 |
 |
|
|
       |
|||
|
|||||||||
|
|
|||||||||
| What was new | |||||||||
OmniMark 5.3
General
OmniMark 5.3 introduces the Binary Coded Decimal (BCD) data type. BCD is the correct data type for performing many mathematical calculations and all types of financial calculations.
OmniMark's BCD implementation supports large numbers, limited only by available memory and 16 decimal places of accuracy. In addition, new formatting operations provide a way to customize the representation of BCD numbers according to the local conventions of most countries.
To use the BCD data type, you must include the public library "ombcd.xin". Using BCD is syntactically similar to mathematical operations on counters, although in some circumstances the BCD number must be "cast".
The IDE now accommodates BCD numbers.
OmniMark language
Heralds removed
Heralds have been removed from the language to make room for "casting". Also removed are the declaration declare heralded-names and the command-line option "-heralds".
V2 programs will no longer compile. To make them compile, you need to declare all variables and remove the words counter, stream, buffer, and switch that precede variables in your program.
Operator syntax
Operator recognition has also been changed to use longest match. For example, prior to OmniMark 5.3 the style of the following find rule was legal:
find "a"+=>x
In OmniMark 5.3, the sequence "+=>" is still recognized as "a"+ followed by =>, but a warning is issued stating that this is deprecated. In a future version of OmniMark this will be reported as an error, unless "+=>" is defined as an operator. To remove the warning, write the code like this:
find "a"+ => x
#program-error location argument
The format of the location argument of the #program-error catch has changed from "on line <line-number> in file <file-name>:%n" to "line <line-number>, file <file-name>".
Integers
OmniMark introduces two new integer data types called integer and int32. Integer will replace the counter data type in all cases except bit operations and "network long" calculations, which will be handled by int32. The default value for both integer and int32 is 0.
The introduction of the integer data type is a phased release. Programmers should use integer instead of counter, and begin phasing out the use of counter in existing programs.
Warnings on deprecated keywords
The several deprecated keywords now cause warning messages. The deprecated keywords, and their replacements, are as follows:
sgml is deprecated in favor of #markup-parser
output is deprecated in favor of #current-output
is equal is deprecated in favor of =
isnt equal is deprecated in favor of !=
is greater-than is deprecated in favor of >
isnt greater-than is deprecated in favor of <=
is greater-equal is deprecated in favor of >=
isnt greater-equal is deprecated in favor of <
is less-than is deprecated in favor of <
isnt less-than is deprecated in favor of >=
is less-equal is deprecated in favor of <=
isnt less-equal is deprecated in favor of >
= (as pattern assignment operator) is deprecated in favor of =>
Libraries
LDAP
LDAP is an acronym for "Lightweight Directory Access Protocol", a client-server protocol to retrieve and manage directory information. The OmniMark LDAP library contains functions that let you create connections to different LDAP data sources and let you display, create, modify, and delete entries in those sources. This library is available only to OMDN members.
OmniMark 5.2
OmniMark language
You can now test to see if a catch is available before initiating a throw using the is catchable test.
You can now test to see if a throw is in progress using the is thrown test.
The using input as keyword has been added to allow you to change the current input scope without initiating a scan.
The with utf-8 modifier has been added to do xml-parse and do sgml-parse to allow you to specify how the parser handles character entities greater than 127.
HTML 4.0
You can now process HTML 4.0 as SGML. The SGML parser in OmniMark 5.2 now supports large character values and other SGML features used in the HTML 4.0 prolog. You will need to use an appropriate HTML 4.0 SGML declaration when parsing HTML 4.0 files.
Case-sensitive XML parsing
XML parsing is now case sensitive.
Namecase general and namecase entity obsolete
OmniMark now reads the NAMECASE GENERAL and NAMECASE ENTITY declarations from the SGML declaration. The OmniMark namecase general and namecase entity declarations are now obsolete.
For this reason it is now a compile-time error to have a namecase general no declaration in a program that uses attribute aliases. If you get this error, check carefully for any ambiguity in your use of attribute aliases before removing namecase general no from your program.
SGML quantities
The OmniMark SGML parser now accepts SGML files that use element names longer than eight characters without the need to supply an SGML declaration. A number of other SGML quantities also have larger default values.
Empty pattern matching
It is now permissible, in all situations, to write a pattern that matches zero characters. This is particularly useful for detecting the end of nested structures when you are doing nested pattern matching.
include-guard
The new include-guard makes it easier to ensure that a file is not included twice in your source code.
Large file support
On Solaris 2.6, OmniMark will now correctly handle files that contain more than 2 gigabytes of data.
Library changes
Extended Database Xpack
The new Database Extensions library in this OmniMark Developer's Network Xpack offers support for stored procedures and transaction processing.
HTTP library
The HTTP library now supports proxy servers.
Run-time environment
Project files as arguments files
You can now use OmniMark IDE project files as arguments files with OmniMark VM, OmniMark C, and OmniMark C/VM.
Library, include, and temporary file paths
Library, include, and temporary file paths can now be set globally using environment variables and/or Windows registry entries. In Windows, suitable default values are set during installation. See Developing and running OmniMark programs.
-noenvarg command-line option
You can use -noenvarg, or its short form, -noea, to tell OmniMark to ignore the library, include, and temporary file path settings in the environment.
OmniMark 5
OmniMark 5 includes several new language features and development tools.
Language features
Catch & Throw
You can use catch and throw to manage the execution flow in your OmniMark programs. Catch and throw is a powerful addition to the flow handling features of OmniMark, allowing you to make major redirections of program flow in a safe and structured way.
Buffered, Unbuffered, Flush
Buffered and unbuffered are two new stream modifiers that allow you to
control the buffering characteristics of the input and output streams your
OmniMark program uses. The ability to declare the #main-input stream as
unbuffered is one of the key features that allow you to do CGI programming with OmniMark. Accompanying buffered and unbuffered is the new flush keyword, which lets you force a buffered stream to send the data it contains.
Nested pattern matching
The new keyword #current-input has been added. This lets you use the technique of nested pattern matching to simplify the processing of data that contains nested structures.
Pattern matching functions
You can now write pattern matching functions, functions that participate in the pattern matching process. This greatly enhances the power and flexibility of OmniMark pattern matching.
Direct support for CGI Programming
OmniMark now supports the directive used by UNIX shells and
some web servers to associate a script with the appropriate runtime. This
greatly eases the writing of CGI programs in OmniMark.
Default variable shelf size is now 0
The default size of a shelf declared variable is now 0. It
was 1 in previous versions. 0 is the appropriate default in almost all
programming situations.
Command-line option changes
-sb command-line option added
The new -sb command-line option is a combination of "-s" and "-brief". You can use this option to specify the name of the OmniMark program to execute, and to suppress the OmniMark banner information normally output when you run an OmniMark program from the command line. This command-line option was included to use in CGI programs.
-expand command-line option removed
The -expand command-line option has been removed from this release. This was done to address security concerns when using OmniMark to run CGI programs. The functionality provided by -expand is no longer required now that the IDE is available.
Libraries
Database
A new high-level database library is included in this release. It is much easier to use than the ODBC library, which is still available for use in applications that require its low level functionality.
CGI
Developing CGI applications with OmniMark is a simple process made simpler with the addition of the CGI library. This library provides routines to retrieve and decode the information that the web server sends using the CGI protocol.
OMX components
We have renamed opaque objects OmniMark eXtension (OMX) components. OMX components extend the functionality of OmniMark the same way Active X and other component types are used to extend other programming languages. OMX components perform the same essential function as other component technologies: they encapsulate complex functionality and allow the OmniMark programmer to easily deal with complex systems such as FTP servers, databases, and TCP/IP connections. In keeping with the streaming nature of the OmniMark language, many OMX components are capable of acting as sources and sinks. You can read from or write to the OMX just as you would to a file or buffer. This greatly simplifies many aspects of network programming.
Development tools
Integrated Development Environment
The OmniMark IDE (available on Windows 95/98/NT) is a graphical user interface application that includes a full suite of editing and debugging tools.
The IDE editing environment:
The IDE debugging tools let you:
OmniMark VM
OmniMark VM is a stand-alone version of the OmniMark virtual machine. You can use it to run compiled OmniMark byte codes files (*.xvc).
OmniMark 4.0
OmniMark 4.0 includes numerous new features and changes that have been implemented to extend functionality and increase program flexibility.
XML document processing
OmniMark 4.0 supports XML parsing, a variant of SGML parsing that supports the key features of the Extensible Markup Language (XML) specification (as defined in WD-xml-970807). The addition of this feature allows OmniMark to successfully parse a valid XML document. With the addition of this feature, some changes and additions were made to the OmniMark language. Most significantly, a set of generic "markup" keywords was introduced to generalize, as much as possible, OmniMark functionality that was previously specific to the SGML parser. With this generalization, programmers can interact with a parser without having to specify which one, SGML or XML.
Opaque types
<Opaque data types have been added, allowing programmers to define their own data types to be used with external functions.
The "no-default-io" declaration
This declaration has been added to simplify the process of writing and testing server programs. OmniMark programs without default input/output sources and destinations are expected to be servers with their own explicitly defined and controlled inputs and outputs. Using the no-default-io declaration allows you to run an OmniMark CI program in a simulated virtual machine environment.
Format items and the format operator
The format operator has been added to increase the flexibility of format item usage. This feature reduces the amount of coding required in complex output formatting tasks. With the implementation of this operator, format modifiers can also be dynamically generated rather than having to be statically coded into a program.
Changes to "find" and "submit"
With this release of OmniMark, find rules and the submit action can be used almost anywhere. The most important effect of this change is that submit can now be used in the body of process rules.
New date format item
An "=t" has been added to the format items that can be used with the date and compiled-date keywords. This format item produces the offset in hours and minutes from Greenwich Mean Time.
New command-line options
An "-x" command-line option has been added. This option allows programmers to specify the path of external function libraries from the command line rather than having to "hard code" them into an OmniMark program. This option can also be used to increase cross-platform portability of OmniMark programs that use external function libraries. The file-name extension of the external function library (which is different on most platforms) can also now be specified on the command line rather than in the OmniMark program itself.
The "-stats" command-line option has also been added. When specified, this option causes an OmniMark program to generate and output data regarding memory usage and execution times.
External functions changes
Two changes have been made to external functions. External functions now accept remainder and optional arguments. Additionally, external functions support generic functions that make use of various return types, argument counts, and argument types determined at run-time. The external functions API has also undergone some changes.
Function libraries
The following have been added to the OmniMark function libraries suite. Please note that some of these libraries can be used only in OmniMark 4.0 and beyond because they utilize opaque types:
| ---- |
