Dear Ehshan, dear Kenneth,

I managed to build RStan 2.32.  However, this was before I went on
holiday for two weeks, so I am a little hazy about exactly why I did
what. 

I created an EC

  RStan-2.32.6-gfbf-2023a.eb

which maybe should be renamed to 

  RStan-2.32.6-gfbf-2023a-R-4.3.2.eb

For some reason I did not go down the R-bundle-CRAN path, but managed
with 33 R packages.  I am guessing that I encountered some issue and
reverted to my original approach with the individual packages.

I'll try to open a pull request.

Regards

Loris
 
Ehsan Moravveji <[email protected]> writes:

> Dear Loris
>
> It happens that I was also looking into it, but mainly to build a new
> R-bundle-CRAN module which also includes RStan version 2.32.6.
> I have an open PR (still in progress) which is far from complete, but
> it might give you some starting point. If interested, check out:
> https://github.com/easybuilders/easybuild-easyconfigs/pull/20756
>
> Regards
> Ehsan
>
> -----Original Message-----
> From: [email protected] <[email protected]> On 
> Behalf Of Kenneth Hoste
> Sent: woensdag 5 juni 2024 22:03
> To: Loris Bennett <[email protected]>
> Cc: [email protected]
> Subject: Re: [easybuild] EC for RStan 2.32.6?
>
> Hi Loris,
>
> Yes, you should use R-bundle-CRAN as a dependency, with recent R
> easyconfigs we have started to only include a limited amount of
> extensions with R itself, and moving the bulk of them to the separate
> R-bundle-CRAN easyconfig.
>
>
> regards,
>
> Kenneth
>
> On 08/05/2024 14:04, Loris Bennett wrote:
>> Hi,
>> 
>> Two years ago I created an EC for RStan 2.21.5 with foss/2021b, which 
>> depended only on 'R' and the R packages 'inline', 'loo', 'StanHeaders'
>> and 'rstan' itself.
>> 
>> I am now trying to do an updated version for RStan 2.32.6 with 
>> gfbf/2023a.  However, here I seem to need a whole lot more R packages 
>> (I am up to 20 and still more packages are needed).
>> 
>> Has any one else tried this?  Is the quickest solution just to use
>> 
>>    R-bundle-CRAN-2023.12-foss-2023a
>> 
>> as the dependency?
>> 
>> Cheers,
>> 
>> Loris
>> 
>
-- 
Dr. Loris Bennett (Herr/Mr)
FUB-IT (ex-ZEDAT), Freie Universität Berlin

Reply via email to