Commits

Branch selector

 julia0.5

User selector

All users

Datepicker

All time

Commit History

Commits on Aug 7, 2017

• Merge pull request #51 from bbolker/julia0.5

  Show description for c47de8d
  

  Peter Haverty

  authored
  c47de8d

  Copy full SHA for c47de8d

Commits on Jul 17, 2017

• hack jl_is_float{32,64}

  bbolker

  committed
  a7e22a6

  Copy full SHA for a7e22a6

Commits on Mar 15, 2017

• Copied jDo and friends from rjulia2. This way seems to be more robust to errors on the julia side.

  

  phaverty

  committed
  8c55824

  Copy full SHA for 8c55824

Commits on Dec 14, 2016

• More GC PUSH simplifications

  

  phaverty

  committed
  fb6ddb6

  Copy full SHA for fb6ddb6

Commits on Dec 13, 2016

• Safer way to do GC in top level r2j functions

  

  phaverty

  committed
  5de58d2

  Copy full SHA for 5de58d2

Commits on Nov 26, 2016

• Really added version check to julia_init

  

  phaverty

  committed
  43ad892

  Copy full SHA for 43ad892

• Added version check to julia_init

  

  phaverty

  committed
  5ecf729

  Copy full SHA for 5ecf729

Commits on Oct 29, 2016

• change default branch to julia0.5

  

  Yu Gong

  committed
  b6c61ed

  Copy full SHA for b6c61ed

Commits on Oct 26, 2016

• .julia_init_if_necessary was calling julia_init with one unnamed arg, which was taken as a path rather than the gc flag. Also simplified loading the rjulia shared object

  

  phaverty

  committed
  60ef39c

  Copy full SHA for 60ef39c

Commits on Aug 20, 2016

• Merge pull request #43 from phaverty/julia0.5

  Show description for 7352b68
  

  Yu Gong

  authored
  7352b68

  Copy full SHA for 7352b68

Commits on Aug 17, 2016

• Had GC turned off for tests, turned back on and tests still pass.

  

  Peter M. Haverty

  committed
  43ffb14

  Copy full SHA for 43ffb14

• r2j and j2r now work for data.frames, and tests pass. I am using the original implemnation of R_Julia_MD_NA_DataFrame as creating a non-empty DataFrame with jl_call2 fails for mysterious reasons.

  

  Peter M. Haverty

  committed
  62775cd

  Copy full SHA for 62775cd

• Fix for returning a data frame with PooledDataArray columns. These are no longer DataArrays according to isa.

  

  Peter M. Haverty

  committed
  9824cdf

  Copy full SHA for 9824cdf

• typo in stringsAsFactors that was ignored by partial matching it seems

  

  Peter M. Haverty

  committed
  2a3bac7

  Copy full SHA for 2a3bac7

• went back to old r2j definition for DataFrame

  

  Peter M. Haverty

  committed
  198bd9f

  Copy full SHA for 198bd9f

Commits on Aug 14, 2016

• r2j and j2r on a data.frame work, but only with gc disabled.

  

  Peter M. Haverty

  committed
  f1bf947

  Copy full SHA for f1bf947

Commits on Aug 13, 2016

• Sending and receiving a zero row, zero col data.frame works using jl_get_function

  

  Peter M. Haverty

  committed
  8bd5018

  Copy full SHA for 8bd5018

Commits on Aug 8, 2016

• had to turn on biginttodouble to make return of Int64 dataframe columns work. We are converting Int64 to REALSXP for now, correct?

  

  Peter M. Haverty

  committed
  92ab5cb

  Copy full SHA for 92ab5cb

Commits on Aug 6, 2016

• Fixed bug in Julia_R_MD_NA that made returning STRSXPs crash. Now the only problem is with sending and maybe returning data.frames

  

  Peter M. Haverty

  committed
  b8e5414

  Copy full SHA for b8e5414

• more write barriers. R_Julia_MD and R_Julia_MD_NA and R_Julia_MD_NA_Factor work for any input. Still trouble with dataframe input and j2r for strings with NAs.

  

  Peter M. Haverty

  committed
  2bb5f91

  Copy full SHA for 2bb5f91

Commits on Aug 5, 2016

• data frames always need names

  

  Peter M. Haverty

  committed
  64f97d7

  Copy full SHA for 64f97d7

Commits on Aug 4, 2016

• Added GC write barriers for creating an array of jl_string_type and the VECSXP equivalent Array{Any,1}. Still crash problems with DataArrays of character and logicals.

  

  Peter M. Haverty

  committed
  dbfc99c

  Copy full SHA for dbfc99c

Commits on Aug 3, 2016

• oops, left code in debug state in previous checkin

  

  Peter M. Haverty

  committed
  a4ad927

  Copy full SHA for a4ad927

• GC push/pop in CreateArray

  

  Peter M. Haverty

  committed
  357b1d2

  Copy full SHA for 357b1d2

• More tests, specifically factor and data.frame tests

  

  Peter M. Haverty

  committed
  f1c8aaf

  Copy full SHA for f1c8aaf

• Dropped Julia_R_MD_INT as it is unused. Maybe we still want this if we go back to converting PooledDataArray.refs to INTEGER in our C code?

  

  Peter M. Haverty

  committed
  918f098

  Copy full SHA for 918f098

• Julia_R_MD_NA_Factor works now doing conversion to Int32 in julia via evaluating a string (for now?)

  

  Peter M. Haverty

  committed
  9b4ff0a

  Copy full SHA for 9b4ff0a

• Julia_R_MD_NA_Factor in progress. Getting levels by calling 'levels' works. Some trouble converting integer values in factor vector.

  

  Peter M. Haverty

  committed
  08c497a

  Copy full SHA for 08c497a

Commits on Aug 2, 2016

• bitunpack is deprecated, generating a warning in Julia_R_MD_NA. Used Array(bitvector) as suggested by the warning message.

  

  Peter M. Haverty

  committed
  254612e

  Copy full SHA for 254612e

• Julia_MD_NA was not filling it its data and NA vectors because I was calling NewArray rather than Julia_MD. Fixed. Julia_MD_DataFrame_NA now fully functional as a result. All r2j functionality shou…

  Show description for 26dd6ce
  

  Peter M. Haverty

  committed
  26dd6ce

  Copy full SHA for 26dd6ce

• PooledDataArray creation does not need R to provide the NAs. It is simple enough to just compare to NA_INTEGER.

  

  Peter M. Haverty

  committed
  e5028f8

  Copy full SHA for e5028f8

• PooledDataArray creation now starts with making a DataArray. armgong had a nice optimization for this, but I do not know how to make it work without evaluating strings.

  

  Peter M. Haverty

  committed
  dbb200f

  Copy full SHA for dbb200f

• Closer to creating data.frame. Fixed my bug where I was using wrong variable as incoming list/data.frame. Also, simpler way to tell if column is factor.

  

  Peter M. Haverty

  committed
  cf1b97b

  Copy full SHA for cf1b97b

• DataFrames and DataArrays are pretty central, so we should probably load them at init time

  

  Peter M. Haverty

  committed
  6f28f30

  Copy full SHA for 6f28f30

• Fixed warning about windows-only initialization

  

  Peter M. Haverty

  committed
  0a34a5f

  Copy full SHA for 0a34a5f

Pagination

Previous

Next