Syntactically, the compiler allows "deprecated alias foo bar;"  but the deprecated attribute is apparently ignored?

Is this intentional?

kris escribió:
kris escribió:
> Syntactically, the compiler allows "deprecated alias foo bar;"  but the deprecated attribute is apparently ignored?
> 
> Is this intentional?

> Syntactically, the compiler allows "deprecated alias foo bar;"  but the deprecated attribute is apparently ignored?
> 
> Is this intentional?

As far as I know, *any* declaration (imports, typedefs, static asserts, debug, version, etc.) allow *any* modifier. I don't know if it's intentional, but at least that's the way the DMD parser works. I guess this makes the parser faster than worrying about this things, but that's just a guess.

Ary Manzana wrote:
> kris escribió:
> kris escribió:
> 
>> Syntactically, the compiler allows "deprecated alias foo bar;"  but the deprecated attribute is apparently ignored?
>>
>> Is this intentional?
> 
> 
>> Syntactically, the compiler allows "deprecated alias foo bar;"  but the deprecated attribute is apparently ignored?
>>
>> Is this intentional?
> 
> 
> As far as I know, *any* declaration (imports, typedefs, static asserts, debug, version, etc.) allow *any* modifier. I don't know if it's intentional, but at least that's the way the DMD parser works. I guess this makes the parser faster than worrying about this things, but that's just a guess.


Hehe ... just want to know if deprecated is intended to work with alias or not? And if so, can we expect a fix for it soon?

:)

On Fri, 16 Feb 2007 16:37:40 -0500, kris <foo@bar.com> wrote:

> Syntactically, the compiler allows "deprecated alias foo bar;"  but the deprecated attribute is apparently ignored?
>
> Is this intentional?

I did notice deprecated isn't working in all cases, and it seemed to me that it's only working when accessing deprecated things in other code. It might just be a bug and isn't quite happening this way, but it's what I got from it, and I realized it would be a decent way of handling it.
If my code accesses a deprecated thing from within the same file, it's quite OK; like private. It's probably just happening to support the deprecated feature. e.g. a deprecated function calling another deprecated function in the same file shouldn't require -d; it might be just a stub with different parameters calling the full implementation, or calling a helper function that is also deprecated, in order to continue supporting a deprecated feature.

Chris Miller wrote:

> On Fri, 16 Feb 2007 16:37:40 -0500, kris <foo@bar.com> wrote:
> 
>> Syntactically, the compiler allows "deprecated alias foo bar;"  but the deprecated attribute is apparently ignored?
>>
>> Is this intentional?
> 
> I did notice deprecated isn't working in all cases, and it seemed to me
> that it's only working when accessing deprecated things in other code. It
> might just be a bug and isn't quite happening this way, but it's what I
> got from it, and I realized it would be a decent way of handling it.
> If my code accesses a deprecated thing from within the same file, it's
> quite OK; like private. It's probably just happening to support the
> deprecated feature. e.g. a deprecated function calling another deprecated
> function in the same file shouldn't require -d; it might be just a stub
> with different parameters calling the full implementation, or calling a
> helper function that is also deprecated, in order to continue supporting a
> deprecated feature.

I tried to test this, and the compiler did not pick up the deprecated alias if used in the client code. An additional factor that may have played a role in this particular test was that it was a .di file.

-- 
Lars Ivar Igesund
blog at http://larsivi.net
DSource, #d.tango & #D: larsivi
Dancing the Tango

Lars Ivar Igesund wrote:

> Chris Miller wrote:
> 
>> On Fri, 16 Feb 2007 16:37:40 -0500, kris <foo@bar.com> wrote:
>> 
>>> Syntactically, the compiler allows "deprecated alias foo bar;"  but the deprecated attribute is apparently ignored?
>>>
>>> Is this intentional?
>> 
>> I did notice deprecated isn't working in all cases, and it seemed to me
>> that it's only working when accessing deprecated things in other code. It
>> might just be a bug and isn't quite happening this way, but it's what I
>> got from it, and I realized it would be a decent way of handling it.
>> If my code accesses a deprecated thing from within the same file, it's
>> quite OK; like private. It's probably just happening to support the
>> deprecated feature. e.g. a deprecated function calling another deprecated
>> function in the same file shouldn't require -d; it might be just a stub
>> with different parameters calling the full implementation, or calling a
>> helper function that is also deprecated, in order to continue supporting
>> a deprecated feature.
> 
> I tried to test this, and the compiler did not pick up the deprecated alias if used in the client code. An additional factor that may have played a role in this particular test was that it was a .di file.
> 

Actually it did, but not the way I expected. If you use the declared symbol, not the alias, then you will get a deprecation message. Starts to sound like a bug.

-- 
Lars Ivar Igesund
blog at http://larsivi.net
DSource, #d.tango & #D: larsivi
Dancing the Tango