Sub definition
in [VB.NET]
|
From: Scott M. on 13 Sep 2009 15:35 I refer to them as compiler attributes because the compiler turns them into assembly metadata to then use as necessary. http://ondotnet.com/pub/a/dotnet/excerpt/vbnut_8/index1.html -Scott "Tom Shelton" <tom_shelton(a)comcastXXXXXXX.net> wrote in message news:erLUEDKNKHA.3412(a)TK2MSFTNGP04.phx.gbl... > On 2009-09-13, Scott M. <s-mar(a)nospam.nospam> wrote: >> The are markers that delinate a compiler "attribute", which is basically >> a >> special instruction just for the compiler. They are a way to mark, not >> only >> a Sub (you can use them on almost all class members), with extra >> information >> that the compiler needs to know. >> >> -Scott >> > > Hmmm.. While there are some attributes that affect compile time behavior > (for > example, ObsoleteAttribute) - most attributes are processed at runtime. > They > are markers that can be used by other classes to gain additional > information > about the class, property, or method. For instance, the attributes that > you > can place on the properties of a class to influence the behavior of the > PropertyGrid. > > One place I use them is for mapping object properties to fields in a > dataset > at runtime - sort of a simplified ORM tool. > > -- > Tom Shelton
From: Tom Shelton on 13 Sep 2009 20:32 On 2009-09-13, Scott M. <s-mar(a)nospam.nospam> wrote: > I refer to them as compiler attributes because the compiler turns them into > assembly metadata to then use as necessary. > > http://ondotnet.com/pub/a/dotnet/excerpt/vbnut_8/index1.html > > -Scott > Ok... that's true. They are put into the metadata of the assembly - but, I was mostly refering to your remarks about them being special instructions for the compiler. There are very few attributes that are recognized to have meaning to the compiler - beyond informing it to store the metadata. It seems a little unclear to describe a mostly runtime concept this way - at least it did to me, hence my comments :) -- Tom Shelton
From: Scott M. on 13 Sep 2009 22:14 But, technically, they are all instructions for the compiler. They tell the compiler to store this metadata about the type. I get what you are saying as far as what the attribute actually does for/to the type, but it is the compiler that creates this metadata in the first place. -Scott "Tom Shelton" <tom_shelton(a)comcastXXXXXXX.net> wrote in message news:erk1gLNNKHA.5072(a)TK2MSFTNGP05.phx.gbl... > On 2009-09-13, Scott M. <s-mar(a)nospam.nospam> wrote: >> I refer to them as compiler attributes because the compiler turns them >> into >> assembly metadata to then use as necessary. >> >> http://ondotnet.com/pub/a/dotnet/excerpt/vbnut_8/index1.html >> >> -Scott >> > > Ok... that's true. They are put into the metadata of the assembly - but, > I was > mostly refering to your remarks about them being special instructions for > the compiler. > > There are very few attributes that are recognized to have meaning to the > compiler - beyond informing it to store the metadata. It seems a little > unclear to describe a mostly runtime concept this way - at least it did to > me, > hence my comments :) > > -- > Tom Shelton >
From: Jack Jackson on 15 Sep 2009 00:48 Sure, but: x = 1 is an instruction to the compiler to generate some IL. By your definition, how is that different from what an attribute does, except that one affects metadata and the other IL. On Sun, 13 Sep 2009 22:14:11 -0400, "Scott M." <s-mar(a)nospam.nospam> wrote: >But, technically, they are all instructions for the compiler. They tell the >compiler to store this metadata about the type. > >I get what you are saying as far as what the attribute actually does for/to >the type, but it is the compiler that creates this metadata in the first >place. > >-Scott > >"Tom Shelton" <tom_shelton(a)comcastXXXXXXX.net> wrote in message >news:erk1gLNNKHA.5072(a)TK2MSFTNGP05.phx.gbl... >> On 2009-09-13, Scott M. <s-mar(a)nospam.nospam> wrote: >>> I refer to them as compiler attributes because the compiler turns them >>> into >>> assembly metadata to then use as necessary. >>> >>> http://ondotnet.com/pub/a/dotnet/excerpt/vbnut_8/index1.html >>> >>> -Scott >>> >> >> Ok... that's true. They are put into the metadata of the assembly - but, >> I was >> mostly refering to your remarks about them being special instructions for >> the compiler. >> >> There are very few attributes that are recognized to have meaning to the >> compiler - beyond informing it to store the metadata. It seems a little >> unclear to describe a mostly runtime concept this way - at least it did to >> me, >> hence my comments :) >> >> -- >> Tom Shelton >> >
From: Scott M. on 15 Sep 2009 02:11
Well sure, but attributes are a *different* way to talk to the compiler. Saying x =1 is using the main programming language. Attributes do not. From the article I referenced earlier: "You may wonder why attributes are used on the .NET platform and why they are not simply implemented as language elements. The answer comes from the fact that attributes are stored as metadata in an assembly, rather than as part of its executable code. As an item of metadata, the attribute describes the program element to which it applies and is available through reflection both at design time (if a graphical environment such as Visual Studio .NET is used), at compile time (when the compiler can use it to modify, customize, or extend the compiler's basic operation), and at runtime (when it can be used by the Common Language Runtime or by other executable code to modify the code's ordinary runtime behavior)." -Scott "Jack Jackson" <jjackson-nsp(a)cinnovations.net> wrote in message news:r17ua519ce3bnavp93aseh968gbdp9s54p(a)4ax.com... > Sure, but: > x = 1 > is an instruction to the compiler to generate some IL. By your > definition, how is that different from what an attribute does, except > that one affects metadata and the other IL. > > On Sun, 13 Sep 2009 22:14:11 -0400, "Scott M." <s-mar(a)nospam.nospam> > wrote: > >>But, technically, they are all instructions for the compiler. They tell >>the >>compiler to store this metadata about the type. >> >>I get what you are saying as far as what the attribute actually does >>for/to >>the type, but it is the compiler that creates this metadata in the first >>place. >> >>-Scott >> >>"Tom Shelton" <tom_shelton(a)comcastXXXXXXX.net> wrote in message >>news:erk1gLNNKHA.5072(a)TK2MSFTNGP05.phx.gbl... >>> On 2009-09-13, Scott M. <s-mar(a)nospam.nospam> wrote: >>>> I refer to them as compiler attributes because the compiler turns them >>>> into >>>> assembly metadata to then use as necessary. >>>> >>>> http://ondotnet.com/pub/a/dotnet/excerpt/vbnut_8/index1.html >>>> >>>> -Scott >>>> >>> >>> Ok... that's true. They are put into the metadata of the assembly - >>> but, >>> I was >>> mostly refering to your remarks about them being special instructions >>> for >>> the compiler. >>> >>> There are very few attributes that are recognized to have meaning to the >>> compiler - beyond informing it to store the metadata. It seems a little >>> unclear to describe a mostly runtime concept this way - at least it did >>> to >>> me, >>> hence my comments :) >>> >>> -- >>> Tom Shelton >>> >> |