Ruby 1.4.6 - trouble with require path

ruby-talk
1

I inherited a project that uses Ruby 1.4.6 on a RedHat Linux 6.1J. I
copied it onto my RedHat Linux 6.1J machine, but there is a problem
when I try to build it; the ruby scripts are unable to find required
files (located in the same directory as the scripts) when the script
is executed from a different directory. I upgraded to Ruby 1.6.7,
which fixed that problem, but the scripts are not quite compatible
with the updated Ruby, so I had to revert to the version that was
originally used.

As an example, lets say I have two ruby scripts in a directory:
     /scripts/script1.rb:
          #!/usr/bin/ruby
          require "script2"
     
     /scripts/script2.rb:
          print "Testing...\n"
     
     Command Prompt:
     [testing]# cd /scripts
     [scripts]# script1.rb
     Testing...
     [scripts]# cd /testing
     [testing]# script1.rb
     script1.rb:2:in `require': No such file to load -- script2
(LoadError)

With Ruby 1.6.7, it works fine. With Ruby 1.4.6, it can find
script1.rb, but not script2.rb.

There must be a way to correct this, as the project apparently worked
for the original developers. I'm using the same OS, and I added the
same ruby options and script directory to my path as specified in
their readme (well almost, they said put it in .cshrc, but I had to
put it in .tcshrc for it to work):
     setenv RUBYOPT '-Ke -rkconv'
     set path = ( $path /scripts)

I also tried the following:
     setenv RUBYOPT '-Ke -rkconv -S'
     setenv LOAD_PATH /scripts
     setenv RUBYLIB /scripts

So what am I missing here? Does anyone know how I can get this to work
correctly WITHOUT ALTERING THE SCRIPTS (because there are hundreds of
scripts)?

2

i can't help you at all, but i'm dying to know: what project have you
inherited with hundreds of 1.4.6?

-a

···

On Thu, 4 Nov 2004, Ken Innes wrote:

I inherited a project that uses Ruby 1.4.6 on a RedHat Linux 6.1J. I
copied it onto my RedHat Linux 6.1J machine, but there is a problem
when I try to build it; the ruby scripts are unable to find required
files (located in the same directory as the scripts) when the script
is executed from a different directory. I upgraded to Ruby 1.6.7,
which fixed that problem, but the scripts are not quite compatible
with the updated Ruby, so I had to revert to the version that was
originally used.

As an example, lets say I have two ruby scripts in a directory:
    /scripts/script1.rb:
         #!/usr/bin/ruby
         require "script2"

    /scripts/script2.rb:
         print "Testing...\n"

    Command Prompt:
    [testing]# cd /scripts
    [scripts]# script1.rb
    Testing...
    [scripts]# cd /testing
    [testing]# script1.rb
    script1.rb:2:in `require': No such file to load -- script2
(LoadError)

With Ruby 1.6.7, it works fine. With Ruby 1.4.6, it can find
script1.rb, but not script2.rb.

There must be a way to correct this, as the project apparently worked
for the original developers. I'm using the same OS, and I added the
same ruby options and script directory to my path as specified in
their readme (well almost, they said put it in .cshrc, but I had to
put it in .tcshrc for it to work):
    setenv RUBYOPT '-Ke -rkconv'
    set path = ( $path /scripts)

I also tried the following:
    setenv RUBYOPT '-Ke -rkconv -S'
    setenv LOAD_PATH /scripts
    setenv RUBYLIB /scripts

So what am I missing here? Does anyone know how I can get this to work
correctly WITHOUT ALTERING THE SCRIPTS (because there are hundreds of
scripts)?

--

EMAIL :: Ara [dot] T [dot] Howard [at] noaa [dot] gov
PHONE :: 303.497.6469
When you do something, you should burn yourself completely, like a good
bonfire, leaving no trace of yourself. --Shunryu Suzuki

===============================================================================

3

I inherited a project that uses Ruby 1.4.6 on a RedHat Linux 6.1J. I
copied it onto my RedHat Linux 6.1J machine, but there is a problem
when I try to build it; the ruby scripts are unable to find required
files (located in the same directory as the scripts) when the script
is executed from a different directory. I upgraded to Ruby 1.6.7,
which fixed that problem, but the scripts are not quite compatible
with the updated Ruby, so I had to revert to the version that was
originally used.

As an example, lets say I have two ruby scripts in a directory:
     /scripts/script1.rb:
          #!/usr/bin/ruby
          require "script2"
     
     /scripts/script2.rb:
          print "Testing...\n"
     
     Command Prompt:
     [testing]# cd /scripts
     [scripts]# script1.rb
     Testing...
     [scripts]# cd /testing
     [testing]# script1.rb
     script1.rb:2:in `require': No such file to load -- script2
(LoadError)

With Ruby 1.6.7, it works fine. With Ruby 1.4.6, it can find
script1.rb, but not script2.rb.

While I believe you, I don't see how ruby can see from /testing into
/scripts.

There must be a way to correct this, as the project apparently worked
for the original developers. I'm using the same OS, and I added the
same ruby options and script directory to my path as specified in
their readme (well almost, they said put it in .cshrc, but I had to
put it in .tcshrc for it to work):
     setenv RUBYOPT '-Ke -rkconv'
     set path = ( $path /scripts)

I also tried the following:
     setenv RUBYOPT '-Ke -rkconv -S'
     setenv LOAD_PATH /scripts

     setenv RUBYLIB /scripts

       ^^^^^^^^^^^^^^^^^^^^^^^

This should do the trick.

You can verify this with something like:

cd /testing

ruby14 -e 'p $:' -r /scripts/script1.rb

/scripts should be the first thing in the array.

So what am I missing here? Does anyone know how I can get this to work
correctly WITHOUT ALTERING THE SCRIPTS (because there are hundreds of
scripts)?

PS: I now have a ruby14 installed!

···

Ken Innes (primehalo@hotmail.com) wrote:

--
Eric Hodel - drbrain@segment7.net - http://segment7.net
All messages signed with fingerprint:
FEC2 57F1 D465 EB15 5D6E 7C11 332A 551C 796C 9F04

4

heh, I'm suprised they kept that many scripts without adjusting to other releases. I'm curious myself as well. Insane it is =]

David Ross

···

Ara.T.Howard@noaa.gov wrote:

On Thu, 4 Nov 2004, Ken Innes wrote:

I inherited a project that uses Ruby 1.4.6 on a RedHat Linux 6.1J. I
copied it onto my RedHat Linux 6.1J machine, but there is a problem
when I try to build it; the ruby scripts are unable to find required
files (located in the same directory as the scripts) when the script
is executed from a different directory. I upgraded to Ruby 1.6.7,
which fixed that problem, but the scripts are not quite compatible
with the updated Ruby, so I had to revert to the version that was
originally used.

As an example, lets say I have two ruby scripts in a directory:
    /scripts/script1.rb:
         #!/usr/bin/ruby
         require "script2"

    /scripts/script2.rb:
         print "Testing...\n"

    Command Prompt:
    [testing]# cd /scripts
    [scripts]# script1.rb
    Testing...
    [scripts]# cd /testing
    [testing]# script1.rb
    script1.rb:2:in `require': No such file to load -- script2
(LoadError)

With Ruby 1.6.7, it works fine. With Ruby 1.4.6, it can find
script1.rb, but not script2.rb.

There must be a way to correct this, as the project apparently worked
for the original developers. I'm using the same OS, and I added the
same ruby options and script directory to my path as specified in
their readme (well almost, they said put it in .cshrc, but I had to
put it in .tcshrc for it to work):
    setenv RUBYOPT '-Ke -rkconv'
    set path = ( $path /scripts)

I also tried the following:
    setenv RUBYOPT '-Ke -rkconv -S'
    setenv LOAD_PATH /scripts
    setenv RUBYLIB /scripts

So what am I missing here? Does anyone know how I can get this to work
correctly WITHOUT ALTERING THE SCRIPTS (because there are hundreds of
scripts)?

i can't help you at all, but i'm dying to know: what project have you
inherited with hundreds of 1.4.6?

-a
--

--
Hazzle free packages for Ruby?
RPA is available from http://www.rubyarchive.org/

5

PS: I now have a ruby14 installed!

If you go to the list archive at
  http://blade.nagaokaut.ac.jp/ruby/ruby-talk/index.shtml
and do a "(Subjects, Regular Expression)" search, the bottom of the second
frame still carries a graphic saying:

"Ruby 1.1: Enjoy Multi-thread Programming by Ruby."

:slight_smile:

6

Well, "setenv RUBYLIB /scripts" was the right solution. I don't quite
understand what was going on; RUBYLIB was set (as verified by typing
"setenv"), but Ruby couldn't find the scripts until I restarted the
computer. Then it seemed to work fine.

But now I'm having another weird problem where accessing the PATH
environment variable is causing Ruby not find scripts. For example:
     /scripts/script1.rb:
          #!/usr/bin/ruby
          print `script3.rb`

     /scripts/script2.rb:
          #!/usr/bin/ruby
          ENV["PATH"].split(/:/).each { |i|
               #do nothing
          }
          print `script3.rb`

     /scripts/script3.rb:
          #!/usr/bin/ruby
          print "Testing...\n"

     Command Prompt:
          [scripts]# cd /testing
          [testing]# script1.rb
          Testing....
          [testing]# script2.rb
          /testing/script2:4: command not found: script3

Any access to ENV["PATH"] seems to cause the problem, for example:
print ENV["PATH"]
tmp = ENV["PATH"]
ENV["PATH"].chomp!

Any ideas on this one?

7

David Ross <dross@code-exec.net> wrote in message news:<418AE839.9040401@code-exec.net>...

···

Ara.T.Howard@noaa.gov wrote:

> On Thu, 4 Nov 2004, Ken Innes wrote:
>
>> I inherited a project that uses Ruby 1.4.6 on a RedHat Linux 6.1J. I
>> copied it onto my RedHat Linux 6.1J machine, but there is a problem
>> when I try to build it; the ruby scripts are unable to find required
>> files (located in the same directory as the scripts) when the script
>> is executed from a different directory. I upgraded to Ruby 1.6.7,
>> which fixed that problem, but the scripts are not quite compatible
>> with the updated Ruby, so I had to revert to the version that was
>> originally used.
>>
>> As an example, lets say I have two ruby scripts in a directory:
>> /scripts/script1.rb:
>> #!/usr/bin/ruby
>> require "script2"
>>
>> /scripts/script2.rb:
>> print "Testing...\n"
>>
>> Command Prompt:
>> [testing]# cd /scripts
>> [scripts]# script1.rb
>> Testing...
>> [scripts]# cd /testing
>> [testing]# script1.rb
>> script1.rb:2:in `require': No such file to load -- script2
>> (LoadError)
>>
>> With Ruby 1.6.7, it works fine. With Ruby 1.4.6, it can find
>> script1.rb, but not script2.rb.
>>
>> There must be a way to correct this, as the project apparently worked
>> for the original developers. I'm using the same OS, and I added the
>> same ruby options and script directory to my path as specified in
>> their readme (well almost, they said put it in .cshrc, but I had to
>> put it in .tcshrc for it to work):
>> setenv RUBYOPT '-Ke -rkconv'
>> set path = ( $path /scripts)
>>
>> I also tried the following:
>> setenv RUBYOPT '-Ke -rkconv -S'
>> setenv LOAD_PATH /scripts
>> setenv RUBYLIB /scripts
>>
>> So what am I missing here? Does anyone know how I can get this to work
>> correctly WITHOUT ALTERING THE SCRIPTS (because there are hundreds of
>> scripts)?
>
>
> i can't help you at all, but i'm dying to know: what project have you
> inherited with hundreds of 1.4.6?
>
> -a
> --

heh, I'm suprised they kept that many scripts without adjusting to other
releases. I'm curious myself as well. Insane it is =]

David Ross

Can't go into details, but it's a project from Japan that was finished
years ago, probably when version 1.4.6 was current, or at least
near-current.

8

Looking at the code for ENV., I don't see anything that would cause
this, unless there's some bug in your libc with getenv.

Perhaps you should try a bash-like shell instead of a csh-like shell?

Oh, try this:

puts ENV["PATH"]
puts `env | grep path`

See if your PATH is really destroyed.

···

Ken Innes (primehalo@hotmail.com) wrote:

Well, "setenv RUBYLIB /scripts" was the right solution. I don't quite
understand what was going on; RUBYLIB was set (as verified by typing
"setenv"), but Ruby couldn't find the scripts until I restarted the
computer. Then it seemed to work fine.

But now I'm having another weird problem where accessing the PATH
environment variable is causing Ruby not find scripts. For example:
     /scripts/script1.rb:
          #!/usr/bin/ruby
          print `script3.rb`

     /scripts/script2.rb:
          #!/usr/bin/ruby
          ENV["PATH"].split(/:/).each { |i|
               #do nothing
          }
          print `script3.rb`

     /scripts/script3.rb:
          #!/usr/bin/ruby
          print "Testing...\n"

     Command Prompt:
          [scripts]# cd /testing
          [testing]# script1.rb
          Testing....
          [testing]# script2.rb
          /testing/script2:4: command not found: script3

Any access to ENV["PATH"] seems to cause the problem, for example:
print ENV["PATH"]
tmp = ENV["PATH"]
ENV["PATH"].chomp!

Any ideas on this one?

--
Eric Hodel - drbrain@segment7.net - http://segment7.net
All messages signed with fingerprint:
FEC2 57F1 D465 EB15 5D6E 7C11 332A 551C 796C 9F04

9

Eric Hodel <drbrain@segment7.net> wrote in message news:<20041106073536.GK93473@segment7.net>...

> Well, "setenv RUBYLIB /scripts" was the right solution. I don't quite
> understand what was going on; RUBYLIB was set (as verified by typing
> "setenv"), but Ruby couldn't find the scripts until I restarted the
> computer. Then it seemed to work fine.
>
> But now I'm having another weird problem where accessing the PATH
> environment variable is causing Ruby not find scripts. For example:
> /scripts/script1.rb:
> #!/usr/bin/ruby
> print `script3.rb`
>
> /scripts/script2.rb:
> #!/usr/bin/ruby
> ENV["PATH"].split(/:/).each { |i|
> #do nothing
> }
> print `script3.rb`
>
> /scripts/script3.rb:
> #!/usr/bin/ruby
> print "Testing...\n"
>
> Command Prompt:
> [scripts]# cd /testing
> [testing]# script1.rb
> Testing....
> [testing]# script2.rb
> /testing/script2:4: command not found: script3
>
> Any access to ENV["PATH"] seems to cause the problem, for example:
> print ENV["PATH"]
> tmp = ENV["PATH"]
> ENV["PATH"].chomp!
>
> Any ideas on this one?

Looking at the code for ENV., I don't see anything that would cause
this, unless there's some bug in your libc with getenv.

Perhaps you should try a bash-like shell instead of a csh-like shell?

Oh, try this:

puts ENV["PATH"]
puts `env | grep path`

See if your PATH is really destroyed.

My default shell is bash, but I change to csh to run their scripts (as
that's what it said to do in their readme file). Still, I have tried
it in both shells with the same results.

Anyway, I tried your suggestion, and here's what happens:

Try 1:
     /scripts/script2.rb:
          #!/usr/bin/ruby
          puts ENV["PATH"]
          puts `env | grep path`

     Command Line:
          [testing]% script2.rb
          /bin:/usr
          sh: env: command not found
          [testing]%

Didn't work because I referenced ENV["PATH"] so it couldn't find
"env".

Try 2:
     /scripts/script2.rb:
          #!/usr/bin/ruby
          puts ENV["PATH"]
          puts `/usr/bin/env | grep path`

     Command Line:
          [testing]% script2.rb
          /bin:/usr
          
          [testing]%

Probably put a blank line because of a difference in case.

Try 3:
     /scripts/script2.rb:
          #!/usr/bin/ruby
          puts ENV["PATH"]
          puts `/usr/bin/env | grep PATH`

     Command Line:
          [testing]% script2.rb
          /bin:/usr
          PATH=/bin:/usr
          [testing]%

The PATH environment variable still seems to be in tact, but Ruby
still can't find any files that are in PATH directories after
ENV["PATH"] is accessed.

···

Ken Innes (primehalo@hotmail.com) wrote:

10

Oops, actually, my path did get messed up. I didn't notice before
(must be Monday), but there are several directories missing from the
PATH, including the /scripts directory. So I just did another test:

     /scripts/script2.rb:
          #!/usr/bin/ruby
          puts `env | grep PATH`
          puts ENV["PATH"]
          puts `/usr/bin/env | grep PATH`

     Command Line:
          [testing]% script2.rb
          /bin:/usr/bin:/usr/local/bin:.:/scripts
          /bin:/usr
          PATH=/bin:/usr
          [testing]%

And I notice that there is no "/usr" initially, but it's there after
the ENV["PATH"] is referenced. I've searched all the files in my home
directory, and don't see it setting just "/usr" to the path anywhere.
I also search through the files in the Ruby directory, and also never
see a "/usr" by itself.

11

Hrm...

I'm not sure what the equivalent is on RedHat, but what about looking /etc/csh.*

There should be default a default .cshrc and .login there, do they have any odd paths?

···

On Nov 8, 2004, at 2:53 PM, Ken Innes wrote:

Oops, actually, my path did get messed up. I didn't notice before
(must be Monday), but there are several directories missing from the
PATH, including the /scripts directory. So I just did another test:

     /scripts/script2.rb:
          #!/usr/bin/ruby
          puts `env | grep PATH`
          puts ENV["PATH"]
          puts `/usr/bin/env | grep PATH`

     Command Line:
          [testing]% script2.rb
          /bin:/usr/bin:/usr/local/bin:.:/scripts
          /bin:/usr
          PATH=/bin:/usr
          [testing]%

And I notice that there is no "/usr" initially, but it's there after
the ENV["PATH"] is referenced. I've searched all the files in my home
directory, and don't see it setting just "/usr" to the path anywhere.
I also search through the files in the Ruby directory, and also never
see a "/usr" by itself.

12

I did more testing, and found that the path is getting truncated after
"/bin:/usr". My .tcshrc looked like so:
     set path = ( /bin )
     set path = ( $path /usr/bin )
     set path = ( $path /usr/local/bin )
     set path = ( $path . )
     set path = ( $path /scripts )

Then I added a few lines, and it still truncated at the same spot:
     set path = ( /bin )
     set path = ( $path /test1/usr )
     set path = ( $path /test1/usr/test2 )
     set path = ( $path /usr/bin )
     set path = ( $path /usr/local/bin )
     set path = ( $path . )
     set path = ( $path /scripts )

Gave this after executing the script:
     PATH=/bin:/test1/usr:/test1/usr/test2:/usr

However, it does not seem to truncate at all if put /usr as the first
one:
     set path = ( /usr )
     set path = ( $path /bin )
     set path = ( $path /usr/bin )
     set path = ( $path /usr/local/bin )
     set path = ( $path . )
     set path = ( $path /scripts )

I really don't understand this at all.