Use alternative header for nightly builds

[wilzbach]

I didn't realize that when we provide different Makefile targets (with BUILD_NIGHTLY) we can override even more logic :)

[CyberShadow]

How weird that this is causing a diff in the autotester.

It seems DDox has a race condition or something regarding which homonym symbol it chooses to link to.

[wilzbach]

How weird that this is causing a diff in the autotester.

Btw as this could also lead to unrelated diffs, should we disable the Bugzilla query for bugs for the changelog builder in "nightly mode"?

[CyberShadow]

Hmm, yeah, changes that shouldn't result in documentation changes should not have documentation changes indicated in the status line, because then inadvertent documentation changes will get through. I guess with this and dlang/dlang.org#1537 we should expand NODATETIME to something more generic.

[CyberShadow]

If you want to update the posix.mak side I'll update DAutoTest. I'm thinking a DIFFABLE makefile parameter which will automatically enable NODATETIME and the other stuff.

[wilzbach]

If you want to update the posix.mak side I'll update DAutoTest. I'm thinking a DIFFABLE makefile parameter which will automatically enable NODATETIME and the other stuff.

Alrighty :)

[wilzbach]

Alrighty :)

-> dlang/dlang.org#1588

[CyberShadow]

I understand this change will never be visible in the documentation tester because it applies to the Bugzilla header.

[wilzbach]

I understand this change will never be visible in the documentation tester because it applies to the Bugzilla header.

Well in theory it will be visible, e.g. here:
http://dlang.org/changelog/2.074.0.html

But of course only if the --nightly mode is enabled:

dlang/dlang.org#1586

[CyberShadow]

Right, I meant the documentation tester specifically (because we excluded Bugzilla entries when DIFFABLE is on).