I updated Ruby. When I request Ruby version in CLI, it works, but when I request for the Gem version, it returns the following error:
C:/Ruby23/lib/ruby/2.3.0/rubygems/config_file.rb:90:in `join': no implicit conversion of nil into String (TypeError)
The error is found on a portable version on Windows, as well as on an installed version.
I don't know what to do to run a working version of Ruby. Does anyone already got this bug or have a clue to resolve this?
You need to somehow set the environment variable SYSTEM_CONFIG_PATH
Here's line 90 of rubygems/config_file.rb:
SYSTEM_WIDE_CONFIG_FILE = File.join SYSTEM_CONFIG_PATH, 'gemrc'
That fails when SYSTEM_CONFIG_PATH is nil.
Searching the exact error message "no implicit conversion of nil into String", I finally found a working solution (may not be the best).
I replace the line 90:
SYSTEM_WIDE_CONFIG_FILE = File.join SYSTEM_CONFIG_PATH, 'gemrc'
With:
SYSTEM_WIDE_CONFIG_FILE = File.join SYSTEM_CONFIG_PATH.to_s, 'gemrc'
Related
I have the following ruby code:
EmailTemplate.for(mailer).each do |template|
begin
print '.'
template.upload(publish)
rescue Mandrill::UnknownTemplateError
failed.push(mailer)
end
end
Rubocop corrected my code to:
EmailTemplate.for(mailer).each do |template|
print '.'
template.upload(publish)
rescue Mandrill::UnknownTemplateError
failed.push(mailer)
end
and now it returns following error:
syntax error, unexpected keyword_rescue, expecting keyword_end
How can I fix that?
Rubocop warnings was:
C: Style/RedundantBegin: Redundant begin block detected.
Ruby 2.5.0 added a feature:
rescue/else/ensure are now allowed to be used directly with do/end blocks. [Feature #12906]
But before that, it was not allowed. So syntax error will be there.
Lets do syntax test for the code in sample.rb:
[].each do |a|
# ops
rescue Exception => ex
puts ex.inspect
end
From terminal:
Ruby$ ruby -c sample.rb
sample.rb:3: syntax error, unexpected keyword_rescue
rescue Exception => ex
^
sample.rb:5: syntax error, unexpected keyword_end, expecting end-of-input
Ruby$ rvm use 2.5.1
Using /Users/aruprakshit/.rvm/gems/ruby-2.5.1
Ruby$ ruby -c sample.rb
Syntax OK
See the News. So before 2.5.0, you need to write it like:
[].each do |a|
begin
# ops
rescue => Exception
puts ex.inspect
end
end
You can configure Rubocop to select the version of Ruby you want by following Setting the target Ruby version.
Some checks are dependent on the version of the Ruby interpreter which
the inspected code must run on. For example, enforcing using Ruby 2.3+
safe navigation operator rather than try can help make your code
shorter and more consistent... unless it must run on Ruby 2.2.
If .ruby-version exists in the directory RuboCop is invoked in,
RuboCop will use the version specified by it. Otherwise, users may let
RuboCop know the oldest version of Ruby which your project supports
with:
AllCops:
TargetRubyVersion: 2.4
For some reason, Rubocop thinks you're running Ruby 2.5, not Ruby 2.4.1.
You can fix this one of two ways:
1) Create a file .ruby-version with content 2.4.1. Rubocop should pick up your Ruby version from this file.
2) Add the following to your .rubocop.yml:
AllCops:
TargetRubyVersion: 2.4
I am having different output of following line with different version of ruby:
puts "#{"%c"%[254]}"
Output from ruby-2.4.1 : þ
Output from ruby-1.8.7 : �
It looks like issue with encoding, as there is no encoding in Ruby-1.8.7
Can someone please help me to get same result as Ruby-1.8.7 in Ruby-2.4.1 also?
This result is desirable because Ruby-1.8.7 gives me "254" after decoding while Ruby-2.4.1 gives me following error
`%': invalid value for Integer(): "þ" (ArgumentError)
I don't know much about Ruby 1.8.7, as I started Ruby with version 2.3, but this looks promising: https://gist.github.com/afair/2911107
But as everyone else is saying it would be best to upgrade your Ruby version if at all possible. Older versions of Ruby had limited Unicode support.
The following custom fact:
# returns latest packerversion, e.g. 0.10.1
Facter.add("latest_packerversion") do
setcode do
url="https://www.packer.io/downloads.html"
file = open("#{url}")
contents = file.read()
match = contents.match(/Latest\sversion:\s(.*)</)
match[1]
end
end
worked using puppet 3.6.2, but since the upgrade to 4.5.2 the following issue occurs:
Error: Facter: error while resolving custom fact "latest_packerversion":
No such file or directory # rb_sysopen - https://www.packer.io/downloads.html
Analysis
It seems that the rb_sysopen could not be found anymore for some reason (No such file or directory # rb_sysopen) since the upgrade to Puppet 4.
Puppet4 seems to use an embedded ruby version instead of the one installed on the host (Puppet3):
Puppet 4, both Facter 2.4 and CFacter 0.4, the latest Hiera and
Mcollective, as well Ruby 2.1.5, OpenSSL 1.0.0r, and our gem
dependencies.
Does rb_sysopen not exist in Ruby 2.1.5? No evidence was found.
Perhaps a change related to facts have occurred that could cause the issue? Nothing related was found in the release notes.
Question
Why could rb_sysopen not be found anymore by the custom fact since the upgrade to Puppet4?
Concise
I will include require 'open-uri' in both facts, but I do not understand why this is required since the upgrade to Puppet4
Verbose
Once require 'open-uri' is included in one of the custom facts the issue is solved.
# returns latest gitversion, e.g. 2.8.2
Facter.add("latest_gitversion") do
setcode do
require 'open-uri'
url="https://git-scm.com/downloads"
file = open("#{url}")
contents = file.read()
match = contents.match(/RelNotes.*((\d\.){2}\d)/)
match[1]
end
end
As soon as the require 'open-uri' has been commented out, the issue occurs again:
Error: Facter: error while resolving custom fact "latest_gitversion": No such file or directory # rb_sysopen - https://git-scm.com/downloads
Error: Facter: error while resolving custom fact "latest_packerversion": No such file or directory # rb_sysopen - https://www.packer.io/downloads.html
At the moment it is unclear what is causing the issue.
I am very new with GO, and when I tried to install the framwork beego on my Windows 8 Desktop or in a VirtualBox with Debian8 I had the same error:
sergi#odoo:~$ go get github.com/astaxie/beego
# github.com/astaxie/beego
work/src/github.com/astaxie/beego/tree.go:144: syntax error: unexpected range, expecting {
work/src/github.com/astaxie/beego/tree.go:155: syntax error: unexpected else, expecting semicolon or newline
work/src/github.com/astaxie/beego/tree.go:157: non-declaration statement outside function body
work/src/github.com/astaxie/beego/tree.go:158: non-declaration statement outside function body
work/src/github.com/astaxie/beego/tree.go:159: non-declaration statement outside function body
work/src/github.com/astaxie/beego/tree.go:160: syntax error: unexpected }
work/src/github.com/astaxie/beego/tree.go:257: syntax error: unexpected range, expecting {
work/src/github.com/astaxie/beego/tree.go:267: syntax error: unexpected else, expecting semicolon or newline or }
work/src/github.com/astaxie/beego/tree.go:283: syntax error: unexpected }
You seem to be using an old Go version which is failing to build the beego package.
Note that the package is only tested against go 1.5.1. So try upgrading to go 1.5.x and attempt to fetch it again.
https://github.com/astaxie/beego/blob/master/.travis.yml#L4
I'm pretty sure the upgrade will fix your issues.
The empty for range loops were introduced in Go 1.4. Your Go version seems to be older. Check your version with
go version
and if it's older than 1.4, install a newer version from the official website.
I encountered error:
"FATAL ERROR: Could not find system resources"
I then added the following code:
putenv("SWI_HOME_DIR=/usr/lib/swi-prolog");
I ran into another problem of:
"$c_call_prolog...Undefined procedure"
(More details here)
I ran the below command on Ubuntu 10.12
sudo apt-get install swi-prolog build-essential.
I wrote a simple C program to test it but was not able to retrieve the return value from prolog. Suppose I want to interact with mortal.pl, I compiled with:
swipl-ld -o mmm mortal.c mortal.pl
no error but like I mentioned. I did not get what I expected.
I obtained another code, nothing weird on the code. This time I got the first error. It seems that if swi-prolog is set to that directory, it won't search for my predicate.
Where and what should I set this SWI_HOME_DIR?
(Recap) with
putenv("SWI_HOME_DIR=/usr/lib/swi-prolog");
I called
mmm socrates
I got
$c_call_prolog...Undefined procedure
w/o putenv, the same call returned
FATAL ERROR: Could not find system resources
I finally changed
PL_initialise(1, plav)
where plav[0] = "mortal"; plav[1] = NULL; to
PL_initialise(1, argv)
I am able to obtain True/False from
rval = PL_call_predicate(NULL, PL_Q_NORMAL, pred, h0);
Though this confuses myself. What does PL_initialise() expect then?...not a to be called predicate?
No solution. I ended up installing ubuntu 12.10 64 bits (previously 32 bit). All those weird behavior are gone.