June 23, 2023
On Thursday, 22 June 2023 at 20:48:09 UTC, Jonathan M Davis wrote:

> I've never understood why some folks have kept pushing for us to use github issues. They're nice if you're setting up a github repo and don't want to go to the trouble of setting up and hosting a bugzilla instance elsewhere, but they're not as sophisticated as bugzilla.
>
> - Jonathan M Davis

Because its not about the features, its about the people. D's bugzilla has none, github has all of them.
June 23, 2023

On Thursday, 22 June 2023 at 15:50:06 UTC, Vladimir Panteleev wrote:

>

I still see some issues:

  • Please import one BZ comment as one GH comment, so that it's still possible to link to a comment. Comments to links should be converted to point to the right comment, or else we may lose context.

No, github is rate limited, 100 writes per ip per hour. One comment at a time
will end this migration project.

>
  • Please put the description and every comment in a Markdown block, otherwise things are going to be corrupted due to the Markdown formatting.

  • Please try it with some bugs from the DMD project before doing the real migration. You should see a much higher variety of "interesting" bug reports that you would not see otherwise.

  • Make sure to actually try moving ALL bugs to a private test repository before doing the real migration, for the same reasons.

I think this is just stalling stuff, nobody has time to take a meaningful look at all the dmd issues. Nothing is lost in the migration as the old data still exists in the same place.

>
  • Have you considered using the same software that the LLVM project used for their migration? Writing our own tool seems wasteful if existing proven software exists.

They have no software they have people https://github.com/llvm/llvm-iwg/issues/56#issuecomment-1001707975 basically they have github people doing the have lifting.

>
  • We have agreed that a prerequisite of doing this migration was ensuring that we can still download the entire data off of GitHub, and/or that we maintain a live independent copy/backup in case we lose access to the data. Has this been done?

Live was never on the table. I have not written the tool. But that tool is easy.

June 23, 2023

On Thursday, 22 June 2023 at 15:50:06 UTC, Vladimir Panteleev wrote:

>

I still see some issues:

  • Please import one BZ comment as one GH comment, so that it's still possible to link to a comment. Comments to links should be converted to point to the right comment, or else we may lose context.

  • Please put the description and every comment in a Markdown block, otherwise things are going to be corrupted due to the Markdown formatting.

Sounds good. I'll discuss it with Robert.

>
  • Please try it with some bugs from the DMD project before doing the real migration. You should see a much higher variety of "interesting" bug reports that you would not see otherwise.

  • Make sure to actually try moving ALL bugs to a private test repository before doing the real migration, for the same reasons.

Agreed.

>
  • Have you considered using the same software that the LLVM project used for their migration? Writing our own tool seems wasteful if existing proven software exists.

I didn't look into it. I think Mathias did when he first explored this a couple years ago, but I may be wrong.

>
  • We have agreed that a prerequisite of doing this migration was ensuring that we can still download the entire data off of GitHub, and/or that we maintain a live independent copy/backup in case we lose access to the data. Has this been done?

There are a few different options out there for this. We can download everything anytime via GitHub's API. GH actually released a tool for this a couple of years ago, but archived the repo a few months back. However, I'm leaning toward a paid service that makes regular backups of designated repositories to a cloud storage provider (which would be Google Drive in our case).

June 23, 2023

On Friday, 23 June 2023 at 08:34:43 UTC, Robert Schadek wrote:

>

No, github is rate limited, 100 writes per ip per hour. One comment at a time will end this migration project.

[...]

They have no software they have people https://github.com/llvm/llvm-iwg/issues/56#issuecomment-1001707975 basically they have github people doing the have lifting.

It's surely both, right?

In any case, have we tried reaching out to GitHub to see if they can do the same for us?

If that doesn't work, we can do an incremental migration, with issues being moved one by one. Then, it is not time constrained.

>

nobody has time to take a meaningful look at all the dmd issues.

We definitely shouldn't look at all of them, but a good sample of combinations of various interesting things like attachments or different formatting of example programs would be useful to check.

>

Nothing is lost in the migration as the old data still exists in the same place.

You're right that it should be possible to improve the formatting of imported issues after-the-fact, by editing the issue descriptions and comments. So, we should focus on things that will be difficult to change after-the-fact, such as URLs, preserving issue numbers, and preserving the comments-to-comments mapping.

>

I think this is just stalling stuff

No. Please, let's get this right. If you are not interested in ensuring we do everything to get it right, please let me know and I'll continue from here.

As I mentioned at the meeting from 2021, I'm not against migrating, but it needs to be done well so that we have no regrets afterwards.

June 23, 2023
On Friday, June 23, 2023 2:14:29 AM MDT Robert Schadek via Digitalmars-d wrote:
> On Thursday, 22 June 2023 at 20:48:09 UTC, Jonathan M Davis wrote:
> > I've never understood why some folks have kept pushing for us to use github issues. They're nice if you're setting up a github repo and don't want to go to the trouble of setting up and hosting a bugzilla instance elsewhere, but they're not as sophisticated as bugzilla.
> >
> > - Jonathan M Davis
>
> Because its not about the features, its about the people. D's bugzilla has none, github has all of them.

I'm afraid that I don't understand your logic. When you want to report a bug for a project, you go to wherever that project handles bug reports. In that respect, whether it's bugzilla or github or wherever is irrelevant. I don't see how the fact that a lot of people use github for code has anything do with how we handle bug reports.

Walter and company will decide whatever they decide for this, and if that means going to github issues, I guess that it means going to github issues, but it seems to me that it's making things worse, because we'll be switching to an inferior tool. And I truly don't understand how any of it is about the people. Are you afraid that people are too lazy to report bugs via bugzilla or something and that using github issues will somehow overcome that?

- Jonathan M Davis



June 23, 2023
On Friday, 23 June 2023 at 08:59:56 UTC, Jonathan M Davis wrote:
>
> I'm afraid that I don't understand your logic. When you want to report a bug for a project, you go to wherever that project handles bug reports. In that respect, whether it's bugzilla or github or wherever is irrelevant. I don't see how the fact that a lot of people use github for code has anything do with how we handle bug reports.

I fear you're wrong.
If a regular (not on this forum person) goes to actually write a bug-report.
They go like.

1. google dlang
2. WAT issues are in a tool I have to log in that is not github; I can not be bothered.

>
> Walter and company will decide whatever they decide for this, and if that means going to github issues, I guess that it means going to github issues, but it seems to me that it's making things worse, because we'll be switching to an inferior tool. And I truly don't understand how any of it is about the people. Are you afraid that people are too lazy to report bugs via bugzilla or something and that using github issues will somehow overcome that?
>

I'm not afraid that people are too lazy, I'm convinced, and IMHO right, that people are too lazy report anywhere but github.
Github helps as everybody knows it, and there is no friction at all.




June 23, 2023
On 6/23/23 10:42, Vladimir Panteleev wrote:
> 
>> I think this is just stalling stuff
> 
> No. Please, let's get this right. If you are not interested in ensuring we do everything to get it right, please let me know and I'll continue from here.
> 
> As I mentioned at the meeting from 2021, I'm not against migrating, but it needs to be done well so that we have no regrets afterwards.

+1.
June 23, 2023
On 6/23/23 11:53, Robert Schadek wrote:
>>
>> Walter and company will decide whatever they decide for this, and if that means going to github issues, I guess that it means going to github issues, but it seems to me that it's making things worse, because we'll be switching to an inferior tool. And I truly don't understand how any of it is about the people. Are you afraid that people are too lazy to report bugs via bugzilla or something and that using github issues will somehow overcome that?
>>
> 
> I'm not afraid that people are too lazy, I'm convinced, and IMHO right, that people are too lazy report anywhere but github.
> Github helps as everybody knows it, and there is no friction at all.
> 

Yes, unfortunately it is true that many people won't report issues on bugzilla. Preserving the existing issues is still really important as well. Those are issues reported by _enthusiastic people who let themselves be assed to report on bugzilla_, and we owe them to treat their issues well. In particular, I think moving issues to github issues and giving all the github contributions/authorship for those issues to yourself and Mike is a terrible way to move forward, even if the issues are technically in the public domain and/or belong to DLF.
June 23, 2023
On Friday, 23 June 2023 at 12:31:39 UTC, Timon Gehr wrote:
> On 6/23/23 11:53, Robert Schadek wrote:
>>>

>
> Yes, unfortunately it is true that many people won't report issues on bugzilla. Preserving the existing issues is still really important as well. Those are issues reported by _enthusiastic people who let themselves be assed to report on bugzilla_, and we owe them to treat their issues well. In particular, I think moving issues to github issues and giving all the github contributions/authorship for those issues to yourself and Mike is a terrible way to move forward, even if the issues are technically in the public domain and/or belong to DLF.

For the record, I intend to create a special account for adding the issues to GitHub (e.g., "bugzilla_migration"). I certainly don't want them all in my name. I don't see how we can post them as the original user, though. From what I understand, it requires the id and a token.
June 23, 2023

On Friday, 23 June 2023 at 08:59:56 UTC, Jonathan M Davis wrote:

>

And I truly don't understand how any of it is about the people. Are you afraid that people are too lazy to report bugs via bugzilla or something and that using github issues will somehow overcome that?

While my original comment about the spam defense issue wasn’t directed at you and I’ve been only using your comment as reference for my sarcastic remark, I’d like to invite you to come to Discord and throw “Gmail” into the search there.

… or actually… I don’t want to bother you with this… so I did the busywork for you :)

2023-06-07
>

For bugzilla, besides gmail restriction, are there any others?

2023-06-01
>

Because all our code is at GH, if you have a GH account you don't need to create a Bugzilla account just to file, some developers won't touch Bugzilla, we don't have direct admin access to our Bugzilla (the admin disabled account creation with gmail addresses because of spam), etc.

^ which btw also implies that the Bugzilla instance isn’t even exactly run by the DLF

>

The gmail thing is truly egregious. I still don't understand it.

>

Tons of people have gmail addresses […]. To exclude all of them from posting bugs gives off the impression that we don’t care about our bugs.

2023-03-13
>

i don't have bugzilla account 😦 and I remember it had some troubles with gmail.. […]

2023-01-07
>

I bet the number of would-be bug reporters that have gmail addresses that just don't do it is in the dozens by now. Is this seriously still an issue?

>

gmail can't post issue on zilla y?

2022-11-19
>

So I create a gmail account to register to the bugzilla, I mean here it says that you can use Gmail: [screenshot]

>

We recommend using a secondary account or free web email service (such as Gmail, Yahoo, Hotmail, or similar) to avoid receiving spam at your primary email address.

So I have to go through the hassle of making a new google account linking it with my phone to get notifications from the bugzilla just to get told that they don't accept gmail after it was clearly stated that I could use gmail?
I'm done with this

2022-09-21
>

Okay, but I can't do that since I can't make an account there (since it doesn't accept gmail)

2022-09-15
>

We really need to fix the "no gmail addresses on bugzilla" problem.

2022-09-04
  1. >

    Is bugzilla the place to report phobos bugs?

  2. >

    yes
    issues.dlang.org

  3. >

    The e-mail address you entered (blah.blah@gmail.com) didn't pass our syntax checking for a legal email address. A legal address must contain exactly one '@', and at least one '.' after the @. Currently, registering using Gmail addresses is not allowed due to spam. It also must not contain any illegal characters.

  4. >

    you will need to post on the N.G. to get an account created manually

  5. >

    Well in case anyone who has an account feels like reporting it.
    std.experimental.logger broke

>

Hey Mike, we had somebody get a deny on creating a bugzilla account due to gmail email account

2022-03-24
>

Robert's argument for the move (and those who supported on it) is that GitHub is where the developers are. We currently require anyone submitting Bugzilla issues to create an account, and account creation is currently disabled by the maintainer for anyone with a gmail address because of spam!

2021-12-20
>

the issue tracker: "we recommend using gmail to avoid spam"
also the issue tracker, but only after you enter a gmail address: "Currently, registering using Gmail addresses is not allowed due to spam."
trolled

>

blocking gmail adresses seems like a great idea

Long story, short

It comes up regulary.