Permalink
Commits on Apr 25, 2012
  1. validate sub-messages

    codekitchen committed Apr 25, 2012
    fixes #14
  2. remove the no-longer-used gen_methods! call in .pb.rb files

    codekitchen committed Apr 25, 2012
    the method is left on the Message class, so that already-generated
    .pb.rb files don't break
Commits on Jan 21, 2012
  1. Merge pull request #18 from mculp/master

    codekitchen committed Jan 21, 2012
    add an attr_reader for proxy_class in message fields
Commits on Jan 16, 2012
  1. Merge pull request #17 from iseatz/master

    codekitchen committed Jan 16, 2012
    Fix a few bugs. Thanks Mark!
Commits on Jan 9, 2012
Commits on Jan 4, 2012
Commits on Jan 3, 2012
  1. Adds Rspec tests to demonstrate various bugs in the protobuffer

    mherman-iseatz committed Jan 3, 2012
    library.
    
    On code without the fixes, certain of these tests are expected to
    fail.  All tests should pass after merging in the fix branch.
  2. Fixes Ruby 1.9 encoding on returned Strings.

    mherman-iseatz committed Jan 3, 2012
    Ruby 1.9 associates an Encoding object with each string.  Strings were getting an ASCII Encoding before.  Protobuffers are only supposed to return UTF-8 strings.
  3. Uses bytesize to determine string length in encoder rather than length.

    mherman-iseatz committed Jan 3, 2012
    The length function on Ruby 1.9 Strings returns the number of characters rather than the number of bytes.  Ruby 1.8 and 1.9 Strings both have a bytesize function that returns the number of bytes.  This makes the encoder use bytesize.
    
    It should fix decoding errors on UTF-8 strings that contain multi-byte characters.
Commits on Dec 28, 2011
  1. Better yet, a one-line negative fix. This preserves max range checking.

    mherman-iseatz committed Dec 28, 2011
    If an int32 exceeds it's proper max range, the 64-bit max is subtracted
    from it.
  2. Fixes bugs in previously committed negative number handling fix.

    mherman-iseatz committed Dec 28, 2011
    After further evaluation, I found that my new encoder was making the
    numbers one bit too small.  The problem with negative number decoding
    wasn't that either the encoder or decoder was not working.  The
    negative number problem turned out to be specifically for 32-bit numbers.
    Negative int32 numbers encode to a full 64-bits.  The deserializer was
    assuming they were 32-bit.  This commit assumes all negative int32/64
    numbers are 64-bit.
Commits on Dec 23, 2011
  1. Fixes handling of negative numbers in int32 fields.

    mherman-iseatz committed Dec 23, 2011
    Fixes handling of negative numbers in int32 fields.  I'm sure this can be optimized, but it should at least give direction on how things should work.
Commits on Dec 22, 2011
  1. Merge pull request #15 from iseatz/master

    codekitchen committed Dec 22, 2011
    More Descriptive Error Handling On Encoding Failures
Commits on Dec 21, 2011
  1. Fixes bug in handling of nil fields

    mherman-iseatz committed Dec 21, 2011
    Fixes a bug that was causing the encoder to skip fields after other fields were set to nil.
Commits on Dec 4, 2011
  1. Bubble up the invalid field in the EncodeError

    beckel-iseatz committed Dec 4, 2011
    We often get invalid messages and the error does not say anything about
    which field was invalid. This patch adds a validate! method to Message
    and Message's instance variables that throws an exception rather than
    returning false like Message#valid? . It also adds the invalid field to
    the EncodeError so that it can be inspected.
  2. Bubble up the invalid field in the EncodeError

    beckel-iseatz committed Dec 4, 2011
    We often get invalid messages and the error does not say anything about
    which field was invalid. This patch adds a validate! method to Message
    and Message's instance variables that throws an exception rather than
    returning false like Message#valid? . It also adds the invalid field to
    the EncodeError so that it can be inspected.
Commits on Oct 1, 2011
  1. version 1.2.1

    codekitchen committed Oct 1, 2011
  2. version 1.2.0

    codekitchen committed Oct 1, 2011
  3. fix ruby 1.8.7 compat

    codekitchen committed Oct 1, 2011
  4. underscores and dots in package names

    rmarable-iseatz authored and codekitchen committed Sep 30, 2011
    camelcase according to the common rails convention for underscores, and
    treat dots as a new submodule,
    
    e.g. package.sub_package -> Package::SubPackage
    
    Squashed commit of the following:
    
    commit 5279440
    Author: Rob Marable <rmarable@iseatz.com>
    Date:   Fri Sep 30 18:25:25 2011 -0500
    
        Fixed my slopdoodle
    
    commit 7492656
    Author: Rob Marable <rmarable@iseatz.com>
    Date:   Fri Sep 30 17:03:44 2011 -0500
    
        Added another test
    
    commit 9da5970
    Author: Rob Marable <rmarable@iseatz.com>
    Date:   Fri Sep 30 15:49:03 2011 -0500
    
        Tweak to #field_typename
    
    commit 9fe22e8
    Author: Rob Marable <rmarable@iseatz.com>
    Date:   Fri Sep 30 14:44:19 2011 -0500
    
        Simplified camelize
    
    commit 852c3b9
    Author: Rob Marable <rmarable@iseatz.com>
    Date:   Fri Sep 30 14:21:08 2011 -0500
    
        Adding compiler support for underscores and dots in package declarations
    
    Signed-off-by: Brian Palmer <brian@codekitchen.net>
  5. don't automatically include proto file dirs as -I options

    codekitchen committed Sep 30, 2011
    This can cause errors in protoc in certain situations, and doesn't match
    the behavior of protoc very well anyway
Commits on Sep 30, 2011
  1. Merge pull request #12 from pcasaretto/master

    codekitchen committed Sep 30, 2011
    Fixing #11
Commits on Sep 29, 2011
  1. ruby-protoc now generates directories mirroring the target files

    pcasaretto committed Sep 29, 2011
    This fixes #11
    Its a it ugly, could probably use some refactoring
Commits on Sep 20, 2011
  1. version 1.1.0

    codekitchen committed Sep 20, 2011
  2. remove package unloading support, fixes #7

    codekitchen committed Sep 20, 2011
    This breaks loading multiple .pb.rb files into the same package
    (module) namespace, and is better done by your loading framework like
    ActiveSupport anyway.
Commits on Mar 29, 2011