On Monday, 20 July 2015 at 11:45:31 UTC, Gary Willoughby wrote:
> On Monday, 20 July 2015 at 04:02:04 UTC, Walter Bright wrote:
>> Perhaps we should name this 2.100, to signify such a milestone.
>
> 2.1 Sounds good!

2.1.0 sounds even better. 2.100 - not.

> But going forward can we stick to a sane versioning system like what dub uses:
>
> http://semver.org/

+1

On Monday, 20 July 2015 at 13:16:46 UTC, sigod wrote:
> On Monday, 20 July 2015 at 11:45:31 UTC, Gary Willoughby wrote:
>> On Monday, 20 July 2015 at 04:02:04 UTC, Walter Bright wrote:
>>> Perhaps we should name this 2.100, to signify such a milestone.
>>
>> 2.1 Sounds good!
>
> 2.1.0 sounds even better. 2.100 - not.

That'd probably cause confusion, people wondering if there'd been a mistake and an old version was put in the release. 2.7.0 would probably be better. Otherwise, +1 for semver.

On Monday, 20 July 2015 at 13:52:16 UTC, Meta wrote:
> On Monday, 20 July 2015 at 13:16:46 UTC, sigod wrote:
>> On Monday, 20 July 2015 at 11:45:31 UTC, Gary Willoughby wrote:
>>> On Monday, 20 July 2015 at 04:02:04 UTC, Walter Bright wrote:
>>>> Perhaps we should name this 2.100, to signify such a milestone.
>>>
>>> 2.1 Sounds good!
>>
>> 2.1.0 sounds even better. 2.100 - not.
>
> That'd probably cause confusion, people wondering if there'd been a mistake and an old version was put in the release. 2.7.0 would probably be better. Otherwise, +1 for semver.

2.69.0 for DDMD release, then? Or does 3rd digit means PATCH version in current versioning system?

On Monday, 20 July 2015 at 13:58:02 UTC, sigod wrote:
> 2.69.0 for DDMD release, then? Or does 3rd digit means PATCH version in current versioning system?

I don't really know how the current versioning system works. I think it just increments by 1 every release, and patches get .1, etc. added on.

On Monday, 20 July 2015 at 14:07:12 UTC, Meta wrote:
> On Monday, 20 July 2015 at 13:58:02 UTC, sigod wrote:
>> 2.69.0 for DDMD release, then? Or does 3rd digit means PATCH version in current versioning system?
>
> I don't really know how the current versioning system works. I think it just increments by 1 every release, and patches get .1, etc. added on.

Yes, you're right. I forgot about current version - 2.067.1
It seems not so different from semver.

On Monday, 20 July 2015 at 14:31:38 UTC, sigod wrote:
> On Monday, 20 July 2015 at 14:07:12 UTC, Meta wrote:
>> On Monday, 20 July 2015 at 13:58:02 UTC, sigod wrote:
>>> 2.69.0 for DDMD release, then? Or does 3rd digit means PATCH version in current versioning system?
>>
>> I don't really know how the current versioning system works. I think it just increments by 1 every release, and patches get .1, etc. added on.
>
> Yes, you're right. I forgot about current version - 2.067.1
> It seems not so different from semver.

It's missing the most important piece of semver, the fact that the major version number denotes backwards incompatible changes. If D were using semver, we would be somewhere between D10 and D15.

On Monday, 20 July 2015 at 16:32:08 UTC, Jack Stouffer wrote:
> It's missing the most important piece of semver, the fact that the major version number denotes backwards incompatible changes. If D were using semver, we would be somewhere between D10 and D15.

D2 is the product name. We're on major version 67 of D2. :P

On Monday, 20 July 2015 at 11:45:31 UTC, Gary Willoughby wrote:
> 2.1 Sounds good!

That version was already released in 2007 [1]. Version numbers are not floats, they are period separated lists of integers. So `assert(2.1 == 2.01 && 2.01 == 2.001);`. The leading zero is actually forbidden in several important places (such as Linux package version numbers), and is one of the more head-scratching stylistic choices in D. Bumping to 2.100 will appease those who insist on a three-digit second component for reasons I will never understand, and make it so that D versions can actually be uniformly represented everywhere they show up.

[1]: http://dlang.org/changelog.html#new2_001

On 20 July 2015 at 19:10, Whatever via Digitalmars-d <digitalmars-d@puremagic.com> wrote:
> On Monday, 20 July 2015 at 11:45:31 UTC, Gary Willoughby wrote:
>>
>> 2.1 Sounds good!
>
>
> That version was already released in 2007 [1]. Version numbers are not floats, they are period separated lists of integers. So `assert(2.1 == 2.01 && 2.01 == 2.001);`.

You might be able to get away with 2.001 == 2.0.01 and 2.1 == 2.1.00. :o)

On a serious note though, the versioning could be better, and I think the current suggested version bump from Walter is a missed opportunity to set things right.  To quote anther project leader though: "Let’s face it - what’s the point of being in charge if you can’t pick the bike shed color without holding a referendum on it?"

Iain

On 2015-07-20 18:32, Jack Stouffer wrote:

> It's missing the most important piece of semver, the fact that the major
> version number denotes backwards incompatible changes. If D were using
> semver, we would be somewhere between D10 and D15.

Since we're about to release 2.068.0, I would say D68 ;). I'm actually quite serious. Every single release since at least 2.050 has broken at least one of my projects.

-- 
/Jacob Carlborg