Hi all,

I'm a little perplexed b/c I can't seem to find anything that could tell me where this ends:

version(something):
code
code
code
\eof

How do you stop statements from belonging to the specific version of code without using brackets?

Thanks!

On Friday, 28 February 2014 at 04:19:47 UTC, Etienne Cimon wrote:
> Hi all,
>
> I'm a little perplexed b/c I can't seem to find anything that could tell me where this ends:
>
> version(something):
> code
> code
> code
> \eof
>
> How do you stop statements from belonging to the specific version of code without using brackets?
>
> Thanks!

add "version(all):" after code where specific version ends.

On Friday, 28 February 2014 at 04:31:03 UTC, evilrat wrote:
> On Friday, 28 February 2014 at 04:19:47 UTC, Etienne Cimon wrote:
>> Hi all,
>>
>> I'm a little perplexed b/c I can't seem to find anything that could tell me where this ends:
>>
>> version(something):
>> code
>> code
>> code
>> \eof
>>
>> How do you stop statements from belonging to the specific version of code without using brackets?
>>
>> Thanks!
>
> add "version(all):" after code where specific version ends.

nope

On 2014-02-28 7:53 AM, anonymous wrote:
> nope

nope?

On 2014-03-01 4:14 PM, Etienne wrote:
> On 2014-02-28 7:53 AM, anonymous wrote:
>> nope
>
> nope?

Can someone refute this anonymous nope?

On Saturday, 1 March 2014 at 21:14:53 UTC, Etienne wrote:
> On 2014-02-28 7:53 AM, anonymous wrote:
>> nope
>
> nope?

yep ;)

The nope was directed at this statement specifically:

On Friday, 28 February 2014 at 04:31:03 UTC, evilrat wrote:
> add "version(all):" after code where specific version ends.

I.e. "version(all):" doesn't cancel a former "version(foo):".
There may or may not be a way to cancel a "version(foo):". I
can't think of anything. Also, I don't see the problem with
brackets.

On Saturday, 1 March 2014 at 21:42:56 UTC, anonymous wrote:
> On Saturday, 1 March 2014 at 21:14:53 UTC, Etienne wrote:
>> On 2014-02-28 7:53 AM, anonymous wrote:
>>> nope
>>
>> nope?
>
> yep ;)
>
> The nope was directed at this statement specifically:
>
> On Friday, 28 February 2014 at 04:31:03 UTC, evilrat wrote:
>> add "version(all):" after code where specific version ends.
>
> I.e. "version(all):" doesn't cancel a former "version(foo):".
> There may or may not be a way to cancel a "version(foo):". I
> can't think of anything. Also, I don't see the problem with
> brackets.

of course it doesn't cancel it. it just re-enables code after descending version operator. versions can only be enabled via compiler args or CTFE.

On Sunday, 2 March 2014 at 05:23:21 UTC, evilrat wrote:
> On Saturday, 1 March 2014 at 21:42:56 UTC, anonymous wrote:
>> On Saturday, 1 March 2014 at 21:14:53 UTC, Etienne wrote:
>>> On 2014-02-28 7:53 AM, anonymous wrote:
>>>> nope
>>>
>>> nope?
>>
>> yep ;)
>>
>> The nope was directed at this statement specifically:
>>
>> On Friday, 28 February 2014 at 04:31:03 UTC, evilrat wrote:
>>> add "version(all):" after code where specific version ends.
>>
>> I.e. "version(all):" doesn't cancel a former "version(foo):".
>> There may or may not be a way to cancel a "version(foo):". I
>> can't think of anything. Also, I don't see the problem with
>> brackets.
>
> of course it doesn't cancel it. it just re-enables code after descending version operator. versions can only be enabled via compiler args or CTFE.

oh. sorry, i remembere i have same problems with version:, back then i switched to brackets.

"evilrat"  wrote in message news:mxhmgkljrzqhaymeclvk@forum.dlang.org...
On Friday, 28 February 2014 at 04:19:47 UTC, Etienne Cimon wrote:
> >
> > How do you stop statements from belonging to the specific version of code without using brackets?
>
> add "version(all):" after code where specific version ends.

This is incorrect, there is no way to do this.

The compiler translates
version(...):
...

into
version(...)
{
...
}

So any version labels after the first are moved inside the first.

You can see it with this code:

version = y;
version(x):
pragma(msg, "versionx");
version(y):
pragma(msg, "versiony");

Nothing is printed, because the version(y) block is inside the version(x) block, and version(x) is not set.

The answer: use braces. 

On 2014-03-01 16:42, anonymous wrote:
>
> I.e. "version(all):" doesn't cancel a former "version(foo):".
> There may or may not be a way to cancel a "version(foo):". I
> can't think of anything. Also, I don't see the problem with
> brackets.

Brackets (braces) with "good practice" requires changing the indentation of everything in it :/

To keep it clean I decided to put the different sources in separate files and put the version clause near the import or with version (DEFINES): at the top of the page..