An elaborate wrapper around the Define column is jedi.inc which is used in many projects (both open source and closed source) to distinguish between various Delphi versions, libraries and platforms at compile time (URL: github.com/project-jedi/jedi/blob/master/jedi.inc)
Empty arrays are not used often as arrays usually are about the presence data, not about the absence.
Here are two ways based on the int data type in C# (the original [WayBack] examples [WayBack] are using string, but since string itself is also a kind of array…):
Specify a size of zero:
int[] a = new int[0];
Specify an empty initialisation:
int[] a = new int[] { };
Though many people think arrays are a thing of the past, I think it is one of the first generic types and have their place. For one, enumerating over arrays using foreach is a lot faster in many environments than enumerating over other data types. Another thing is that the fixed nature of arrays can be very beneficial in setting constraints.
It comes down to these cases for XML elements having maxOccurs="1" (which the default for maxOccurs):
adding nillable="true" will convert from a regular type to a nullable type.
adding minOccurs="0" will add boolean …Specified properties in the generated C# for each element.
you can have both nillable="true" and minOccurs="0" in an element which gets you a nullable type and a …Specified property.
Note I’m not considering fixed or default here, nor attributes (that have use instead of minOccurs/maxOccurs, but do not allow for nillable) nor larger values of maxOccurs (which both xsd.exe and xsd2code regard as unbounded).
From the above, XML has a richer type system than C#, so in XML there are subtle a differences between:
an explicit nil in the XML element
the XML element being absent
the XML element being empty.
Hopefully later more text and examples to show how to actually work with this.
A while ago I bumped into this interesting bit: LLLPG (Loyc LL(k) Parser Generator) is a new recursive-decent parser generator for C#, with a feature set better than ANTLR version 2.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters. Learn more about bidirectional Unicode characters
It is next to the “Project Dependencies” in this image from Sara Ford:
Sara Ford: change “Project Build Order”
In the resulting dialog, you can change the build order within your solution.
This can be very useful when – for various reasons – you cannot have Project Level dependencies for an assembly, but have to have Assembly Reference dependencies for individual assemblies.
At a client I bumped into this, and this dialog was a life saver for us.
I tend to forget this: xsd.exe can resolve xs:include, but not xs:import. When using xs:import it will complain about missing types.
Simple solution: reference all imported XSDs (but not included XSDs) on the same command-line:
I was trying to create a C# class with xsd.exe from an schema that imports others schemas like that xsd.exe /c schema.xsd and it fails miserably reporting an error like “The datatype ‘xxx’ is missing.”. Basically xsd.exe doesn’t resolve the schemaLocation attribute.
The solution isn’t very nice, but at least it’s simple. You have to tell xsd.exe all the schema referenced, xsd.exe /c schema.xsd importedSchema1.xsd importedSchema2.xsd
Be aware that the C# filename is sometimes generated from all XSD filenames on the command-line (I’ve noticed this when the import is specified before the actual XSD file, if the import is at the end, the name of the import is used.
If my memory serves me right, older versions of XSD.exe could not resolve unix-style relative paths on the command-line, but newer versions do.
The Wiert Corner - irregular stream of stuff 