Code Review: MutableString5

grujildo · May 9, 2008, 8:13pm

tfpt review /shelveset:MutableString5;REDMOND\tomat

A new implementation for Ruby MutableString and Ruby regular expression
wrappers.
This is just the first pass, w/o optimizations and w/o encodings
(Default system encoding is used for all strings).
Many improvements and adjustments will come in future, some hacks will
be removed.

Basic architecture:
MutableString holds on Content and Encoding. Content is an abstract
class that has three subclasses:

```
 StringContent
```

     Holds on an instance of System.String - an immutable .NET

string. This is the default representation for strings coming from CLR
methods and for Ruby string literals.

     A textual write operation on the mutable string that has this

content representation will cause implicit conversion of the
representation to StringBuilderContent.

     A binary read/write operation triggers a transition to

BinaryContent using the Encoding stored on the owning MutableString.

```
 StringBuilderContent
```

     Holds on an instance of System.Text.StringBuilder - a mutable

Unicode string.

     A binary read/write operation transforms the content to

BinaryContent representation.

     StringBuilder is not optimal for some operations (requires

unnecessary copying), we may consider to replace it with resizable
char[].

```
 BinaryContent
```

     A textual read/write operation transforms the content to

StringBuilderContent representation.

     List<byte> is currently used, but it doesn't fit many

operations very well. We should replace it by resizable byte[].

The content representation is changed based upon operations that are
performed on the mutable string. There is currently no limit on number
of content type switches, so if one alternates binary and textual
operations the conversion will take place for each one of them. Although
this shouldn’t be a common case we may consider to add some counters and
keep the representation binary/textual based upon their values.

The design assumes that the nature of operations implemented by library
methods is of two kinds: textual and binary. And that data that are once
treated as text are not usually treated as raw binary data later. Any
text in the IronRuby runtime is represented as a sequence of 16bit
Unicode characters (standard .NET representation). Each binary data
treated as text is converted to this representation, regardless of the
encoding used for storage representation in the file. The encoding is
remembered in the MutableString instance and the original representation
could be always recreated. Not all Unicode characters fit into 16 bits,
therefore some exotic ones are represented by multiple characters
(surrogates). If there is such a character in the string, some
operations (e.g. indexing) might not be precise anymore - the n-th item
in the char[] isn’t the n-th Unicode character in the string (there
might be escape characters). We believe this impreciseness is not a real
world issue and is worth performance gain and implementation simplicity.

Tomas

grujildo · May 10, 2008, 11:28am

This is a big old diff to search through. I couldn’t work out a way of
easily patching it onto my source at home due to the folder differences.

I really like this hybrid idea and it looks like it will work well. I
have
one question with regards to encodings and KCODE.

I appreciate that String is changing between Ruby 1.8 and 1.9. It
appears
that this MutableString implementation is leaning more toward the 1.9
implementation (i.e. holding on to an Encoding within the String
itself).

1.8 does hold the encoding and as I understand it the implicit encoding
of
the bytes held in a String is driven off KCODE. Is that correct? If so
you
have a number of scenarios which I think could cause problems with
MutableString holding on to its own Encoding, which stem from times when
KCODE is changed at runtime. I’ll try to describe a concrete example
and
you can tell me where I am going wrong…

Assume that KCODE is set to UTF8. If you create a String from an array
of
bytes in Ruby, the bytes are just stored as-is. You can do stuff which
is
encoding dependent and UTF8 is assumed.

If you now change KCODE to say EUC, then the bytes in the String are
unchanged but now encoding dependent operations will possibly produce
different results on the same string since they interpret the bytes
differently.

The worry I have with MutableString, is that if you create a string from
bytes but then do an operation that requires it to be converted to a CLR
string internally. What happens when you change KCODE? You can’t
simply
change the Encoding value of the MutableString, since if you then access
the
bytes you will not get the same bytes back as were originally put in. I
suppose, on changing KCODE, you could go through all the strings in
memory,
which have been converted from binary to CLR strings, and convert them
(i.e.
back to bytes via the old encoding and then to CLR strings via the new
encoding). What would be the optimal solution in this case?

Again, I am not talking from a position of deep knowledge here so I may
be
missing something really obvious. But I thought it was worth asking the
question.

Regards,

Pete

From: [email protected]
[mailto:[email protected]] On Behalf Of Tomas M.
Sent: Friday,09 May 09, 2008 19:08
To: IronRuby External Code R.
Cc: [email protected]
Subject: [Ironruby-core] Code Review: MutableString5

tfpt review /shelveset:MutableString5;REDMOND\tomat

A new implementation for Ruby MutableString and Ruby regular expression
wrappers.

This is just the first pass, w/o optimizations and w/o encodings
(Default
system encoding is used for all strings).

Many improvements and adjustments will come in future, some hacks will
be
removed.

Basic architecture:

MutableString holds on Content and Encoding. Content is an abstract
class
that has three subclasses:

```
 StringContent
```

     Holds on an instance of System.String - an immutable .NET

string.
This is the default representation for strings coming from CLR methods
and
for Ruby string literals.

     A textual write operation on the mutable string that has this

content representation will cause implicit conversion of the
representation
to StringBuilderContent.

     A binary read/write operation triggers a transition to

BinaryContent using the Encoding stored on the owning MutableString.

```
 StringBuilderContent
```

     Holds on an instance of System.Text.StringBuilder - a mutable

Unicode string.

     A binary read/write operation transforms the content to

BinaryContent representation.

     StringBuilder is not optimal for some operations (requires

unnecessary copying), we may consider to replace it with resizable
char[].

```
 BinaryContent
```

     A textual read/write operation transforms the content to

StringBuilderContent representation.

     List<byte> is currently used, but it doesn't fit many

operations
very well. We should replace it by resizable byte[].

The content representation is changed based upon operations that are
performed on the mutable string. There is currently no limit on number
of
content type switches, so if one alternates binary and textual
operations
the conversion will take place for each one of them. Although this
shouldn’t
be a common case we may consider to add some counters and keep the
representation binary/textual based upon their values.

The design assumes that the nature of operations implemented by library
methods is of two kinds: textual and binary. And that data that are once
treated as text are not usually treated as raw binary data later. Any
text
in the IronRuby runtime is represented as a sequence of 16bit Unicode
characters (standard .NET representation). Each binary data treated as
text
is converted to this representation, regardless of the encoding used for
storage representation in the file. The encoding is remembered in the
MutableString instance and the original representation could be always
recreated. Not all Unicode characters fit into 16 bits, therefore some
exotic ones are represented by multiple characters (surrogates). If
there is
such a character in the string, some operations (e.g. indexing) might
not be
precise anymore - the n-th item in the char[] isn’t the n-th Unicode
character in the string (there might be escape characters). We believe
this
impreciseness is not a real world issue and is worth performance gain
and
implementation simplicity.

Tomas

grujildo · May 10, 2008, 11:43pm

$KCODE is orthogonal to the encoding in MutableString. $KCODE seems to
be just a value that is used by some library methods that perform binary
operations on textual data. MutableString.Encoding is encoding of the
representation. If a MutableString instance is created from .NET string
an encoding that is associated with it is used whenever the string is
consumed by a binary data operation. We could represent all strings as
byte[], but then you’d need to convert .NET strings to byte[] at the
construction time. MutableString allows you to be lazy and perhaps not
perform the conversion at all if not needed.

Could you give some code sample that you think could be broken?

Tomas

From: [email protected]
[mailto:[email protected]] On Behalf Of Peter Bacon
Darwin
Sent: Saturday, May 10, 2008 2:27 AM
To: [email protected]
Subject: Re: [Ironruby-core] Code Review: MutableString5

This is a big old diff to search through. I couldn’t work out a way of
easily patching it onto my source at home due to the folder differences.
I really like this hybrid idea and it looks like it will work well. I
have one question with regards to encodings and KCODE.
I appreciate that String is changing between Ruby 1.8 and 1.9. It
appears that this MutableString implementation is leaning more toward
the 1.9 implementation (i.e. holding on to an Encoding within the String
itself).

1.8 does hold the encoding and as I understand it the implicit encoding
of the bytes held in a String is driven off KCODE. Is that correct? If
so you have a number of scenarios which I think could cause problems
with MutableString holding on to its own Encoding, which stem from times
when KCODE is changed at runtime. I’ll try to describe a concrete
example and you can tell me where I am going wrong…

Assume that KCODE is set to UTF8. If you create a String from an array
of bytes in Ruby, the bytes are just stored as-is. You can do stuff
which is encoding dependent and UTF8 is assumed.
If you now change KCODE to say EUC, then the bytes in the String are
unchanged but now encoding dependent operations will possibly produce
different results on the same string since they interpret the bytes
differently.
The worry I have with MutableString, is that if you create a string from
bytes but then do an operation that requires it to be converted to a CLR
string internally. What happens when you change KCODE? You can’t
simply change the Encoding value of the MutableString, since if you then
access the bytes you will not get the same bytes back as were originally
put in. I suppose, on changing KCODE, you could go through all the
strings in memory, which have been converted from binary to CLR strings,
and convert them (i.e. back to bytes via the old encoding and then to
CLR strings via the new encoding). What would be the optimal solution
in this case?

Again, I am not talking from a position of deep knowledge here so I may
be missing something really obvious. But I thought it was worth asking
the question.

Regards,

Pete

From: [email protected]
[mailto:[email protected]] On Behalf Of Tomas M.
Sent: Friday,09 May 09, 2008 19:08
To: IronRuby External Code R.
Cc: [email protected]
Subject: [Ironruby-core] Code Review: MutableString5

tfpt review /shelveset:MutableString5;REDMOND\tomat

A new implementation for Ruby MutableString and Ruby regular expression
wrappers.
This is just the first pass, w/o optimizations and w/o encodings
(Default system encoding is used for all strings).
Many improvements and adjustments will come in future, some hacks will
be removed.

Basic architecture:
MutableString holds on Content and Encoding. Content is an abstract
class that has three subclasses:

```
 StringContent
```

     Holds on an instance of System.String - an immutable .NET

string. This is the default representation for strings coming from CLR
methods and for Ruby string literals.

     A textual write operation on the mutable string that has this

content representation will cause implicit conversion of the
representation to StringBuilderContent.

     A binary read/write operation triggers a transition to

BinaryContent using the Encoding stored on the owning MutableString.

```
 StringBuilderContent
```

     Holds on an instance of System.Text.StringBuilder - a mutable

Unicode string.

     A binary read/write operation transforms the content to

BinaryContent representation.

     StringBuilder is not optimal for some operations (requires

unnecessary copying), we may consider to replace it with resizable
char[].

```
 BinaryContent
```

     A textual read/write operation transforms the content to

StringBuilderContent representation.

     List<byte> is currently used, but it doesn't fit many

operations very well. We should replace it by resizable byte[].

The content representation is changed based upon operations that are
performed on the mutable string. There is currently no limit on number
of content type switches, so if one alternates binary and textual
operations the conversion will take place for each one of them. Although
this shouldn’t be a common case we may consider to add some counters and
keep the representation binary/textual based upon their values.

The design assumes that the nature of operations implemented by library
methods is of two kinds: textual and binary. And that data that are once
treated as text are not usually treated as raw binary data later. Any
text in the IronRuby runtime is represented as a sequence of 16bit
Unicode characters (standard .NET representation). Each binary data
treated as text is converted to this representation, regardless of the
encoding used for storage representation in the file. The encoding is
remembered in the MutableString instance and the original representation
could be always recreated. Not all Unicode characters fit into 16 bits,
therefore some exotic ones are represented by multiple characters
(surrogates). If there is such a character in the string, some
operations (e.g. indexing) might not be precise anymore - the n-th item
in the char[] isn’t the n-th Unicode character in the string (there
might be escape characters). We believe this impreciseness is not a real
world issue and is worth performance gain and implementation simplicity.

Tomas

grujildo · May 11, 2008, 2:31pm

One thing that MutableString could do with is

    public static MutableString/*!*/ CreateBinary(byte[]/*!*/ bytes,

int
start, int length) {

At the moment you have to do something like:

    MutableString str = MutableString.CreateBinary();

    str.Append(buffer, 0, received);

Pete

From: [email protected]
[mailto:[email protected]] On Behalf Of Tomas M.
Sent: Saturday,10 May 10, 2008 22:42
To: [email protected]
Subject: Re: [Ironruby-core] Code Review: MutableString5

$KCODE is orthogonal to the encoding in MutableString. $KCODE seems to
be
just a value that is used by some library methods that perform binary
operations on textual data. MutableString.Encoding is encoding of the
representation. If a MutableString instance is created from .NET string
an
encoding that is associated with it is used whenever the string is
consumed
by a binary data operation. We could represent all strings as byte[],
but
then you’d need to convert .NET strings to byte[] at the construction
time.
MutableString allows you to be lazy and perhaps not perform the
conversion
at all if not needed.