I've just done an updated version of Dfilter geared towards Doxygen 1.4.2.  It's actually turned out to be more lines of code than the previous version - as it happens, Doxygen is still even less capable of parsing D code than I thought it would be.

http://bugzilla.gnome.org/show_bug.cgi?id=173045

Main changes:
- merged in James Dunne's additions to convert versions to preprocessor
directives (pending a better solution)
- got rid of filtering details that aren't necessary anymore
- completely revamped support for attribute blocks - now works on all
attributes, not just access ones
- removes more things that Doxygen can't make any sense of (align and
pragma attributes, /+...+/ comments)

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on the 'group where everyone may benefit.

Stewart Gordon wrote:

> I've just done an updated version of Dfilter geared towards Doxygen 1.4.2.  It's actually turned out to be more lines of code than the previous version - as it happens, Doxygen is still even less capable of parsing D code than I thought it would be.

Nice, here is a patch to make it compile with GDC too...

Fixes:
	expression expected, not 'TOK36' (#36 is dollar, by the way)
	function std.c.stdio.fwrite conflicts with std.stdio.fwrite

--anders

Anders F Björklund wrote:
<snip>
> Nice, here is a patch to make it compile with GDC too...
> 
> Fixes:
>     expression expected, not 'TOK36' (#36 is dollar, by the way)

Looks like GDC has some catching up to do....

>     function std.c.stdio.fwrite conflicts with std.stdio.fwrite
<snip>

Why the two functions, I wonder?

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on the 'group where everyone may benefit.

Stewart Gordon wrote:

>> Nice, here is a patch to make it compile with GDC too...
>>
>> Fixes:
>>     expression expected, not 'TOK36' (#36 is dollar, by the way)
> 
> Looks like GDC has some catching up to do....

Indeed, but if you glance over at the D.gnu newsgroup,
you will see that David Friedman is doing a great job!

Besides, $ is an experimental feature of DMD anyway...
Simply using array.length, is less likely to go away ?

>>     function std.c.stdio.fwrite conflicts with std.stdio.fwrite
> 
> <snip>
> 
> Why the two functions, I wonder?

Phobos tends to run into a few conflicts with the standard headers.

In this case, it's because std.stdio imports std.c.stdio always ?

--anders

Anders F Björklund wrote:
> Stewart Gordon wrote:
> 
>>> Nice, here is a patch to make it compile with GDC too...
>>>
>>> Fixes:
>>>     expression expected, not 'TOK36' (#36 is dollar, by the way)
>>
>> Looks like GDC has some catching up to do....
> 
> Indeed, but if you glance over at the D.gnu newsgroup,
> you will see that David Friedman is doing a great job!
> 
> Besides, $ is an experimental feature of DMD anyway...
> Simply using array.length, is less likely to go away ?

Guess you're right....

>>>     function std.c.stdio.fwrite conflicts with std.stdio.fwrite
>>
>> <snip>
>>
>> Why the two functions, I wonder?
> 
> Phobos tends to run into a few conflicts with the standard headers.
> 
> In this case, it's because std.stdio imports std.c.stdio always ?

That shouldn't have anything to do with it.  If std.stdio imports std.c.stdio, which declares fwrite, then it's still std.c.stdio.fwrite, not std.stdio.fwrite, isn't it?

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on the 'group where everyone may benefit.

Stewart Gordon wrote:

>> Phobos tends to run into a few conflicts with the standard headers.
>> In this case, it's because std.stdio imports std.c.stdio always ?
> 
> That shouldn't have anything to do with it.  If std.stdio imports std.c.stdio, which declares fwrite, then it's still std.c.stdio.fwrite, not std.stdio.fwrite, isn't it?

To be honest, I am not sure. I just know that those two are
always getting mixed up, in GDC. Maybe I should ask David...

Hey, wait a second. Your dfilter program did import both:
import std.file, std.ctype, std.string, std.c.stdio, std.stdio;

Wonder why it works on DMD ?

--anders

Anders F Björklund wrote:
> Stewart Gordon wrote:
<snip>
>> That shouldn't have anything to do with it.  If std.stdio imports std.c.stdio, which declares fwrite, then it's still std.c.stdio.fwrite, not std.stdio.fwrite, isn't it?
> 
> To be honest, I am not sure. I just know that those two are
> always getting mixed up, in GDC. Maybe I should ask David...
> 
> Hey, wait a second. Your dfilter program did import both:
> import std.file, std.ctype, std.string, std.c.stdio, std.stdio;

As I said, that shouldn't have anything to do with it.

> Wonder why it works on DMD ?

Because the bug is in the GDC-specific parts, I guess.

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on the 'group where everyone may benefit.

Anders F Björklund wrote:
<snip>
> Nice, here is a patch to make it compile with GDC too...
> 
> Fixes:
>     expression expected, not 'TOK36' (#36 is dollar, by the way)
>     function std.c.stdio.fwrite conflicts with std.stdio.fwrite
<snip>

I've put this in and also fixed a few more bugs:

- removal of /+...+/ comments (had forgotten to update the end pointer)
- tokenising '\'' (I should've tried running dfilter on its own code
sooner!)

Attached is a combined patch.

I've also put the whole program here....

http://smjg.port5.com/pr/d/

(Maybe the version on Dsource could be updated...?)

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on the 'group where everyone may benefit.