Recent changes to feature-requests

Add arithmetic functions

Kevin Rogers — Fri, 28 Sep 2007 13:09:51 -0000

MillScript is currently missing the simple arithmetic functions, i.e. -, +, *, etc.

This can trigger an obscure bug with subtraction, e.g.

var x = 7;
x - 7;
There is 1 result
0

x- 7;
There is 1 result
0

x -7;
MISHAP : Could not find the definition for this variable
NAME : -
ORIGIN : origin
LINE NO. : 2
PHASE : Compile
Resetting input and resuming execution

x-7;
MISHAP : Could not find the definition for this variable
NAME : -
ORIGIN : origin
LINE NO. : 5
PHASE : Compile
Resetting input and resuming execution

This happens because the "-7" gets parsed as a negative integer, when it should be "- 7". The way this is dealt with in MillScript requires that the subtract function exists and is bound to "-".

Introduce context specific tokenization

Kevin Rogers — Thu, 27 Sep 2007 21:08:43 -0000

Following on from tracker item 1105231, we should update the tokenizer to handle context in a slightly more friendly way.

The current method of using the setWhere() and checkWhere() methods feels like a bodge, rather than a proper solution. Despite the potential of having to duplicate some code, I think the end result will be easier to understand and maintain.

I see two options:

1) Introduce a abstract tokenizer with three concrete sub-classes.
2) Introduce new methods to handle specific cases.

Improved multi-level indexByPrimaryKey function

Kevin Rogers — Fri, 21 Oct 2005 10:47:00 -0000

Hi,

It would be very useful to extend the indexByPrimaryKey
function to support multiple primary keys at the same
time, e.g.

var index = records.indexByPrimaryKey( "key1",
"key2" );

index[ "key1value" ][ "key2value" ];

suchthat only one record can exist for any combination
of primary key field values. This would help to avoid the
common situation where you know this to be the case but
have to use the group function instead, thus:

var index = records.group( "key1", "key2" );

index[ "key1value" ][ "key2value"][ 1 ];

It's only a subtle change, but it does make the code
neater.

Add propertiesFile function

Kevin Rogers — Fri, 16 Sep 2005 10:53:58 -0000

Add a propertiesFile function to match the properties file
loader in MillScript. This new function should either take
a map, or build one from it's arguments, then write the
map out as per Java's java.util.Properties#store() method.

This request arises as at the moment we can't properly
write a properties file where unsupported characters are
escaped.

Add support for Fat strings

Kevin Rogers — Tue, 13 Sep 2005 15:27:32 -0000

Steve has requested we add support for fat strings in the
next version. These are simple multi-line strings and
should be fairly easy to implement.

OpenSpice docs at:

http://www.openspice.org/docs/moin.cgi/DesignNotes_2fStrings

Complete XML namespace support in MillScript

Kevin Rogers — Mon, 05 Sep 2005 10:38:48 -0000

Hi,

Before the 10.2.0 release we need to complete XML
namespace support. This requires the following two
points to be addressed:

1) MillScripts internal structures need to be upgraded to
support namespaces, including remembering any prefix
specified in the source XML. It would be nice to
preserve the prefixes for the programmers benefit.

2) The new XML parser needs to have its namespace
support completed. Currently it parses qualified names
properly, but it doesn't recognise the xmlns attributes.
Namespace support should be achieved through a map
which can hold multiple values for a given key, e.g. a
"stack" hash map, where you can push and pop values
for a given key.

Add relational chains

Kevin Rogers — Fri, 05 Aug 2005 13:47:33 -0000

Add support for relational chains, e.g.

if 5 < x < 72 and a < b < c then
...
endif

This is already planned for the next release and is part of
the OpenSpice specification, but this RFE is just so I
don't forget to do it :-)

Add support for loading Excel spreadsheets

Kevin Rogers — Fri, 05 Aug 2005 13:41:39 -0000

It would be very useful to be able to autoload Excel
spreadsheets. There is a suitable piece of LGPL
software, called ExcelRead:

http://www.andykhan.com/jexcelapi/index.html

This also allows generating and modifying spreadsheets
too, so we should provide support for that too.

I would suggest a similar approach to datafiles, i.e. a
spreadsheet is a list of lists of lists, or less concisely a
list of worksheets where each item is a list of rows where
each item is a list of cells.

Add command line completion support to interpreter

Kevin Rogers — Fri, 22 Apr 2005 12:23:28 -0000

It would be great to add support for command line
completion to the interpreter. I'd consider the following
improvements:

1) Add completion for any variables in the standard
package. i.e. just make a completor for the standard
package.
2) Add completion for any variables in other packages.
This is a bit harder, we need a completor that can
understand qualified names. e.g. somepkg::avar. In this
case the completion needs to search the current
package, then imported packages, unless it's a qualified
name that's being completed(in which case the search is
specific to that package)

Logging support

Kevin Rogers — Wed, 13 Apr 2005 16:48:04 -0000

We need to start investigating good logging support.

NOTES:
1) Log to the VFS so we can use any target supported
by it. This may introduce a problem as some protocols
can't complete a file until it has been finished. Depending
on the log target, e.g. HTTP, this may cause the log to
be lost in the event of failure as the request cannot be
sent until the body content is complete

2) Consider how to handle common logging ideas without
having to insert too much explicit logging code. e.g.
specific function entry/exit and associated arguments
and return values

3) Consider supporting logging as core syntax, by that I
mean allow MillScript to optimise away any unused
logging methods, so they don't even get past the Expr
stage.