presidentbeef / ffi-gdbm Goto Github PK

In other words, it is segfaulting on a 64bit system and not on a 32bit system.

No idea why, yet.

I'm getting this odd error now when I try to use the GDBM gem. The gem used to work. Now when I invoke the GDBM gem I get this error (I get it across many different versions of ruby):

dyld: lazy symbol binding failed: Symbol not found: __gdbm_lock_file
  Referenced from: /usr/local/lib/libgdbm.dylib
  Expected in: flat namespace

dyld: Symbol not found: __gdbm_lock_file
  Referenced from: /usr/local/lib/libgdbm.dylib
  Expected in: flat namespace

I've tried numerous ruby versions and installing/uninstalling GDBM from Brew.

The only major thing I can think has changed is that I'm now running on OSX El Capitan 10.11.6. Or that I somehow made a change to my config or PATH files that could have brought this on.

My GDBM Brew installation is at /usr/local/Cellar/gdbm/1.12

Hi there,

I was trying to follow this general guide on scraping, and had gotten everything working until I tried using this gem to dump JSON data to it.

After a lot of investigation I realized the issue is due to GDBM saving files in ASCII format, while ruby now relies primarily on UTF-8 encoding.

This code below independently replicates the problem I'm encountering:

#!/usr/bin/env ruby
# encoding: UTF-8
require 'gdbm'
require 'json'


# # Data Storage Formats
testdb = GDBM.new("test.db")       # GDBM file
dump   = {}                        # in memory

# # Test Data
str    = '√'                       # #<Encoding:UTF-8>



# # stays UTF-8
puts "-------Without GDBM:"
str.encoding                       # => #<Encoding:UTF-8>
dump[str] = JSON.dump(str)
puts dump[str].encoding            # => #<Encoding:UTF-8>


# # GDBM saves to ASCII-8BIT
puts "-------With GDBM:"
str.encoding                       # => #<Encoding:UTF-8>
testdb[str] = JSON.dump(str)
puts testdb[str].encoding          # => #<Encoding:ASCII-8BIT>

JSON.load(testdb[str])             # `parse': 795: unexpected token at '"?' (JSON::ParserError)

At the end you can see how the incorrect encoding will trigger an error when the text being parsed has UTF-8 characters.

Is there a way to have GDBM play nice with UTF-8?