=slepc-3.6.2.tar.gz'),
     ('otf', '1.12.5salmon',
      'http://wwwpub.zih.tu-dresden.de/%7Emlieber/dcount/dcount.php?package=otf&get=OTF-1.12.5salmon.tar.gz'),
+    ('eospac', '6.4.0beta.1',
+     'http://laws-green.lanl.gov/projects/data/eos/get_file.php?package=eospac&filename=eospac_v6.4.0beta.1_r20171213193219.tgz'),
+    ('vampirtrace', '5.14.4',
+     'http://wwwpub.zih.tu-dresden.de/~mlieber/dcount/dcount.php?package=vampirtrace&get=VampirTrace-5.14.4.tar.gz'),
+    # (we don't actually look for these, they are picked up
+    #  during the preliminary stem parsing)
+    ('octopus', '6.0', 'http://octopus-code.org/down.php?file=6.0/octopus-6.0.tar.gz'),
+    ('cloog', '0.18.1', 'http://www.bastoul.net/cloog/pages/download/count.php3?url=./cloog-0.18.1.tar.gz'),
+    ('libxc', '2.2.2', 'http://www.tddft.org/programs/octopus/down.php?file=libxc/libxc-2.2.2.tar.gz'),
+    ('cistem', '1.0.0-beta',
+     'https://cistem.org/system/tdf/upload3/cistem-1.0.0-beta-source-code.tar.gz?file=1&type=cistem_details&id=37&force=0'),
+    ('Magics', '4.1.0',
+     'https://confluence.ecmwf.int/download/attachments/3473464/Magics-4.1.0-Source.tar.gz?api=v2'),
+    ('grib_api', '1.17.0',
+     'https://software.ecmwf.int/wiki/download/attachments/3473437/grib_api-1.17.0-Source.tar.gz?api=v2'),
+    ('eccodes', '2.2.0',
+     'https://software.ecmwf.int/wiki/download/attachments/45757960/eccodes-2.2.0-Source.tar.gz?api=v2'),
+    ('SWFFT', '1.0',
+     'https://xgitlab.cels.anl.gov/api/v4/projects/hacc%2FSWFFT/repository/archive.tar.gz?sha=v1.0'),
+
+    # 9th Pass: Version in path
+
+    # github.com/repo/name/releases/download/name-vver/name
+    ('nextflow', '0.20.1', 'https://github.com/nextflow-io/nextflow/releases/download/v0.20.1/nextflow'),
+    # ver/name
+    ('ncbi', '2.2.26', 'ftp://ftp.ncbi.nlm.nih.gov/blast/executables/legacy.NOTSUPPORTED/2.2.26/ncbi.tar.gz'),
+
+    # Other tests for corner cases
+
     # single character name
     ('R', '3.3.2', 'https://cloud.r-project.org/src/base/R-3/R-3.3.2.tar.gz'),
     # name starts with digit
diff --git a/lib/spack/spack/url.py b/lib/spack/spack/url.py
index 11b289a0b1..a298d44ea2 100644
--- a/lib/spack/spack/url.py
+++ b/lib/spack/spack/url.py
@@ -153,13 +153,14 @@ def strip_version_suffixes(path):
         r'[Ii]nstall',
         r'all',
         r'code',
-        r'src(_0)?',
         r'[Ss]ources?',
         r'file',
         r'full',
         r'single',
-        r'public',
         r'with[a-zA-Z_-]+',
+        r'rock',
+        r'src(_0)?',
+        r'public',
         r'bin',
         r'binary',
         r'run',
@@ -189,15 +190,24 @@ def strip_version_suffixes(path):
         r'ia32',
         r'intel',
         r'amd64',
+        r'linux64',
         r'x64',
+        r'64bit',
         r'x86[_-]64',
+        r'i586_64',
         r'x86',
         r'i[36]86',
         r'ppc64(le)?',
         r'armv?(7l|6l|64)',
 
+        # Other
+        r'cpp',
+        r'gtk',
+        r'incubating',
+
         # OS
         r'[Ll]inux(_64)?',
+        r'LINUX',
         r'[Uu]ni?x',
         r'[Ss]un[Oo][Ss]',
         r'[Mm]ac[Oo][Ss][Xx]?',
@@ -208,14 +218,18 @@ def strip_version_suffixes(path):
         r'[Ww]in(64|32)?',
         r'[Cc]ygwin(64|32)?',
         r'[Mm]ingw',
+        r'centos',
 
         # Arch
         # Needs to come before and after OS, appears in both orders
         r'ia32',
         r'intel',
         r'amd64',
+        r'linux64',
         r'x64',
+        r'64bit',
         r'x86[_-]64',
+        r'i586_64',
         r'x86',
         r'i[36]86',
         r'ppc64(le)?',
@@ -270,31 +284,41 @@ def strip_name_suffixes(path, version):
         # name-ver
         # name_ver
         # name.ver
-        r'[._-]v?' + str(version) + '.*',
+        r'[._-][rvV]?' + str(version) + '.*',
 
         # namever
-        str(version) + '.*',
+        r'V?' + str(version) + '.*',
 
         # Download type
         r'install',
-        r'src',
+        r'[Ss]rc',
         r'(open)?[Ss]ources?',
+        r'[._-]open',
         r'[._-]archive',
         r'[._-]std',
+        r'[._-]bin',
+        r'Software',
 
         # Download version
         r'release',
         r'snapshot',
         r'distrib',
+        r'everywhere',
+        r'latest',
 
         # Arch
-        r'Linux64',
+        r'Linux(64)?',
+        r'x86_64',
 
         # VCS
         r'0\+bzr',
 
         # License
         r'gpl',
+
+        # Needs to come before and after gpl, appears in both orders
+        r'[._-]x11',
+        r'gpl',
     ]
 
     for regex in suffix_regexes:
@@ -407,7 +431,7 @@ def parse_version_offset(path):
     # 3. names can contain A-Z, a-z, 0-9, '+', separators
     # 4. versions can contain A-Z, a-z, 0-9, separators
     # 5. versions always start with a digit
-    # 6. versions are often prefixed by a 'v' character
+    # 6. versions are often prefixed by a 'v' or 'r' character
     # 7. separators are most reliable to determine name/version boundaries
 
     # List of the following format:
@@ -450,7 +474,7 @@ def parse_version_offset(path):
         (r'^[a-zA-Z+-]*(\d[\da-zA-Z-]*)$', stem),
 
         # name_name_ver_ver
-        # e.g. tinyxml_2_6_2, boost_1_55_0, tbb2017_20161128, v1_6_3
+        # e.g. tinyxml_2_6_2, boost_1_55_0, tbb2017_20161128
         (r'^[a-zA-Z+_]*(\d[\da-zA-Z_]*)$', stem),
 
         # name.name.ver.ver
@@ -476,6 +500,10 @@ def parse_version_offset(path):
         # e.g. fer_source.v696
         (r'^[a-zA-Z\d+_]+\.v?(\d[\da-zA-Z.]*)$', stem),
 
+        # name_ver-ver
+        # e.g. Bridger_r2014-12-01
+        (r'^[a-zA-Z\d+]+_r?(\d[\da-zA-Z-]*)$', stem),
+
         # name-name-ver.ver-ver.ver
         # e.g. sowing-1.1.23-p1, bib2xhtml-v3.0-15-gf506, 4.6.3-alpha04
         (r'^(?:[a-zA-Z\d+-]+-)?v?(\d[\da-zA-Z.-]*)$', stem),
@@ -507,19 +535,17 @@ def parse_version_offset(path):
         # e.g. STAR-CCM+11.06.010_02
         (r'^[a-zA-Z+-]+(\d[\da-zA-Z._]*)$', stem),
 
+        # name-name_name-ver.ver
+        # e.g. PerlIO-utf8_strict-0.002
+        (r'^[a-zA-Z\d+_-]+-v?(\d[\da-zA-Z.]*)$', stem),
+
         # 7th Pass: Specific VCS
 
         # bazaar
         # e.g. libvterm-0+bzr681
         (r'bzr(\d[\da-zA-Z._-]*)$', stem),
 
-        # 8th Pass: Version in path
-
-        # github.com/repo/name/releases/download/vver/name
-        # e.g. https://github.com/nextflow-io/nextflow/releases/download/v0.20.1/nextflow
-        (r'github\.com/[^/]+/[^/]+/releases/download/[a-zA-Z+._-]*v?(\d[\da-zA-Z._-]*)/', path),  # noqa
-
-        # 9th Pass: Query strings
+        # 8th Pass: Query strings
 
         # e.g. https://gitlab.cosma.dur.ac.uk/api/v4/projects/swift%2Fswiftsim/repository/archive.tar.gz?sha=v0.3.0
         (r'\?sha=[a-zA-Z+._-]*v?(\d[\da-zA-Z._-]*)$', suffix),
@@ -528,13 +554,24 @@ def parse_version_offset(path):
         (r'\?ref=[a-zA-Z+._-]*v?(\d[\da-zA-Z._-]*)$', suffix),
 
         # e.g. http://apps.fz-juelich.de/jsc/sionlib/download.php?version=1.7.1
-        (r'\?version=v?(\d[\da-zA-Z._-]*)$', suffix),
+        # e.g. https://software.broadinstitute.org/gatk/download/auth?package=GATK-archive&version=3.8-1-0-gf15c1c3ef
+        (r'[?&]version=v?(\d[\da-zA-Z._-]*)$', suffix),
 
         # e.g. http://slepc.upv.es/download/download.php?filename=slepc-3.6.2.tar.gz
-        (r'\?filename=[a-zA-Z\d+-]+-v?(\d[\da-zA-Z.]*)$', stem),
+        # e.g. http://laws-green.lanl.gov/projects/data/eos/get_file.php?package=eospac&filename=eospac_v6.4.0beta.1_r20171213193219.tgz
+        (r'[?&]filename=[a-zA-Z\d+-]+[_-]v?(\d[\da-zA-Z.]*)', stem),
 
         # e.g. http://wwwpub.zih.tu-dresden.de/%7Emlieber/dcount/dcount.php?package=otf&get=OTF-1.12.5salmon.tar.gz
-        (r'\?package=[a-zA-Z\d+-]+&get=[a-zA-Z\d+-]+-v?(\d[\da-zA-Z.]*)$', stem),  # noqa
+        (r'&get=[a-zA-Z\d+-]+-v?(\d[\da-zA-Z.]*)$', stem),  # noqa
+
+        # 9th Pass: Version in path
+
+        # github.com/repo/name/releases/download/vver/name
+        # e.g. https://github.com/nextflow-io/nextflow/releases/download/v0.20.1/nextflow
+        (r'github\.com/[^/]+/[^/]+/releases/download/[a-zA-Z+._-]*v?(\d[\da-zA-Z._-]*)/', path),  # noqa
+
+        # e.g. ftp://ftp.ncbi.nlm.nih.gov/blast/executables/legacy.NOTSUPPORTED/2.2.26/ncbi.tar.gz
+        (r'(\d[\da-zA-Z._-]*)/[^/]+$', path),
     ]
 
     for i, version_regex in enumerate(version_regexes):
@@ -662,6 +699,9 @@ def parse_name_offset(path, v=None):
         # e.g. http://wwwpub.zih.tu-dresden.de/%7Emlieber/dcount/dcount.php?package=otf&get=OTF-1.12.5salmon.tar.gz
         (r'\?package=([A-Za-z\d+-]+)', stem),
 
+        # ?package=name-version
+        (r'\?package=([A-Za-z\d]+)', suffix),
+
         # download.php
         # e.g. http://apps.fz-juelich.de/jsc/sionlib/download.php?version=1.7.1
         (r'([^/]+)/download.php$', path),