>
>
>
> On 16/03/15 17:35, Malcolm Cook wrote:
>
>
> On Mon, Mar 16, 2015 at 2:19 AM, Stijn De Weirdt <[email protected]>
> wrote:
>
>> hi malcolm,
>>
>> can you first clarify what this implies wrt accepting the license? (what
>> is the user agreeing to by setting this cookie?).
>>
>>
> My understanding is that you in effect  accepting the "Oracle Binary Code
> License Agreement for the Java SE Platform Products and JavaFX":
> http://www.oracle.com/technetwork/java/javase/terms/license/index.html
>
>
> The question is whether this would be clear to someone install Java using
> EasyBuild...
>
> Is it OK to just accept this for them, programmatically, by default?
>
>
>
>> wrt the implementation, i'd suggest to start extending the java easyblock
>> with this functionality. if requires for other packages too, we can always
>> promote the code to some generic block or framework (but i'm afraid oracle
>> will change this method every other month or so...)
>>
>>  If  you are suggesting that I change just the java.eb to shell out and
> download the tarball (if not already present) to "the right place", well,
> then, in fact, I have already done exactly that for myself. But I see that
> as a hack/temporary workaround - not something in tune with the 'EasyBuild
> way' (such as I am trying to pick up on).  Based on my quick review of the
> source code, the 'EasyBuild way' to implement this is as I suggest,
> requiring changes in a few different coordinated places.   I'm not sure I
> like/approve of this 'way', and thought someone might advise that there is
> a better way to isolate the needed changes just to the easyconfig file, but
> I certainly don't see it.
>
>  Or - what am I missing?
>
>
> Stijn is suggesting to tackle this in the easy*block* (
> https://github.com/hpcugent/easybuild-easyblocks/blob/master/easybuild/easyblocks/j/java.py),
> not the easyconfig file (since that would require including it in all Java
> easyconfigs, which results in lots of duplication).
>
>
Aha!  I get it now.  Thanks.


> Question remains whether it's OK to do this by default, though.
>
>
I now see this concern too.  We don't want EasyBuild to blindly chose to
opt someone into this license.  I suppose there is an actual legal issue
here that I am not qualified to decide.  What about the downstream users of
the installed Java?



> We can certainly support it, and require that people add something like
> 'auto_accept_license = True' in their Java easyconfig files...
>
>
>
 Fotis suggested a generalization that might work.... I'll take it up in
his response....

~Malcolm

regards,
>
> Kenneth
>
>
>
>  stijn
>>
>> On 03/16/2015 05:36 AM, Malcolm Cook wrote:
>>
>>> Hi,
>>>
>>> eb Java-1.8.0_31.eb
>>>
>>> Fails
>>>
>>> Reading the easyconfig source code, I see it advises
>>>
>>> “# download the tar.gz directly from
>>> http://www.oracle.com/technetwork/java/javase/downloads/index.html“
>>>
>>> So I download it (following
>>>
>>> http://stackoverflow.com/questions/10268583/how-to-automate-download-and-installation-of-java-jdk-on-linux
>>> )
>>> as:
>>>
>>> wget -P  ${PREFIX}/source/j/Java/  --no-check-certificate --no-cookies
>>> --header "Cookie: oraclelicense=accept-securebackup-cookie"
>>>
>>> http://download.oracle.com/otn-pub/java/jdk/8u31-b13/jdk-8u31-linux-x64.tar.gz
>>>
>>> and the eb now succeeds. Yay.
>>>
>>> But, surely this should not persist. Something must be done! Perhaps by
>>> me!?!
>>>
>>> Presumably the challenge is the check box which is opaque to the
>>> easybuild
>>> framework, and what is needed is to get such a header declared in the
>>> Java*.eb and contrive for it to be passed down.
>>>
>>> Something like:
>>>
>>>      - change filetools.py download_file(*) to
>>>        - accept a new parameter, namely, headers={}
>>>        - change the call to urlopen as:
>>>
>>> url_fd = urllib2.urlopen(urllib2.Request(url, headers=headers),
>>> timeout=timeout)
>>>
>>>
>>>      - give EasyBlock a new instance variable, say, source_url_headers
>>>     - modify the Java*.eb to include:
>>>
>>>    source_url_headers=[{"Cookie: oraclelicense" :
>>> "accept-securebackup-cookie"}]
>>>
>>> I think this should work and am willing to “cut my teeth” on this project
>>> with this small set of changes if y’all agree it is a good approach to
>>> this
>>> problem.
>>>
>>> Do you?
>>>
>>> Thanks
>>>
>>> ~ Malcolm Cook
>>> ​
>>>
>>>
>
>

Reply via email to