December 01, 2014 Re: Phobos - breaking existing code | ||||
---|---|---|---|---|
| ||||
Posted in reply to Jacob Carlborg | On 12/1/2014 5:36 AM, Jacob Carlborg wrote:
> Every single release since at least DMD 2.050 has broken DWT. Most of these
> issues were language changes and not renamed symbols in Phobos.
My experience with upgrading old D code, even D1 code, is that language changes were easier to deal with than Phobos changes. The reason is because we've done a fairly good job with compiler error messages that not only point out errors, but suggest corrective action.
|
December 02, 2014 Re: Phobos - breaking existing code | ||||
---|---|---|---|---|
| ||||
Posted in reply to bearophile | On 11/29/14 8:09 AM, bearophile wrote: > H. S. Teoh: > >> Huh??! I *really* don't know what you're talking about, because in the >> last few months, I've seen a dramatic *increase* in Phobos activity. > > OK, if I was mistaken then I am happy ::) Sometimes it's good to be wrong. Might be because my participation has been scarce lately (busy with work and family). I'll also be out at the YOW conference (http://yowconference.com.au), talking about D. One interesting tidbit is that downloads have been rising sharply through mid-November, to have a bit of a pullback recently: http://erdani.com/d/downloads.daily.png I wonder if there's some way to filter out robots, build machines etc. Ideas? Andrei |
December 02, 2014 Re: Phobos - breaking existing code | ||||
---|---|---|---|---|
| ||||
Posted in reply to Andrei Alexandrescu | On Tuesday, 2 December 2014 at 00:13:03 UTC, Andrei Alexandrescu
wrote:
>
> I wonder if there's some way to filter out robots, build machines etc. Ideas?
The User-Agent header would be a good place to start.
|
December 02, 2014 Re: Phobos - breaking existing code | ||||
---|---|---|---|---|
| ||||
Posted in reply to Andrei Alexandrescu | On Tuesday, 2 December 2014 at 00:13:03 UTC, Andrei Alexandrescu wrote:
> I wonder if there's some way to filter out robots, build machines etc. Ideas?
We should probably do something to identify all the downloads coming from Travis-CI and other CI solutions.
|
December 02, 2014 Re: Phobos - breaking existing code | ||||
---|---|---|---|---|
| ||||
Posted in reply to Andrei Alexandrescu | On Tuesday, 2 December 2014 at 00:13:03 UTC, Andrei Alexandrescu wrote:
> One interesting tidbit is that downloads have been rising sharply through mid-November, to have a bit of a pullback recently: http://erdani.com/d/downloads.daily.png
One note about this: I'm using D in my projects since last year, but I'm still using DMD version 2.060!
So what I think that would be nice is to find a way to measure the user base beyond the downloads.
But anyway nice to see the number of downloads are rising, I hope that means new users.
PS: Sorry my english!
Matheus.
|
December 02, 2014 Re: Phobos - breaking existing code | ||||
---|---|---|---|---|
| ||||
Posted in reply to Andrei Alexandrescu | On Tuesday, 2 December 2014 at 00:13:03 UTC, Andrei Alexandrescu wrote:
> On 11/29/14 8:09 AM, bearophile wrote:
>> H. S. Teoh:
>>
>>> Huh??! I *really* don't know what you're talking about, because in the
>>> last few months, I've seen a dramatic *increase* in Phobos activity.
>>
>> OK, if I was mistaken then I am happy ::) Sometimes it's good to be wrong.
>
> Might be because my participation has been scarce lately (busy with work and family). I'll also be out at the YOW conference (http://yowconference.com.au), talking about D.
>
> One interesting tidbit is that downloads have been rising sharply through mid-November, to have a bit of a pullback recently: http://erdani.com/d/downloads.daily.png
>
> I wonder if there's some way to filter out robots, build machines etc. Ideas?
>
>
> Andrei
Just a thought but, its pretty close to end of year. Projects will be slowing down/stopping ext. People going on holiday. That also could screw the results a lot.
|
December 02, 2014 Re: Phobos - breaking existing code | ||||
---|---|---|---|---|
| ||||
Posted in reply to Rikki Cattermole | On 12/1/2014 6:19 PM, Rikki Cattermole wrote:
> Just a thought but, its pretty close to end of year. Projects will be slowing
> down/stopping ext. People going on holiday. That also could screw the results a
> lot.
In many organizations, nothing happens from Thanksgiving to the New Years.
|
December 02, 2014 Re: Phobos - breaking existing code | ||||
---|---|---|---|---|
| ||||
Posted in reply to Walter Bright | On 2/12/2014 3:26 p.m., Walter Bright wrote:
> On 12/1/2014 6:19 PM, Rikki Cattermole wrote:
>> Just a thought but, its pretty close to end of year. Projects will be
>> slowing
>> down/stopping ext. People going on holiday. That also could screw the
>> results a
>> lot.
>
> In many organizations, nothing happens from Thanksgiving to the New Years.
I guess things are a little different here in New Zealand, but then again different hemisphere.
E.g. we legally get Christmas day as a public holiday and schools already ended for the year.
|
December 02, 2014 Re: Phobos - breaking existing code | ||||
---|---|---|---|---|
| ||||
Posted in reply to Rikki Cattermole Attachments:
| On 12/01/2014 09:31 PM, Rikki Cattermole wrote: > On 2/12/2014 3:26 p.m., Walter Bright wrote: >> On 12/1/2014 6:19 PM, Rikki Cattermole wrote: >>> Just a thought but, its pretty close to end of year. Projects will be >>> slowing >>> down/stopping ext. People going on holiday. That also could screw the >>> results a >>> lot. >> >> In many organizations, nothing happens from Thanksgiving to the New Years. > > I guess things are a little different here in New Zealand, but then again different hemisphere. E.g. we legally get Christmas day as a public holiday and schools already ended for the year. People might "have work", doesn't mean that they get anything done...unless you're my university and decide to give a week of classes and then finals, as the only thing between Thanksgiving and Christmas. -- Matt Soucy http://msoucy.me/ |
December 02, 2014 Re: Phobos - breaking existing code | ||||
---|---|---|---|---|
| ||||
Posted in reply to Walter Bright | On 2014-11-29 00:33, Walter Bright wrote: > I fully understand how unfriendly this is to users and how discouraging > it can be to have their recently working code shattered and scattered. > We need to do a lot better. When we're talking about breaking code, I just found this regression: https://issues.dlang.org/show_bug.cgi?id=13811 -- /Jacob Carlborg |
Copyright © 1999-2021 by the D Language Foundation