EvilZone
Programming and Scripting => Scripting Languages => : TheWormKill March 12, 2015, 07:02:23 PM
-
Hey everybody!
Slackware hasn't any kind of dependency-resolution for SlackBuilds, the common way to install software.
And that's a good thing (kinda), because this way, you control every single package that is installed.
Still, this has a backside: If you want to install software that has a lot of dependencies and these have their own...
you have to invest some time. Thus, I wrote this simple python script to list dependencies of a package to gain a
quick overview, since I haven't found (or searched) for a project solving this. It might be still buggy and the code looks
terrible, but it works.
UPDATE:
Now it is possible to view the dependencies as a tree or a queue. Apart from that caching was implemented, so processing tangled dependencies should be at least a bit faster now.
import requests
from bs4 import BeautifulSoup
import argparse
BASE_URL = "http://www.slackbuilds.org/result/?search=PACKAGE&sv="
class DummyPackage(object):
def __init__(self, name):
self.name = name
def __eq__(self, other):
return self.name == other.name
class Package(DummyPackage):
def __init__(self, name, cache, tree=False, queue=False):
DummyPackage.__init__(self, name)
self.dependencies = []
self.package_cache = cache # a list of Package objects
self.visited = False
self.fetch()
if queue:
print 'install packages in this order:'
self.traverse_df()
if tree:
print 'dependency tree:'
self.traverse()
def fetch(self):
url = BASE_URL.replace('PACKAGE', self.name)
url2 = ''
site = requests.get(url).content
soup = BeautifulSoup(site)
if not soup.find_all('table', id='results'):
for link in soup.find_all('a'):
if link.string.endswith('.info'):
url2 = 'http://www.slackbuilds.org' + link.get('href')
else:
for link in soup.find_all('a'):
if link.string == ' ' + self.name:
url = 'http://www.slackbuilds.org' + link.get('href')
site = requests.get(url).content
soup = BeautifulSoup(site)
for link in soup.find_all('a'):
if link.string.endswith('.info'):
url2 = 'http://www.slackbuilds.org' + link.get('href')
try:
info = requests.get(url2).content
except requests.exceptions.MissingSchema:
print 'package not found. do you even spell?!'
exit()
for line in info.split('\n'):
if line.startswith('REQUIRES="'):
deps = line[10:-1].split(' ')
deps = [dep for dep in deps if dep]
for dep in deps:
o = DummyPackage(dep)
if o not in self.package_cache:
self.package_cache.append(Package(dep, cache))
self.dependencies.append(self.package_cache[-1])
else:
i = self.package_cache.index(o)
o = self.package_cache[i]
self.dependencies.append(o)
def traverse(self, depth=0):
print ' '*depth + self.name
for dep in self.dependencies:
dep.traverse(depth+1)
def traverse_df(self):
for i in self.dependencies:
if not i.visited:
i.visited = True
i.traverse_df()
print self.name
if __name__ == '__main__':
ARG_PARSER = argparse.ArgumentParser(description='''Find the dependencies
of a Slackbuild package.''')
ARG_PARSER.add_argument('package', help='''the full name of the package
which you want to install''')
ARG_PARSER.add_argument('--slackware-version', default='14.1',
help='''which major release of Slackware are you
using? defaults to 14.1''')
ARG_PARSER.add_argument('-t', help='''show tree''', action='store_true')
ARG_PARSER.add_argument('-q', help='''show queue''', action='store_true')
ARGS = ARG_PARSER.parse_args()
BASE_URL += ARGS.slackware_version
if not ARGS.t and not ARGS.q:
print 'no display mode (either -q or -t) specified. do you even think?!'
exit()
print 'fetching dependencies, please wait...'
cache = []
Package(ARGS.package, cache, ARGS.t, ARGS.q)
The old version (for reference):
import requests
from bs4 import BeautifulSoup
import argparse
BASE_URL = "http://www.slackbuilds.org/result/?search=PACKAGE&sv="
def fetch_deps(package, depth=0):
deps = []
url = BASE_URL.replace('PACKAGE', package)
url2 = ''
site = requests.get(url).content
soup = BeautifulSoup(site)
if not soup.find_all('table', id='results'):
for link in soup.find_all('a'):
if link.string.endswith('.info'):
url2 = 'http://www.slackbuilds.org' + link.get('href')
else:
for link in soup.find_all('a'):
if link.string == ' ' + package:
url = 'http://www.slackbuilds.org' + link.get('href')
site = requests.get(url).content
soup = BeautifulSoup(site)
for link in soup.find_all('a'):
if link.string.endswith('.info'):
url2 = 'http://www.slackbuilds.org' + link.get('href')
info = requests.get(url2).content
for line in info.split('\n'):
if line.startswith('REQUIRES="'):
deps = line[10:-1].split(' ')
ret_str = depth*' ' + package + ':\n'
deps = [dep for dep in deps if dep]
for dep in deps:
ret_str += fetch_deps(dep, depth+1)
return ret_str
if __name__ == '__main__':
ARG_PARSER = argparse.ArgumentParser(description='Find the dependencies of a Slackbuild package.')
ARG_PARSER.add_argument('package', help='the full name of the package which you want to install.')
ARG_PARSER.add_argument('--slackware-version', default='14.1', help='which major release of Slackware are you using? Defaults to 14.1')
ARGS = ARG_PARSER.parse_args()
BASE_URL += ARGS.slackware_version
print 'fetching dependencies, please wait...'
print fetch_deps(ARGS.package)
-
Actually I quite enjoyed this little script, good job.
-
Thank you! At least it's useful when you have to install something like pandoc, which has a shitload of dependencies:
pandoc:
haskell-base64-bytestring:
ghc:
haskell-data-default:
haskell-dlist:
ghc:
haskell-random:
ghc:
haskell-temporary:
ghc:
haskell-zip-archive:
haskell-digest:
ghc:
haskell-mtl:
haskell-transformers:
ghc:
haskell-utf8-string:
ghc:
haskell-zlib:
ghc:
haskell-tagsoup:
haskell-text:
haskell-extensible-exceptions:
ghc:
haskell-citeproc-hs:
haskell-json:
haskell-text:
haskell-extensible-exceptions:
ghc:
haskell-mtl:
haskell-transformers:
ghc:
haskell-syb:
ghc:
haskell-pandoc-types:
haskell-syb:
ghc:
haskell-parsec:
haskell-mtl:
haskell-transformers:
ghc:
haskell-text:
haskell-extensible-exceptions:
ghc:
haskell-utf8-string:
ghc:
haskell-xml:
haskell-text:
haskell-extensible-exceptions:
ghc:
haskell-highlighting-kate:
haskell-blaze-html:
haskell-blaze-markup:
haskell-blaze-builder:
haskell-text:
haskell-extensible-exceptions:
ghc:
haskell-parsec:
haskell-mtl:
haskell-transformers:
ghc:
haskell-text:
haskell-extensible-exceptions:
ghc:
haskell-regex-pcre-builtin:
haskell-regex-base:
haskell-mtl:
haskell-transformers:
ghc:
haskell-texmath:
haskell-parsec:
haskell-mtl:
haskell-transformers:
ghc:
haskell-text:
haskell-extensible-exceptions:
ghc:
haskell-syb:
ghc:
haskell-xml:
haskell-text:
haskell-extensible-exceptions:
ghc:
haskell-HTTP:
haskell-network:
haskell-parsec:
haskell-mtl:
haskell-transformers:
ghc:
haskell-text:
haskell-extensible-exceptions:
ghc:
It might be buggy and slow. though.
-
I haven't looked throught he code.
But i was gonna say why didn't you do it in bash though i have a feeling you are gonna say you don't do bash. Which is ok.
Also you could go ahead and check if the dependency is already satisfied or if it needs an upgrade to a latest release. Blah, blah.
-
I haven't looked throught he code.
But i was gonna say why didn't you do it in bash though i have a feeling you are gonna say you don't do bash. Which is ok.
Well, to be honest, I don't know bash well enough, although I should finally learn it.
Also you could go ahead and check if the dependency is already satisfied or if it needs an upgrade to a latest release. Blah, blah.
That would be a nice feature, but I'm not entirely sure it's possible. slackbuilds.org only mentions the name of the dependency, if a specific version of that package is needed, it will be written in the description, which isn't really machine-readable. On the other hand, checking whether a package is already installed should be possible.