www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - State of std.json?

reply Jonathan M Davis <jmdavisProg gmx.com> writes:
As I understand it, std.json is due for an overhaul. Also, as I recall, Robert 
Jacques was working on doing that. What's the current state of that effort? 
Are we looking at being able to review it anytime soon? Or does more have to 
be done before that can happen?

I really think that the remaining major module overhauls in Phobos need to 
have a high priority so that those portions of Phobos can be appropriately 
stabilized. The longer that we have code sitting around that we're planning to 
replace, the more code is going to have to be changed when we do replace them. 
std.json is on the list of modules that supposedly needs an overhaul (I've 
never used it, so I really don't know anything about its current state), so 
I'm asking about it.

- Jonathan M Davis
Jun 25 2011
next sibling parent reply "Robert Jacques" <sandford jhu.edu> writes:
On Sun, 26 Jun 2011 00:25:44 -0400, Jonathan M Davis <jmdavisProg gmx.com>  
wrote:
 As I understand it, std.json is due for an overhaul. Also, as I recall,  
 Robert
 Jacques was working on doing that. What's the current state of that  
 effort?
 Are we looking at being able to review it anytime soon? Or does more  
 have to
 be done before that can happen?

 I really think that the remaining major module overhauls in Phobos need  
 to
 have a high priority so that those portions of Phobos can be  
 appropriately
 stabilized. The longer that we have code sitting around that we're  
 planning to
 replace, the more code is going to have to be changed when we do replace  
 them.
 std.json is on the list of modules that supposedly needs an overhaul  
 (I've
 never used it, so I really don't know anything about its current state),  
 so
 I'm asking about it.

 - Jonathan M Davis

In and of itself, I feel my std.json patch is ready for review and I've posted the docs/code to the phobos list for preliminary review. The major issue holding a formal review up is that it is now dependent on a patch to std.variant which has suffered scope creep and tertiary bugs. As the review queue has had higher priority items in it and there has been little community interest in either module, I've put the final polishing of std.variant on the back burner. But as the queue is temporarily empty, I'll look over my code tomorrow and see if I can't get variant ready for review.
Jun 25 2011
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 6/26/11 12:54 AM, Jonathan M Davis wrote:
 Well, AFAIK the only other thing which could be up for review at the moment is
 std.benchmark, which arguably doesn't even need a review, which is quite small
 (something like half of it coming from std.datetime) and is definitely less
 important than getting a pre-existing module revamped such as std.json. So, if
 you're nearly ready to have the variant stuff reviewed, then I think that we
 should go for it. If not, then we can look at std.benchmark first.

Since std.benchmark got promoted from small additions to full-fledged module, it needs a few improvements before the formal review. The most important is output in machine readable form such that you can see "before" and "after" comparisons. I'm also thinking of a complementary tools/ script that fetches two given releases of Phobos (or the latest and the head), compares them, and outputs the deltas. Unfortunately, I'm very busy for the time being and I don't see getting to this until at least two weeks from now. Thanks, Andrei
Jun 26 2011
prev sibling next sibling parent Jonathan M Davis <jmdavisProg gmx.com> writes:
On 2011-06-25 21:56, Robert Jacques wrote:
 On Sun, 26 Jun 2011 00:25:44 -0400, Jonathan M Davis <jmdavisProg gmx.com>
 
 wrote:
 As I understand it, std.json is due for an overhaul. Also, as I recall,
 Robert
 Jacques was working on doing that. What's the current state of that
 effort?
 Are we looking at being able to review it anytime soon? Or does more
 have to
 be done before that can happen?
 
 I really think that the remaining major module overhauls in Phobos need
 to
 have a high priority so that those portions of Phobos can be
 appropriately
 stabilized. The longer that we have code sitting around that we're
 planning to
 replace, the more code is going to have to be changed when we do replace
 them.
 std.json is on the list of modules that supposedly needs an overhaul
 (I've
 never used it, so I really don't know anything about its current state),
 so
 I'm asking about it.
 
 - Jonathan M Davis

In and of itself, I feel my std.json patch is ready for review and I've posted the docs/code to the phobos list for preliminary review. The major issue holding a formal review up is that it is now dependent on a patch to std.variant which has suffered scope creep and tertiary bugs. As the review queue has had higher priority items in it and there has been little community interest in either module, I've put the final polishing of std.variant on the back burner. But as the queue is temporarily empty, I'll look over my code tomorrow and see if I can't get variant ready for review.

Well, AFAIK the only other thing which could be up for review at the moment is std.benchmark, which arguably doesn't even need a review, which is quite small (something like half of it coming from std.datetime) and is definitely less important than getting a pre-existing module revamped such as std.json. So, if you're nearly ready to have the variant stuff reviewed, then I think that we should go for it. If not, then we can look at std.benchmark first. - Jonathan M Davis
Jun 25 2011
prev sibling parent Jonathan M Davis <jmdavisProg gmx.com> writes:
On 2011-06-26 07:17, Andrei Alexandrescu wrote:
 On 6/26/11 12:54 AM, Jonathan M Davis wrote:
 Well, AFAIK the only other thing which could be up for review at the
 moment is std.benchmark, which arguably doesn't even need a review,
 which is quite small (something like half of it coming from
 std.datetime) and is definitely less important than getting a
 pre-existing module revamped such as std.json. So, if you're nearly
 ready to have the variant stuff reviewed, then I think that we should go
 for it. If not, then we can look at std.benchmark first.

Since std.benchmark got promoted from small additions to full-fledged module, it needs a few improvements before the formal review. The most important is output in machine readable form such that you can see "before" and "after" comparisons. I'm also thinking of a complementary tools/ script that fetches two given releases of Phobos (or the latest and the head), compares them, and outputs the deltas. Unfortunately, I'm very busy for the time being and I don't see getting to this until at least two weeks from now.

Okay. Good to know. Life does get in the way sometimes. ;) - Jonathan M Davis
Jun 26 2011