Jump to page: 1 25  
Page
Thread overview
moving away from changelog.dd?
Dec 24, 2012
Jonathan M Davis
Dec 24, 2012
Bernard Helyer
Dec 24, 2012
SomeDude
Dec 24, 2012
Jonathan M Davis
Dec 25, 2012
Walter Bright
Dec 25, 2012
Jonathan M Davis
Dec 25, 2012
Walter Bright
Dec 25, 2012
Jonathan M Davis
Dec 25, 2012
Walter Bright
Dec 25, 2012
Jonathan M Davis
Dec 25, 2012
Walter Bright
Dec 26, 2012
foobar
Dec 25, 2012
Leandro Lucarella
Dec 25, 2012
Walter Bright
Dec 26, 2012
Leandro Lucarella
Dec 26, 2012
Walter Bright
Dec 25, 2012
SomeDude
Dec 24, 2012
Andrej Mitrovic
Dec 24, 2012
Jonathan M Davis
Dec 24, 2012
Maxim Fomin
Dec 24, 2012
Jonathan M Davis
Dec 24, 2012
Jonathan M Davis
Dec 24, 2012
Kapps
Dec 24, 2012
monarch_dodra
Dec 24, 2012
Jonathan M Davis
Dec 24, 2012
Jacob Carlborg
Dec 25, 2012
Walter Bright
Dec 25, 2012
Jonathan M Davis
Dec 25, 2012
Walter Bright
Dec 25, 2012
Jonathan M Davis
Dec 24, 2012
Jonathan M Davis
Dec 24, 2012
Dmitry Olshansky
December 24, 2012
Hello,


We've discussed this a bit before. It's becoming obvious that changelog.dd doesn't scale well: it's difficult to keep in sync, people forget to add bugs to it etc.

How about we abolish changelog.dd in favor of bugzilla queries? There's a simple rule to follow: all changes should have a corresponding entry in bugzilla. This is easy to enforce.


Destroy,

Andrei
December 24, 2012
On Sunday, December 23, 2012 20:55:03 Andrei Alexandrescu wrote:
> Hello,
> 
> 
> We've discussed this a bit before. It's becoming obvious that changelog.dd doesn't scale well: it's difficult to keep in sync, people forget to add bugs to it etc.
> 
> How about we abolish changelog.dd in favor of bugzilla queries? There's a simple rule to follow: all changes should have a corresponding entry in bugzilla. This is easy to enforce.
> 
> 
> Destroy,

It's not enough. We need to be able to add items to the changelog separately from bugzilla. There are pretty much always extra items in the changelog which are not bug reports or enhancement requests, and many of them make no sense as items in bugzilla - especially when the whole point of putting them there is to explain something rather than listing a bug number.

So, I fully support doing something to automate the portion of the changelog which contains all of the bugzilla entries, but the portion above that where non-bugzilla stuff is entered still needs to be there. We can handle it in some way other than changelog.dd if we want to, but we still need a way to enter stuff into the changelog by hand.

- Jonathan M Davis
December 24, 2012
On 12/23/12 10:35 PM, Jonathan M Davis wrote:
> So, I fully support doing something to automate the portion of the changelog
> which contains all of the bugzilla entries, but the portion above that where
> non-bugzilla stuff is entered still needs to be there. We can handle it in some
> way other than changelog.dd if we want to, but we still need a way to enter
> stuff into the changelog by hand.

Some other way may as well be introduce entries in bugzilla. Again, dealing with the changelog has become more costly than beneficial.

Andrei


December 24, 2012
On Monday, 24 December 2012 at 03:48:47 UTC, Andrei Alexandrescu
wrote:
> Some other way may as well be introduce entries in bugzilla.

Is adding things that don't belong in bugzilla to bugzilla really
a great idea?
December 24, 2012
On Monday, 24 December 2012 at 01:55:04 UTC, Andrei Alexandrescu wrote:
> Hello,
>
>
> We've discussed this a bit before. It's becoming obvious that changelog.dd doesn't scale well: it's difficult to keep in sync, people forget to add bugs to it etc.
>
> How about we abolish changelog.dd in favor of bugzilla queries? There's a simple rule to follow: all changes should have a corresponding entry in bugzilla. This is easy to enforce.
>
>
> Destroy,
>
> Andrei


Why not enforce github pulls to provide updates to changelog?
December 24, 2012
On Sunday, December 23, 2012 22:48:47 Andrei Alexandrescu wrote:
> On 12/23/12 10:35 PM, Jonathan M Davis wrote:
> > So, I fully support doing something to automate the portion of the changelog which contains all of the bugzilla entries, but the portion above that where non-bugzilla stuff is entered still needs to be there. We can handle it in some way other than changelog.dd if we want to, but we still need a way to enter stuff into the changelog by hand.
> 
> Some other way may as well be introduce entries in bugzilla. Again, dealing with the changelog has become more costly than beneficial.

It's costly because of bugzilla. If you automate that part, then I think we're fine. But i don't think that it makes any sense whatsoever to put the non- bugzilla stuff in bugzilla. For instance, take this line from the top of the 2.060 changelog:

$(LI std.string: $(RED The current implementations of std.string.format and string.sformat are scheduled to be replaced in November 2012 with improved implementations which conform to writef. In some, rare cases, this will break code. Please see the documentation for std.string.format and std.string.sformat for details.))

How would you do that with a bugzilla entry? Try and create a bug with that title? Even assuming that bugzilla allows you to create a bug with a title that long, it'll then be buried among all of the bug fixes and have a pointless bug# on it, when it specifically needed to be put at the top.

I really think that the WHATSNEW section needs to stay as it is, and I don't think that it's really all that big a deal to maintain it. It would be to automate the LIBBUGSFIXED section, but I don't think that it should be mixed in with the WHATSNEW section, and the WHATSNEW section is very ill-suited to being put in bugzilla.

- Jonathan m Davis
December 24, 2012
On Monday, December 24, 2012 05:04:24 Maxim Fomin wrote:
> Why not enforce github pulls to provide updates to changelog?

Because then you get merge conflicts galore.

It makes a _lot_ of sense to automate the bugs section of the changelog. Worst case, it's not all that hard to write a program to grab the data and format it appropriately for changelog.dd, but it should be quite possible to put that query in the changelog html page itself. I think that it would be a mistake to try and enforce more process with regards to changelog.dd in order to get the bug entries in there, especially when it's not all that hard to automate.

It's just the WHATSNEW section that makes no sense to automate.

- Jonathan M Davis
December 24, 2012
On 12/23/12 11:08 PM, Jonathan M Davis wrote:
> It's just the WHATSNEW section that makes no sense to automate.

Enhancement requests may fill most of the bill...

Andrei

December 24, 2012
On Sunday, December 23, 2012 23:13:53 Andrei Alexandrescu wrote:
> On 12/23/12 11:08 PM, Jonathan M Davis wrote:
> > It's just the WHATSNEW section that makes no sense to automate.
> 
> Enhancement requests may fill most of the bill...

They definitely don't fill all of it though. We need to be able to put specific messages in the changelog separate from bugzilla entries.

I really don't see why it's that big a deal that the WHATSNEW section isn't automated though. It's the LIBBUGSFIXED section that causes all the grief, and that's easily automated.

- Jonathan M Davis
December 24, 2012
On Monday, 24 December 2012 at 04:19:44 UTC, Jonathan M Davis wrote:
> On Sunday, December 23, 2012 23:13:53 Andrei Alexandrescu wrote:
>> On 12/23/12 11:08 PM, Jonathan M Davis wrote:
>> > It's just the WHATSNEW section that makes no sense to automate.
>> 
>> Enhancement requests may fill most of the bill...
>
> They definitely don't fill all of it though. We need to be able to put specific
> messages in the changelog separate from bugzilla entries.
>
> I really don't see why it's that big a deal that the WHATSNEW section isn't
> automated though. It's the LIBBUGSFIXED section that causes all the grief, and
> that's easily automated.
>
> - Jonathan M Davis


Personally I think it nicer that things like What's New would be written out by a human. Instead of just "Enhancement #4713: Add user-defined annotations", you can give a quick explanation of what actually was added and a link to get more info about it. Same for phobos modules and the like.
« First   ‹ Prev
1 2 3 4 5