Skip to content

[Bioc-devel] get_dcf_info failed and build intervals

7 messages · Shraddha Pai, Hervé Pagès

Original
Shraddha Pai
# 
Hello BioC dev community,

I just submitted a package to BioC.

1) The automatic build gave it an "ABNORMAL" status with the following
error " Accessing git_info complete. get_dcf_info failed; could not open."
https://github.com/Bioconductor/Contributions/issues/1149
I've Googled a bit and cannot find the solution for what this error means
and how to address it. Can someone please help me?

2) The package above was an AdditionalPackage to a companion dataset. The
first build on the initial dataset package resulted in an error. I address
the error and pushed a git commit 2-3 hours ago, but haven't seen a second
build attempt in response to that push. How often are builds triggered?
Obviously it would be great to do a series of builds and fix errors in a
short span of time.

3) Also how can I search the bioc-devel mailing list?
I only see this archive page: https://stat.ethz.ch/pipermail/bioc-devel/
My apologies if these are trivial questions.

Thanks very much,
Shraddha
Hervé Pagès
# 
Hi,

 From the Linux command line:

   git clone https://github.com/BaderLab/netDx

   R CMD INSTALL netDx
   * installing to library ?/home/hpages/R/R-3.6.r76454/library?
   Error: error reading file '/home/hpages/sandbox/netDx/DESCRIPTION'

   R CMD build netDx
   * checking for file ?netDx/DESCRIPTION? ...Error : file 
'/home/hpages/sandbox/netDx/DESCRIPTION' is not in valid DCF format
    EXISTS but not correct format

 From R:

   library(devtools)
   remotes::install_github("BaderLab/netDx")
   # Error in read.dcf(path) :
   #   Line starting 'The method converts  ...' is malformed!

How about making sure that your package is valid before submitting?

H.
On 6/14/19 15:14, Shraddha Pai wrote:

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Shraddha Pai
    

    

      
      #
    

  


  

      
Hi,

Thanks Herve. I have actually spent weeks making sure my packages passed
check and build following guidelines. So I am actually doing my best and
your snarky remark is not appreciated.

Then at the very end I expanded the DESCRIPTION file by putting in line
breaks in the "description" field, which caused the build to fail. My bad.

Thank you for the feedback, I'm running the build on my machine to make
sure there are no more trivial errors.

Regards
Shraddha

        
On Fri, Jun 14, 2019 at 6:35 PM Pages, Herve <hpages at fredhutch.org> wrote:

        

            

      
      
      
    

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Hervé Pagès
    

    

      
      #
    

  


  

      
On 6/14/19 15:58, Shraddha Pai wrote:

            

      
      
      
    

        
OK maybe. But please note that our guidelines say that your package
should pass 'R CMD check' in less than 10 min. However in your case it
takes more than 20 min (on my laptop). This is the kind of problems that
should preferably be addressed before submission.

Best,
H.

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Shraddha Pai
    

    

      
      #
    

  


  

      
Hi,
Ok thank you.
I should probably comment out some time-consuming steps in the vignettes
and just include the output in the companion data set, read those in.

Thanks, will do that before next push to master.

Best S

        
On Fri, Jun 14, 2019 at 7:37 PM Pages, Herve <hpages at fredhutch.org> wrote:

        

            

      
      
      
    

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Hervé Pagès
    

    

      
      #
    

  


  

      
Sounds good. Also don't forget to run 'R CMD BiocCheck' on the source
tarball of the package. Right now it reports a lot of problems:

   hpages at spectre:~$ R CMD BiocCheck netDx_0.99.0.tar.gz
   ...
   Summary:
   ERROR count: 3
   WARNING count: 7
   NOTE count: 12

Once everything is ready, don't forget to bump the version in order to
trigger a new build by the Single Package Builder. Thanks!

H.

        
On 6/14/19 16:43, Shraddha Pai wrote:

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Shraddha Pai
    

    

      
      #
    

  


  

      
Hi,
Sounds good- that's something I haven't run. I ran R CMD check after adding
the .Rcheck file with BioC flags but this sounds different.
But will look up how to run this.
Thanks, Herve.
Shraddha

        
On Fri, Jun 14, 2019 at 8:02 PM Pages, Herve <hpages at fredhutch.org> wrote: