How to add Unregisted Package in julia with continious Integration(gitlab-ci) - package

I have created a package Xnumber.And I have not registered it.
Now, I am creating another package(SHbundle) which uses the above package.Now in the julia REPL locally when I execute the following code:
julia> add "https://gitlab.com/vyush/Xnumber.jl.git"
julia> using Xnumber
It works fine locally and I can use the function but On pushing to SHbundle after adding Xnumber as dependency.
The pipeline script fail.
The command being executed is
- |
julia --project=#. -e '
using Pkg
Pkg.build()
Pkg.test(coverage=true)'
The error that I get is ERROR: expected package Xnumber [fdc6275c] to be registered. The package works fine locally but is giving error while executing the pipeline script.
I there any workaround for this without registering the packages.
The Links for these packages are: Xnumber,SHbundle

Remove the /Manifest.toml in your .gitignore file
Commit the Manifest.toml file after julia> add "https://gitlab.com/vyush/Xnumber.jl.git"
Add an extra step of Pkg.instantiate() in your pipeline

Related

Failed with error: ‘‘operator’ is not a valid installed package

Currently I am having problems with loading an R package that I downloaded from github (https://github.com/hafen/operator). I downloaded the package and put it in my local R folder: "C:/..../Documents/R/win-library/3.2/operator".
However, I have not been able to load said package into R. I keep on receiving the following errors when trying to load the package. Can anyone help me. I found an error that I though was somewhat relevant that suggest updating "Rcpp" which I have already done with no success. Thank you in advance for any and all help!
I should add. I looked into this a bit more and realized the issue I am having is revolving around the C script that is contained within the package. If I remove the C script I can load the package, however, some of the package functions are dependent on the script and will not run as a result.
> require(operator)
Loading required package: operator
Failed with error: ‘‘operator’ is not a valid installed package’
or
> devtools::load_all("C:/..../Documents/R/win-library/3.2/operator")
Loading operator
Re-compiling operator
"C:/PROGRA~1/R/R-32~1.2/bin/x64/R" --no-site-file --no-environ --no-save --no-restore --quiet CMD INSTALL \
"C:\Users\blakey1\Documents\R\win-library\3.2\operator" \
--library="C:\Users\blakey1\AppData\Local\Temp\Rtmpe611i9\devtools_install_499c5059149c" --no-R --no-data \
--no-help --no-demo --no-inst --no-docs --no-exec --no-multiarch --no-test-load
* installing *source* package 'operator' ...
** libs
Warning: running command 'make -f "C:/PROGRA~1/R/R-32~1.2/etc/x64/Makeconf" -f "C:/PROGRA~1/R/R-32~1.2/share/make/winshlib.mk" SHLIB="operator.dll" WIN=64 TCLBIN=64 OBJECTS="loess_op.o"' had status 127
ERROR: compilation failed for package 'operator'
* removing 'C:/Users/blakey1/AppData/Local/Temp/Rtmpe611i9/devtools_install_499c5059149c/operator'
Error: Command failed (1)
I was facing the same issue with RTextTools, as RTextTools is now not available on CRAN.
I have Download RTextTools From Given Link

https://cran.r-project.org/src/contrib/Archive/RTextTools/
and copy
RTextTools_1.4.2.tar.gz file in project root folder
then
run this command in project folder in terminal
"R CMD INSTALL RTextTools_1.4.2.tar.gz"
After running this command I receive below error
"ERROR: dependencies ‘SparseM’, ‘randomForest’, ‘tree’, ‘e1071’, ‘ipred’, ‘caTools’, ‘maxent’, ‘glmnet’, ‘tau’ are not available for package ‘RTextTools’"
Now install each dependencies from RStudio or RConsole (Any Editor used by you) by simply running this code
install.packages("caTools")
Install all 9 required packages One By One (In My Case it was 9 Packages Dependencies required by RTextTools)
all packages will be installed except 'maxent'
Now download maxent from the given link
https://cran.r-project.org/src/contrib/Archive/maxent/
and copy maxent_1.3.3.1.tar file in project folder then
run this command in project folder in terminal
"R CMD INSTALL maxent_1.3.3.1.tar"
Now For RTextTools Run this command again in Terminal
"R CMD INSTALL RTextTools_1.4.2.tar.gz"
All is done Now..
But the Last Step is
Load the RTextTools using
library(RTextTools)
You will see one more Error: Load SparseM
Now Loading SparseM use code below
library(SparseM)
and in the last Load RTextTools
library(RTextTools)

Error when adding jar dependence in Zeppelin

Spark works normally in Zeppelin but when I add a jar dependency and run something like:
val df = spark.read.json("path/file1")
I got the the following error:
com.fasterxml.jackson.databind.JsonMappingException: Jackson version is too old 2.5.4
If I run it a second time I got:
java.lang.NoClassDefFoundError: Could not initialize class org.apache.spark.rdd.RDDOperationScope$
(Not sure why the error changes for the second run)
If I do not attach the jar the code works normally. Is this connected with conflicts maybe?
Tried to do something like the following in the interpreter option in Spark but still the same error:
artifact: path/utils.jar
exclude: com.fasterxml.jackson.databind (Is this the correct way to write how to exclude jackson?)
Any insights?

Error arose when using Pkg.add() in Julia

I have a problem when trying to add a package in the terminal using Pkg.add, the terminal gives me the following error message:
fatal:your current branch 'master' does not have any commits yet
Error: failed process: Process(`git' --git-dir=````)
How do I fix it?
I've had the same issue with Gadfly. Installation failed and I couldn't load nor reinstall the package.
After looking at the documentation.
Turns out there is a cache folder (with the same name as the package) hidden in Users\my_username\.julia\v0.4\.cache\ I removed this folder and the package installed fine from Julia running Pkg.add("Gadfly").
Try removing the folder and install the package again.
Please give more information about where this package comes from. Github? Your hard drive? A zip file?
If this is one of your packages, just create a commit.
git add .
git commit -m 'First commit'
This should work.

R Missing Package error when uploading to shinyapps.io

My Shiny program works fine locally on my PC (Windows 8, RStudio 0.99.489) but not when I upload it to shinyapps.io . I've tried 2 ways of packaging the data for upload - saveRDS on each object and save.image on the entire environment. Either way, when I upload it to shiny.io I get:
Preparing to deploy application...DONE
Uploading bundle for application: 77966...DONE
Deploying bundle: 350891 for application: 77966 ...
Waiting for task: 132618597
building: Parsing manifest
building: Building image: 344796
building: Installing packages
################################ Begin Task Log ################################
[2016-01-16T22:19:45.818533554+0000] Installing R package: magrittr (1.5)
Error in library(stylo) : there is no package called �stylo�
Execution halted
################################# End Task Log #################################
Error: Unhandled Exception: Child Task 132618599 failed: Error building image: Build exited with non-zero status: 1
Execution halted
It seems that Shiny isn't aware of the Stylo package. I tried to install it in my code, but that didn't help.
1. Does Shiny have all R packages?
2. If not, is there a list of which packages are available?
Thanks very much.
Are you including the stylo package at the top of your server.R file via library("stylo")? If you are doing that and it's giving you the error, try using require("stylo") instead.
From the docs, the rsconnect package is supposed to automatically detect what packages are necessary for your app. Probably worth a read (if you haven't already) just to be sure you're following everything correctly in order for rsconnect to do its job.
I solved the problem by doing a clean build of my environment -- imported data tables and functions from RDS files only, and carefully avoided references to unnecessary packages. I recreated the one function I needed from Stylo locally, so that I could be certain I wouldn't require it.

Post-build arguments to batch script fails to convert build parameter

I'm using TFS 2013 and am building an Azure Cloud Service project that I want to package with Nuget so that I can publish to Octopus Deploy. I can't use octopack, because at the moment this is not supported. I'm trying to pass TFS parameters into a post-build batch script so that I can run nuget with -version parameters (which should change with each build).
The problem I'm having is that the batch script does not recognise the TFS build parameters. for example, in the script I want to pass an argument version $(TeamProject)-1.0.0$(Rev:.r), that would give the script the version to set in the package name.
The full nuget package call in the script is:
%nugetPath% push %packagePath%\Veedyo.%version%.nupkg
Passing this into the post-build script path works:
$/Application1/MAIN/Source/.NET/Application1.Package/package.cmd
This is because the source control path is translated into an actual path just after the build completes (and I can see the real path in the log).
However, the Post-build script arguments property in the build, doesn't convert this "$(TeamProject)"-1.0.0"$(Rev:.r)" to the desired value. this leads to an error executing the batch script:
Exception Message: TF270015: 'package.cmd' returned an unexpected exit code. Expected '0'; actual '1'. See the build logs for more details
So, does anyone have any idea how to convert add build parameter to the post-build script arguments property?
These macros are not available when you run the script. You can use one the the TF_BUILD environment variables listed here.
Probably, you are looking for TF_BUILD_BUILDNUMBER, or maybe you have to extract the data you are looking for; in this latter case Powershell can be simpler to use than cmd.exe interpreter.

Resources