www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Ready for review of Formal Review

reply "Jesse Phillips" <Jesse.K.Phillips+D gmail.com> writes:
Hello All,

I have been working on the Review Process documentation and 
believe it should go through a review similar to the process for 
which it describes. This would mean a review manager and voting. 
After that it is up to the review manager to decide how to run 
the review.

http://wiki.dlang.org/Review/Process

All details are up for discussion, I have placed "Approval 
Wanted" in places likely to be controversial so those are good 
places to jump to if short on time.

I have taken the Boost process[1] and consolidated it to the 
points we currently use and probably should use, while leaving 
out items that at this time have not been used and would likely 
not lead to better results.

Would anyone be willing to be a review manager?

Should this be placed in the review queue? :)

1. http://www.boost.org/community/reviews.html
Jun 18 2013
next sibling parent Timothee Cour <thelastmammoth gmail.com> writes:
Just a concern regarding requirement for portability:
we should strive for portability whenever possible, but this shouldn't
hinder useful library code that works only on a subset of platforms (eg
support for other platforms could come later if at all).

On Tue, Jun 18, 2013 at 7:36 AM, Jesse Phillips <
Jesse.K.Phillips+D gmail.com> wrote:

 Hello All,

 I have been working on the Review Process documentation and believe it
 should go through a review similar to the process for which it describes.
 This would mean a review manager and voting. After that it is up to the
 review manager to decide how to run the review.

 http://wiki.dlang.org/Review/**Process<http://wiki.dlang.org/Review/Process>

 All details are up for discussion, I have placed "Approval Wanted" in
 places likely to be controversial so those are good places to jump to if
 short on time.

 I have taken the Boost process[1] and consolidated it to the points we
 currently use and probably should use, while leaving out items that at this
 time have not been used and would likely not lead to better results.

 Would anyone be willing to be a review manager?

 Should this be placed in the review queue? :)

 1. http://www.boost.org/**community/reviews.html<http://www.boost.org/community/reviews.html>

Jun 18 2013
prev sibling next sibling parent Brad Roberts <braddr puremagic.com> writes:
Any block of code going into phobos (or druntime, but it ought to be extremely
rare that we're 
adding any significant chunks of code there) really should be portable to all
supported platforms 
(the union of dmd, gdc, and ldc).

On 6/18/13 10:18 AM, Timothee Cour wrote:
 Just a concern regarding requirement for portability:
 we should strive for portability whenever possible, but this shouldn't hinder
useful library code
 that works only on a subset of platforms (eg support for other platforms could
come later if at all).

 On Tue, Jun 18, 2013 at 7:36 AM, Jesse Phillips <Jesse.K.Phillips+D gmail.com
 <mailto:Jesse.K.Phillips+D gmail.com>> wrote:

     Hello All,

     I have been working on the Review Process documentation and believe it
should go through a
     review similar to the process for which it describes. This would mean a
review manager and
     voting. After that it is up to the review manager to decide how to run the
review.

     http://wiki.dlang.org/Review/__Process
<http://wiki.dlang.org/Review/Process>

     All details are up for discussion, I have placed "Approval Wanted" in
places likely to be
     controversial so those are good places to jump to if short on time.

     I have taken the Boost process[1] and consolidated it to the points we
currently use and
     probably should use, while leaving out items that at this time have not
been used and would
     likely not lead to better results.

     Would anyone be willing to be a review manager?

     Should this be placed in the review queue? :)

     1. http://www.boost.org/__community/reviews.html
<http://www.boost.org/community/reviews.html>

Jun 18 2013
prev sibling next sibling parent reply Timothee Cour <thelastmammoth gmail.com> writes:
On Tue, Jun 18, 2013 at 10:29 AM, Brad Roberts <braddr puremagic.com> wrote:

 Any block of code going into phobos (or druntime, but it ought to be
 extremely rare that we're adding any significant chunks of code there)
 really should be portable to all supported platforms (the union of dmd,
 gdc, and ldc).

There are corner cases. D shared libraries worked only on windows for a while, then were added to linux (then osx?) D runtime loaded libraries will work on linux only first, IIRC. gdb demangles D symbols on linux but not osx (although not part of phobos) because an extra underscore is added in mangling. sometimes, it's not worth waiting until all OS support a feature to introduce it.
Jun 18 2013
parent "Jesse Phillips" <Jesse.K.Phillips+D gmail.com> writes:
On Tuesday, 18 June 2013 at 17:47:10 UTC, Timothee Cour wrote:
 There are corner cases.

 D shared libraries worked only on windows for a while, then 
 were added to
 linux (then osx?)
 D runtime loaded libraries will work on linux only first, IIRC.
 gdb demangles D symbols on linux but not osx (although not part 
 of phobos)
 because an extra underscore is added in mangling.

 sometimes, it's not worth waiting until all OS support a 
 feature to
 introduce it.

Your example is one I would classify as a bug (we aren't building a Go compiler). Note that I also start with "Requirements can be waved at community discretion." Which is to cover the extreme cases where the community feels a requirement does not apply.
Jun 18 2013
prev sibling parent Brad Roberts <braddr puremagic.com> writes:
On 6/18/13 10:46 AM, Timothee Cour wrote:
 On Tue, Jun 18, 2013 at 10:29 AM, Brad Roberts <braddr puremagic.com
<mailto:braddr puremagic.com>>
 wrote:

     Any block of code going into phobos (or druntime, but it ought to be
extremely rare that we're
     adding any significant chunks of code there) really should be portable to
all supported
     platforms (the union of dmd, gdc, and ldc).


 There are corner cases.

 D shared libraries worked only on windows for a while, then were added to
linux (then osx?)
 D runtime loaded libraries will work on linux only first, IIRC.
 gdb demangles D symbols on linux but not osx (although not part of phobos)
because an extra
 underscore is added in mangling.

 sometimes, it's not worth waiting until all OS support a feature to introduce
it.

And both of those examples are going to be cross platform. It's only a matter of when. I talking about intent. You're correct though that timing does matter.
Jun 18 2013