On Saturday, 28 March 2015 at 04:36:18 UTC, Walter Bright wrote:
> On 3/26/2015 3:53 PM, ketmar wrote:
>> filling bugs like "this huge project not compiling!" is not working, as
>> nobody wants to run dustmite on such projects, people just waiting for
>> issue author to provide more information.
>
> Realistically, people who want to work on bug fixing are going to work on ones that have already been isolated and filed.

There are also people who don't mind helping to reduce bugs so the compiler hackers can fix them quickly.

But honestly, there already exists so much information on how to use DustMite, I don't know what else can be said. ANYONE should be able to use DustMite or Digger to reduce a test case down to reasonable size.

On Fri, 27 Mar 2015 21:36:15 -0700, Walter Bright wrote:

> On 3/26/2015 3:53 PM, ketmar wrote:
>> filling bugs like "this huge project not compiling!" is not working, as nobody wants to run dustmite on such projects, people just waiting for issue author to provide more information.
> 
> Realistically, people who want to work on bug fixing are going to work on ones that have already been isolated and filed.
> 
> If you've got a "huge project that's not compiling" and don't know where to start, that implies it isn't well modularized and encapsulated.

thank you for supporting my POV on not reporting bugs. that is what i wanted to told to John, yes.

i must admit, though, that it wasn't even project of mine. but nevermind.

On Sat, 28 Mar 2015 04:55:47 +0000, Vladimir Panteleev wrote:

> But honestly, there already exists so much information on how to use DustMite...

...that people in bugzilla keep asking what it is.

> ANYONE should be able to
> use DustMite or Digger to reduce a test case down to reasonable size.

having a big codebase that you didn't wrote and never read took 12 hours to dustmite. not that i can just leave it unattended though, as compiler itself segfaults sometimes, and that effectively leaves dustmite frozen. so it not only eats resources of my box (and i have a work to do, and that work involves compiling big codebases too), but it requires my attention. but yes, it's entirely my fault that i cannot afford such resources and asking for help, i know.

On Saturday, 28 March 2015 at 04:36:18 UTC, Walter Bright wrote:
> On 3/26/2015 3:53 PM, ketmar wrote:
>> filling bugs like "this huge project not compiling!" is not working, as
>> nobody wants to run dustmite on such projects, people just waiting for
>> issue author to provide more information.
>
> Realistically, people who want to work on bug fixing are going to work on ones that have already been isolated and filed.
>
> If you've got a "huge project that's not compiling" and don't know where to start, that implies it isn't well modularized and encapsulated.

That being said I rarely face bugs in a single module. Usually bug arise in situation like instantiate the a template from another template in another module by passing an alias parameter from a symbol in a 3rd module.

Dustmite help to get a smaller repro case, but it literally takes ages to get one.

On 3/27/2015 11:06 PM, deadalnix wrote:
> That being said I rarely face bugs in a single module. Usually bug arise in
> situation like instantiate the a template from another template in another
> module by passing an alias parameter from a symbol in a 3rd module.

I've noticed this problem with Phobos, which is why I've been advocating breaking up the modules into encapsulated pieces.

On Saturday, 28 March 2015 at 05:35:57 UTC, ketmar wrote:
> On Sat, 28 Mar 2015 04:55:47 +0000, Vladimir Panteleev wrote:
>
>> But honestly, there already exists so much information on how to use
>> DustMite...
>
> ...that people in bugzilla keep asking what it is.

Not knowing what something is and not wanting to learn how to use it are different things.

>> ANYONE should be able to
>> use DustMite or Digger to reduce a test case down to reasonable size.
>
> having a big codebase that you didn't wrote and never read took 12 hours
> to dustmite. not that i can just leave it unattended though, as compiler
> itself segfaults sometimes, and that effectively leaves dustmite frozen.
> so it not only eats resources of my box (and i have a work to do, and
> that work involves compiling big codebases too), but it requires my
> attention. but yes, it's entirely my fault that i cannot afford such
> resources and asking for help, i know.

Honestly, did you even try?

https://github.com/CyberShadow/DustMite/wiki/Detecting-a-segfault-in-dmd-itself
https://github.com/CyberShadow/DustMite/wiki/Running-commands-with-a-timeout
https://github.com/CyberShadow/DustMite/wiki/Useful-test-scripts

Or did you just give up after the first difficulty, saying, "well, I tried"?

Do you think your time is more valuable than that of D contributors' or something?

On Sat, 28 Mar 2015 14:12:17 +0000, Vladimir Panteleev wrote:


> Honestly, did you even try?

how do you think, where that "12 hours" came from?

> Do you think your time is more valuable than that of D contributors' or something?

sure. main D developers shown that they have no respect for other's work (see Andrei calling H.S.Teoh's work of splitting std.algorithm "useless", or Walter blaming me that "the project is badly designed" when it wasn't even my project and i didn't wrote a single line there, and have no understanding of codebase at all), so why should i think that my time is less valuable than theirs?

On Saturday, 28 March 2015 at 14:12:19 UTC, Vladimir Panteleev wrote:
> On Saturday, 28 March 2015 at 05:35:57 UTC, ketmar wrote:
>> On Sat, 28 Mar 2015 04:55:47 +0000, Vladimir Panteleev wrote:
>>
>>> But honestly, there already exists so much information on how to use
>>> DustMite...
>>
>> ...that people in bugzilla keep asking what it is.
>
> Not knowing what something is and not wanting to learn how to use it are different things.
>
>>> ANYONE should be able to
>>> use DustMite or Digger to reduce a test case down to reasonable size.
>>
>> having a big codebase that you didn't wrote and never read took 12 hours
>> to dustmite. not that i can just leave it unattended though, as compiler
>> itself segfaults sometimes, and that effectively leaves dustmite frozen.
>> so it not only eats resources of my box (and i have a work to do, and
>> that work involves compiling big codebases too), but it requires my
>> attention. but yes, it's entirely my fault that i cannot afford such
>> resources and asking for help, i know.
>
> Honestly, did you even try?
>
> https://github.com/CyberShadow/DustMite/wiki/Detecting-a-segfault-in-dmd-itself
> https://github.com/CyberShadow/DustMite/wiki/Running-commands-with-a-timeout
> https://github.com/CyberShadow/DustMite/wiki/Useful-test-scripts
>
> Or did you just give up after the first difficulty, saying, "well, I tried"?
>
> Do you think your time is more valuable than that of D contributors' or something?

This attitude is crap and is becoming more frequent on the forums.

The D development team is not interested in listening to their user base unless the user base is willing to contribute back to D language development with PRs.

Good luck with that because most end-users will not bother even trying to file a bug report, let alone distill it down with some tool in the compiler download. They'll just move on in another language that doesn't require effort fighting compiler/language bugs.

bye,
lobo

On Sun, 29 Mar 2015 02:47:40 +0000, lobo wrote:

> This attitude is crap and is becoming more frequent on the forums.

the funny thing is that D devs managed to convert me from loyal adopter to "asshole". and this has nothing to do with rejecting my ideas per se, btw. the situation now is... bizarre: i love D with passion, but i don't like W&A "driving force" almost with the same passion. that doesn't mean that i'm not grateful to W&A for creating this piece of art, though.

On Sunday, 29 March 2015 at 02:47:42 UTC, lobo wrote:
> On Saturday, 28 March 2015 at 14:12:19 UTC, Vladimir Panteleev wrote:
>> Do you think your time is more valuable than that of D contributors' or something?
>
> This attitude is crap and is becoming more frequent on the forums.
>
> The D development team is not interested in listening to their user base unless the user base is willing to contribute back to D language development with PRs.

Not sure how this is related to the discussion at hand. Contributors are more likely to have enough experience with the language to make better suggestions, though.

I think that from an outside perspective, it's hard to lose sight that D has no well-defined "devteam". Aside Walter and Andrei, there are no cabals or inner circles. D contributors are all D users who also like to spend some time to improve D.

> Good luck with that because most end-users will not bother even trying to file a bug report, let alone distill it down with some tool in the compiler download. They'll just move on in another language that doesn't require effort fighting compiler/language bugs.

Missing the point.

Here is what Ketmar could have done:

1. Create a regression issue with the unreduced test case. This is OK. I occasionally check for unreduced regressions, though usually someone beats me to it.

2. Ask for help with the reduction (e.g. digitalmars.D.learn). He said he did, but I don't see where! Posting deep in some thread doesn't count, very few people read the entire forum.

3. Contact me directly for assistance in using DustMite. I'd be happy to help.

I don't think these are unreasonable expectations. If you fail or refuse to at least do the above, and then complain how everything and everyone is horrible, I have no sympathy for you.