www.digitalmars.com         C & C++   DMDScript  

D.gnu - 2.066 merge

reply "Iain Buclaw" <ibuclaw gdcproject.org> writes:
Hi,

Looks like its that time of the year again.

Same as last time, will be raising a pull and struggling through 
the last 6 months of dmd development.


Daniel, do you have a list of gluey-breaking changes/PRs at hand? 
  Using mail filters are failing me at this point in time.

Regards
Iain
Jul 14 2014
parent reply "Daniel Murphy" <yebbliesnospam gmail.com> writes:
"Iain Buclaw"  wrote in message news:phejgytsbkojsczbpvrk forum.dlang.org...


 Hi,

 Looks like its that time of the year again.

 Same as last time, will be raising a pull and struggling through the last 
 6 months of dmd development.


Apologies in advance.


 Daniel, do you have a list of gluey-breaking changes/PRs at hand? Using 
 mail filters are failing me at this point in time.


Not at hand, no.  grepping for ": public Visitor" over the glue files will 
tell you which ones I've done fairly easily.
Jul 14 2014
next sibling parent reply "Iain Buclaw via D.gnu" <d.gnu puremagic.com> writes:
On 14 July 2014 16:05, Daniel Murphy via D.gnu <d.gnu puremagic.com> wrote:

 "Iain Buclaw"  wrote in message news:phejgytsbkojsczbpvrk forum.dlang.org...



 Hi,

 Looks like its that time of the year again.

 Same as last time, will be raising a pull and struggling through the last
 6 months of dmd development.


 Apologies in advance.


There's a lot of good stuff in it.  So I'm happy about this, at the
same time I'm going to kill you next time we meet.  (Ha ha ha)




 Daniel, do you have a list of gluey-breaking changes/PRs at hand? Using
 mail filters are failing me at this point in time.


 Not at hand, no.  grepping for ": public Visitor" over the glue files will
 tell you which ones I've done fairly easily.


It's not that which I'm worried about, it's tracking down gluey
changes before/after the conversion that will take the longest.
Jul 14 2014
parent "Daniel Murphy" <yebbliesnospam gmail.com> writes:
"Iain Buclaw via D.gnu"  wrote in message 
news:mailman.4028.1405352460.2907.d.gnu puremagic.com...


 It's not that which I'm worried about, it's tracking down gluey
 changes before/after the conversion that will take the longest.


Maybe it would be worth getting a list of all pull requests in the release 
and work off that?  It should be fairly easy to exclude most as 
frontend-only by looking at the descriptions.
Jul 14 2014
prev sibling next sibling parent "Manu via D.gnu" <d.gnu puremagic.com> writes:
On 15 July 2014 01:40, Iain Buclaw via D.gnu <d.gnu puremagic.com> wrote:


 On 14 July 2014 16:05, Daniel Murphy via D.gnu <d.gnu puremagic.com>
 wrote:

 "Iain Buclaw"  wrote in message


 news:phejgytsbkojsczbpvrk forum.dlang.org...

 Hi,

 Looks like its that time of the year again.

 Same as last time, will be raising a pull and struggling through the




 last

 6 months of dmd development.


 Apologies in advance.


 There's a lot of good stuff in it.  So I'm happy about this, at the
 same time I'm going to kill you next time we meet.  (Ha ha ha)




 Daniel, do you have a list of gluey-breaking changes/PRs at hand? Using
 mail filters are failing me at this point in time.


 Not at hand, no.  grepping for ": public Visitor" over the glue files


 will

 tell you which ones I've done fairly easily.


 It's not that which I'm worried about, it's tracking down gluey
 changes before/after the conversion that will take the longest.


Out of curiosity/ignorance, why do you find it easier to do this in one
huge irregular block, than to merge trunk commits progressively?
Jul 14 2014
prev sibling parent "Iain Buclaw via D.gnu" <d.gnu puremagic.com> writes:
On 15 July 2014 03:18, Manu via D.gnu <d.gnu puremagic.com> wrote:

 On 15 July 2014 01:40, Iain Buclaw via D.gnu <d.gnu puremagic.com> wrote:

 On 14 July 2014 16:05, Daniel Murphy via D.gnu <d.gnu puremagic.com>
 wrote:

 "Iain Buclaw"  wrote in message
 news:phejgytsbkojsczbpvrk forum.dlang.org...



 Hi,

 Looks like its that time of the year again.

 Same as last time, will be raising a pull and struggling through the
 last
 6 months of dmd development.


 Apologies in advance.


 There's a lot of good stuff in it.  So I'm happy about this, at the
 same time I'm going to kill you next time we meet.  (Ha ha ha)




 Daniel, do you have a list of gluey-breaking changes/PRs at hand? Using
 mail filters are failing me at this point in time.


 Not at hand, no.  grepping for ": public Visitor" over the glue files
 will
 tell you which ones I've done fairly easily.


 It's not that which I'm worried about, it's tracking down gluey
 changes before/after the conversion that will take the longest.



 Out of curiosity/ignorance, why do you find it easier to do this in one huge
 irregular block, than to merge trunk commits progressively?



There was a time when it was easier.  Back when I was playing catch-up
with DMD, there were times that I went through 3-4 release merges in
just one week.

I made a graph just after DConf on this
http://dgnu.org/data/gdc-dmd-v2release.svg

The rapid rate of updating the frontend is partially contributed to
the fact that only Walter was committing changes, but I feel that its
mostly because new releases were occurring in a bi-weekly, monthly, or
bi-monthly basis, depending on seriousness of regressions,
implementation of a new feature, an ABI change, etc.

Fact in point.  Releases used to be focused around one major change
and lots of tiny bug fixes.

Now that the release process has grown to 6 months (along with it's
ever growing list of contributors).  It has lost all previous meaning
for me.  DMD releases are now just a sorry mess of changes that have
gone down the ether.  It's not just me who suffers, I think users
suffer just as much with understanding compatibility problems moving
from one to the other.

I have also been telling them since 2.060 (2 years now) to get their
act together.

Iain
Jul 15 2014