Re: [nant-dev] ILASM Task

[Ian M. <ianm@ActiveState.com>]

Gert Driesen wrote:

>----- Original Message -----
>From: "Ian MacLean" <ianm@ActiveState.com>
>To: "Gert Driesen" <ger...@pa...>
>Cc: "Giuseppe Greco" <giu...@ag...>;
><nan...@li...>
>Sent: Wednesday, June 30, 2004 4:17 PM
>Subject: Re: [nant-dev] ILASM Task
>  
>
>>Gert Driesen wrote:
>>
>>    
>>
>>>----- Original Message -----
>>>From: "Giuseppe Greco" <giu...@ag...>
>>>To: <nan...@li...>
>>>Sent: Wednesday, June 30, 2004 3:16 PM
>>>Subject: [nant-dev] ILASM Task
>>>
>>>
>>>
>>>      
>>>
>>>>Hi guys,
>>>>
>>>>I've given a lock at the CompilerBase and
>>>>ExternalProgramBase classes... and I think
>>>>it would be better to modify the CompilerBase
>>>>class instead of inheriting from ExternalProgramBase...
>>>>
>>>>
>>>>        
>>>>
>>>What exact functionality of CompilerBase do you need for the ilasm task ?
>>>The resource compilation stuff ? What else ?
>>>
>>>
>>>
>>>      
>>>
>>actually the resource compilation stuff does not apply. the /resources
>>flag of ilasm is for embedding a win32 resource block. .net resources
>>are defined as metadata in the il source.
>>See http://www.gotdotnet.com/Community/MessageBoard/Thread.aspx?id=151718
>>    
>>
>
>Don't have time to look into this much further, but I don't see CompilerBase
>and a task for ilasm having much in common ...
>
>  
>
I tend to agree. Looking quickly thru compilerBase I see the following 
attributes/elements:

lib
references
resources
modules
main
win32icon
define
target -- ilasm uses either /exe or /dll
warnaserror

In CompilerBase that have no relevance to ilasm. In fact the only thing 
that would be of any possible benefit is the NeedsCompiling method and 
even that is full of checks for resources/references etc.  I think 
you'll be much better off using ExternalProgramBase and copying across 
anything common.

Ian