Jump to page: 1 26  
Page
Thread overview
Send me your list of D gripes and wishes
Dec 24, 2022
Mike Parker
Dec 24, 2022
ag0aep6g
Dec 24, 2022
Mike Parker
Dec 24, 2022
Walter Bright
Dec 24, 2022
Paul Backus
Dec 24, 2022
ag0aep6g
Dec 25, 2022
Jonathan Marler
Dec 25, 2022
Walter Bright
Dec 25, 2022
GrimMaple
Dec 25, 2022
GrimMaple
Dec 25, 2022
Walter Bright
Dec 25, 2022
bachmeier
Dec 25, 2022
Walter Bright
Dec 26, 2022
claptrap
Dec 26, 2022
Walter Bright
Dec 26, 2022
claptrap
Dec 25, 2022
Paul Backus
Dec 25, 2022
Walter Bright
Dec 26, 2022
claptrap
Dec 26, 2022
Paul Backus
Dec 26, 2022
Walter Bright
Dec 26, 2022
Paul Backus
Dec 26, 2022
Mike Parker
Jan 01, 2023
Paul Backus
Dec 31, 2022
Walter Bright
Jan 02, 2023
JN
Jan 05, 2023
John Colvin
Dec 26, 2022
Walter Bright
Dec 26, 2022
Paul Backus
Dec 26, 2022
monkyyy
Dec 25, 2022
Max Samukha
Dec 30, 2022
peterrmaxx15
Dec 31, 2022
Max Samukha
Dec 25, 2022
Walter Bright
Dec 25, 2022
Mike Parker
Dec 25, 2022
RTM
Dec 25, 2022
Hipreme
Dec 25, 2022
GrimMaple
Jan 02, 2023
Dukc
Mar 21, 2023
Kagamin
Jan 01, 2023
John Colvin
Jan 02, 2023
Dukc
December 24, 2022

Everyone in the D community has their own reasons for being here. We each have our own goals and plans, our own likes and dislikes, our own expectations and tolerance levels when those expectations go unmet, and so on.

When any given person's expectations aren't met over a period of time, whatever that period may be, frustration is understandable. But ranting in the forums about how Walter, or anyone involved in furthering D's development, doesn't care about the community is both unwarranted and unproductive. None of us would be here if we didn't care.

Discussions of D's problems and weaknesses, and the pet peeves of different D users, are scattered out across the forums, the community Discord, the IRC channel, Slack, and who knows where. When someone makes a general complaint in the forums, there's a certain amount of surprise and frustration when Walter asks for specific Bugzilla issues or links to past discussions, as if he's expected to have an eidetic memory.

The issues that you encounter in your code and frequently discuss with others in any given communications channel may be at the front of your mind, but that doesn't mean they are at the front of anyone else's mind. I certainly don't remember every complaint I've read about any given D feature. The issues at the front of my mind usually have nothing to do with any of that. I'm totally with Walter when he asks for specific details, or links to past discussions or Bugzilla issues. Otherwise, it's just noise.

So I have a proposal. Let's take a first step to bring some order to the chaos.

I invite every member of the D community to email me at social@dlang.org with your specific gripes about the current state of D and the things you'd like to see in the future (changes, new features, etc). But don't write in general terms. Be specific.

If your code is breaking every release, how is it breaking? If a certain language feature isn't working the way you expect, how do you expect it to work? Which unresolved issues have forced you into annoying workarounds, and what are those workarounds? What missing features would make your life easier? What are some important but unmaintained projects in the ecosystem that need attention?

Any D feature, any D tool or project in the ecosystem, is fair game. The important point here though is please be as detailed and specific as you can. Include any relevant links to Bugzilla issues or past forum discussions. I can always ask for more detail if I need to, but you'll save both of us time if you include it up front.

I will take this information as it comes in and look for commonalities and establish priorities. Over time, this should help me compile a list of actionable items that I can bring to future foundation meetings, and give us a starting point for serious discussions with community members on how we can solve specific issues that aren't so easily solved. And it will give the ecosystem management team a head start once they get off the ground. Moreover, I can post the list somewhere so that everyone can see it, and we'll all have a common point of reference and see the progress being made as issues are resolved.

By doing this, we can work to align the maintainers' goals with those of the community. I can't promise that Walter or anyone else on the team will agree that any given issue should be a priority, but I'm sure we'll identify a number of them. Nor can I promise that we'll be able to resolve any given issue in a reasonable time frame, given our lack of resources, but having them identified and prioritized is an important step toward ultimately resolving them.

I want to remind everyone that we're in this together. I would love it if we had a dedicated team of engineers who could methodically work through task lists. But we don't. Razvan and Dennis are doing a good bit of work beyond their primary job descriptions, like implementing approved DIPs that have been languishing unimplemented, but they don't have infinite time. So at some point, we'll need a few people to step up in one way or another.

But for now, just email me. Everyone in the D community who has the time and ability to post in the forums also has the time and ability to send me an email detailing their complaints. So please, don't be silent. Just redirect all that energy you've put into internet rants, or kept bottled inside as the case may be, into an email telling me all about your pain.

PLEASE NOTE: I won't be looking for your lists in this thread. If it turns into one of those multipage threads, I don't want to have to pick through every post looking for the details. Please email me. Then I can keep everything in one place with nice tags that make it easy for me to reference.

December 24, 2022

On Saturday, 24 December 2022 at 13:46:33 UTC, Mike Parker wrote:

>

PLEASE NOTE: I won't be looking for your lists in this thread. If it turns into one of those multipage threads, I don't want to have to pick through every post looking for the details. Please email me. Then I can keep everything in one place with nice tags that make it easy for me to reference.

A long time ago, I found a bug. Being a good little user/contributor, I reported it. It's here: https://issues.dlang.org/show_bug.cgi?id=15086

I then had to painfully convince the leadership that it is indeed a bug. I'm still not sure if Walter recognizes the actual issue.

Years later, Walter told us to post to the forums if we want to "get action" on issues. I did that. Nothing happened. [1][2]

Now you're asking me to email you. I'm not going to jump through that hoop. I have no reason to believe that this new channel is going to be more effective than the previous ones. Also, I want my airing of grievances to be public, not neatly hidden away in your personal emails.

You have a Bugzilla database full of bugs. You need to figure out how to fix them. You do not need new, less transparent channels to report them.

If Bugzilla seems too big to handle, that should tell you: Stop adding new features; start chipping away at the bug mountain.

[1] https://forum.dlang.org/post/rbeab9$20l1$1@digitalmars.com
[2] https://forum.dlang.org/post/rtppnb$3083$1@digitalmars.com

December 24, 2022

On Saturday, 24 December 2022 at 14:46:42 UTC, ag0aep6g wrote:

>

Now you're asking me to email you. I'm not going to jump through that hoop. I have no reason to believe that this new channel is going to be more effective than the previous ones. Also, I want my airing of grievances to be public, not neatly hidden away in your personal emails.

Whether you participate or not is up to you, but I want to reiterate that I'll be publishing everything I receive (not sure yet of the format). The issues won't be "hidden away" in my personal emails. The point here is for me to be able to stay organized. Anyone who sends me an email is free to copy and paste it in the forums, or shout it to the world if it isn't included on the public list.

And I encourage everyone, whatever your past experience, to think of this as a clean slate. And remember, this isn't just about Bugzilla issues.

Again, I can't promise Walter will agree any given complaint is a bug, or that there will be general agreement that your priorities should be broader priorities, or that any given issue will ever be resolved. However, I do believe that with enough feedback, this should help us to establish a roadmap going forward.

December 24, 2022
On 12/24/2022 6:46 AM, ag0aep6g wrote:
> On Saturday, 24 December 2022 at 13:46:33 UTC, Mike Parker wrote:
>> __PLEASE NOTE__: I won't be looking for your lists in this thread. If it turns into one of those multipage threads, I don't want to have to pick through every post looking for the details. __Please email me__. Then I can keep everything in one place with nice tags that make it easy for me to reference.
> 
> A long time ago, I found a bug. Being a good little user/contributor, I reported it. It's here: https://issues.dlang.org/show_bug.cgi?id=15086
> 
> I then had to painfully convince the leadership that it is indeed a bug. I'm still not sure if Walter recognizes the actual issue.

Your report was not ignored. It's that changing it would break existing code, something I get routinely excoriated about.

The -mv switch was added to deal with this problem, and is mentioned in the comments on the issue.

https://dlang.org/dmd-windows.html#switch-mv

There have also been two PRs by marler8997 about it.

https://github.com/dlang/dmd/pull/7778
https://github.com/dlang/dmd/pull/7878

though I'm not sure why he closed the second.


December 24, 2022
On Saturday, 24 December 2022 at 19:53:34 UTC, Walter Bright wrote:
> There have also been two PRs by marler8997 about it.
>
> https://github.com/dlang/dmd/pull/7778
> https://github.com/dlang/dmd/pull/7878
>
> though I'm not sure why he closed the second.

It seems pretty clear to me from the discussion: he closed it because Andrei told him that the rationale given for the PR was not convincing (reportedly after "a long discussion with @WalterBright"), and he did not feel like arguing the point.
December 24, 2022
On Saturday, 24 December 2022 at 19:53:34 UTC, Walter Bright wrote:
> Your report was not ignored.

I didn't say it was.

> It's that changing it would break existing code, something I get routinely excoriated about.

You use that excuse a lot, and it rarely truly applies. It doesn't apply here. People complain when you break their valid code. They don't complain when you make their buggy code fail to compile. Or rather: If someone complains about that, you should ignore them.

> The -mv switch was added to deal with this problem, and is mentioned in the comments on the issue.
>
> https://dlang.org/dmd-windows.html#switch-mv

That linked documentation is buggy itself. "path/filename" is missing in the syntax line. Someone should fix that, or at least file an issue. Not me though.

Also, that switch seems to be meant to ease the transition for when the bug gets fixed. But the bug hasn't been fixed.

> There have also been two PRs by marler8997 about it.
>
> https://github.com/dlang/dmd/pull/7778
> https://github.com/dlang/dmd/pull/7878
>
> though I'm not sure why he closed the second.

Really? Andrei argued against fixing the issue. Jonathan simply accepted that. Andrei was wrong, of course, but we can't blame Jonathan for yielding to the (supposedly benevolent) dictator.

For me, dealing with D's leadership has been an infuriating experience time and time again. It's the one reason why I have dialed back my contributions to near zero. There is a lot to fix in D, but when I have to fight for the privilege of cleaning up after you guys while you refuse to even recognize your own bugs, that's just not worth it.

I can't say for sure that Jonathan Marler stepped away for the same reason, but I am sure that poor handling of contributors has cost D a lot of talent.
December 25, 2022
On Saturday, 24 December 2022 at 23:22:47 UTC, ag0aep6g wrote:
> On Saturday, 24 December 2022 at 19:53:34 UTC, Walter Bright wrote:
>> Your report was not ignored.
>
> I didn't say it was.
>
>> It's that changing it would break existing code, something I get routinely excoriated about.
>
> You use that excuse a lot, and it rarely truly applies. It doesn't apply here. People complain when you break their valid code. They don't complain when you make their buggy code fail to compile. Or rather: If someone complains about that, you should ignore them.
>
>> The -mv switch was added to deal with this problem, and is mentioned in the comments on the issue.
>>
>> https://dlang.org/dmd-windows.html#switch-mv
>
> That linked documentation is buggy itself. "path/filename" is missing in the syntax line. Someone should fix that, or at least file an issue. Not me though.
>
> Also, that switch seems to be meant to ease the transition for when the bug gets fixed. But the bug hasn't been fixed.
>
>> There have also been two PRs by marler8997 about it.
>>
>> https://github.com/dlang/dmd/pull/7778
>> https://github.com/dlang/dmd/pull/7878
>>
>> though I'm not sure why he closed the second.
>
> Really? Andrei argued against fixing the issue. Jonathan simply accepted that. Andrei was wrong, of course, but we can't blame Jonathan for yielding to the (supposedly benevolent) dictator.
>
> For me, dealing with D's leadership has been an infuriating experience time and time again. It's the one reason why I have dialed back my contributions to near zero. There is a lot to fix in D, but when I have to fight for the privilege of cleaning up after you guys while you refuse to even recognize your own bugs, that's just not worth it.
>
> I can't say for sure that Jonathan Marler stepped away for the same reason, but I am sure that poor handling of contributors has cost D a lot of talent.

PR #7878 fixed a discrepency between "dmd -c main.d foo.d" and "dmd -c main.d && dmd -c foo.d". Andrei didn't think this discrepency was worth fixing. I was dumbfounded by this response and still am to this day.

I have alot of memories like this one from when I use to try to fix issues with D. I have a passion about my tools being correct, but, I found myself spending more time arguing with people than writing code. It was miserable.

Now I'm able to spend time doing what I love, writing robust-maintainable software with good tooling. When there's an issue with my toolchain, I'm able to submit a fix and it gets merged. It's night and day for me. I don't understand why there's was always so much disagreement/resistance when I contributed to D, I suppose I may never know. No hard feelings though, hope y'all are doing well.
December 25, 2022
Yeah, we have had a real problem with engaging contributors long-term. For instance even Walter has been blocking himself contributing lately.

I'm hoping to see this improve, Razvan does an amazing job as PR manager so far which has been a real boost to getting stuff in.

The desire is there to fix things.
December 25, 2022

On Saturday, 24 December 2022 at 13:46:33 UTC, Mike Parker wrote:

>

I invite every member of the D community to email me at social@dlang.org with your specific gripes about the current state of D and the things you'd like to see in the future (changes, new features, etc). But don't write in general terms. Be specific.

Thanks to everyone who has emailed me so far. I've received some long, thoughtful responses. Some of it has gone beyond the actionable and more into the philosophical/ideological, but that's welcome, too.

I hope to see more feedback in the coming days and weeks. I want a lot of feedback from several different people. Every email I've gotten so far has focused on different things. Some of it I've read before, but I need a large sample so I can identify the top most common gripes and wish list items.

So I'll periodically post a reminder here in the forums until I have enough to work with.

In the meantime, I hope everyone enjoys the holidays and has a safe and Happy New Year.

December 25, 2022
On Saturday, 24 December 2022 at 23:22:47 UTC, ag0aep6g wrote:

>
>> The -mv switch was added to deal with this problem, and is mentioned in the comments on the issue.
>>
>> https://dlang.org/dmd-windows.html#switch-mv
>
> That linked documentation is buggy itself. "path/filename" is missing in the syntax line. Someone should fix that, or at least file an issue. Not me though.
>

When the specification is lacking, it is not clear what code is valid. In the case of -mv, the spec says (or rather implies) package.module=path/filename, but the implementation also allows package=path, which I find reasonable and rely upon. If the implementaion gets 'fixed' to conform to the specification, I will be very unhappy.

« First   ‹ Prev
1 2 3 4 5 6