hmm. I agree that this should be handled by ESS and I have not given up yet.
as indicated in my previous mail, I dont know how to test if an
R-session is remote because the command ess-remote deletes all local
variables. a hack would be to let the R-process evaluate
file.exists((file-name-directory org-babel-temp-file))
thanks
Eric Schulte <schulte.e...@gmail.com> writes:
> It is a shame that this can't be handled gracefully either through ESS
> or R code.
>
> I agree it would be nice to raise a warning rather than hang waiting
> for a file which won't ever exist. So, how can we tell from the Babel
> source if the R session is remote?
>
> Thanks,
>
> Thomas Alexander Gerds <t...@biostat.ku.dk> writes:
>
>> yes, I am using ESS. ess-remote allows me to evaluate R-code from
>> the local emacs-session on a remote machine connected to via ssh.
>> there are two problems:
>> 1) the remote machine cannot write to org-babel-temp-file because
>> the tmp-directory exists on the local machine. here we could add
>> if(!file.exists(dirname(transfer.file))){dir.create(dirname(transfer.file))}
>>
>> in the middle of the variable org-babel-R-write-object-command this
>> would achieve that the file is at least generated on the remote
>> host.
>> 2) however, still the transfer file does not exist on the local
>> machine. there are several possiblities:
>> a) tell org-babel-comint-eval-invisibly-and-wait-for-file that the
>> file is remote and then test if (concat "/" username "@" host ":"
>> file) exists instead of file.
>> b) use tramp to transfer the file from the remote to the local
>> machine. the function ssh does define ssh-host and ssh-username,
>> however, calling ess-remote removes these variables again.
>> c) tell org-babel-comint-eval-invisibly-and-wait-for-file not to
>> wait for file if it is remote
>> my conclusion: it would be nice to have this functionality, but
>> perhaps it is not worth the efforts and it would be sufficient to
>> avoid the endless loop when waiting for a file which never will
>> generated.
>> cheers thomas
>>
>>
>>
>> Eric Schulte <schulte.e...@gmail.com> writes:
>>
>>> Bastien <b...@gnu.org> writes:
>>>
>>>> Hi Thomas, thanks for the follow-up. Thomas Alexander Gerds
>>>> <t...@biostat.ku.dk> writes:
>>>>
>>>>> I think that I can describe the problem a bit better now. It is
>>>>> not related to the silent option but occurs whenever ":results
>>>>> value". Emacs freezes due to the following line in
>>>>> org-babel-comint-eval-invisibly-and-wait-for-file (while (not
>>>>> (file-exists-p file)) (sit-for (or period 0.25))) it seems that R
>>>>> cannot transfer the file and hence this is an endless loop.
>>>> I'm not knowledgeable enough in this area to provide a fix, maybe
>>>> someone else will.
>>>>
>>> Could this be a problem with whatever tool (I'm assuming ESS) you
>>> are using to maintain the R session and generate the R file?
>>> Perhaps babel needs to modify the R code used to create the file
>>> (held in the `org-babel-R-write-object-command' variable). Could
>>> you take a shot at providing another version of this variable? I
>>> don't really use R myself.
>>> Thanks,
>> -- Thomas A. Gerds -- Assoc. Prof. Department of Biostatistics
>> University of Copenhagen, Øster Farimagsgade 5, 1014 Copenhagen,
>> Denmark Office: CSS-15.2.07 (Gamle Kommunehospital) tel: 35327914
>> (sec: 35327901)
--
Thomas A. Gerds -- Assoc. Prof. Department of Biostatistics
University of Copenhagen, Øster Farimagsgade 5, 1014 Copenhagen, Denmark
Office: CSS-15.2.07 (Gamle Kommunehospital)
tel: 35327914 (sec: 35327901)
