Bug/feature reqeust move to Sourceforge

[thomas]

I just check our exe installer on SF was affected, the result is: it still download the correct exe (with the name: codeblocks-12.11mingw-setup.exe), but people may concern whether SF may do some modification to this exe, so we need check-sum numbers.

Is this even possible in an easy, straightforward way? I mean, sure it's possible to add any kind of malware to any kind of executable (including an installer), but does our installer support adding a drive-by install once it has been compiled? That's NSIS, right?

In any case, checksumming and signing might rule that out (though of course nobody verifies checksums...).

[oBFusCATed]

As far as I understand it for now the feature is opt in and it is up to the projects to decide if they want it or not.
Also every project has some control on things that are included in the installer.

[thomas]

That's what I understood too, been searching the web for that thing the last hour or so. It really seems you have to opt-in, otherwise nothing happens. I've verified the download just to be 100% sure too, and it looks genuine.

[ollydbg]

Just for the interes, I havee placed a json file with the conten of our bug, feature and patch-tracker.
I used a slightly modified version of forgeplucker (a software that seems to be discontinued).

See: http://apt.jenslody.de/downloads/berlios.json.xz (about 2MB packed).

I create a test project in SF (https://sourceforge.net/p/ticketexport/) , and add some tickets there, then I export them, the result json file attached.

I compared the format of json format of berlios and SF, much similar, but we need a transfer program/script to translate the format.

a ticket of BerliOS:

Code Select Expand


{
                    "URL": "https://developer.berlios.de/bugs/?func=detailbug&bug_id=19187&group_id=5358",
                    "assigned_to": "jenslody",
                    "category": "Application::Crash",
                    "class": "ARTIFACT",
                    "comments": [
                        {
                            "class": "COMMENT",
                            "comment": "The current SVN revision 9435 crashes when I start C::B without options and wants to send a debug log. (Which has only modules information in it.)\n\nSVN revision 9425 worked fine.\n\nWhen starting with \ncodeblocks -d -v\nit works and no longer crashes after\n\n- A message on the console:\n(codeblocks:10887): Gtk-WARNING **: gtk_widget_size_allocate(): attempt to allocate widget with width 18 and height -16\n\n- A popup with the text\n\"iCCP: known incorrect sRGB profile\"\n-> Details shows the same message 9 times.\n\nAfter clicking \"OK\" on the popup I can work with C::B normally.\n",
                            "date": "2013-11-08T10:44:00Z",
                            "submitter": {
                                "class": "IDENTITY",
                                "nick": "yamakuzure"
                            }
                        }
                    ],
                    "dependents": [],
                    "group": "Platform:Linux",
                    "history": [
                        {
                            "by": "jenslody",
                            "class": "FIELDCHANGE",
                            "date": "2013-11-08T21:46:00Z",
                            "field": "assigned_to",
                            "new": "jenslody",
                            "old": "none"
                        }
                    ],
                    "id": 19187,
                    "priority": "5 - Medium",
                    "resolution": "None",
                    "status": "Open",
                    "summary": "[SVN 9435] Crash without debug mode (iCCP, Gtk)",
                    "type": "bugs"
                },
               


a ticket of SF:

Code Select Expand


{
  "status": "open",
  "reported_by_id": "4d215fc8b9363c7d7a0004b9",
  "related_artifacts": [],
  "attachments": [],
  "reported_by": "ollydbg",
  "description": "Report a bug.",
  "labels": [
    "CodeCompletion"
  ],
  "assigned_to": "nobody",
  "assigned_to_id": null,
  "private": false,
  "summary": "Second ticket",
  "discussion_thread": {
    "_id": "3a7bb9c2",
    "posts": [
      {
        "text": "Add comments in second ticket",
        "attachments": [],
        "author": "ollydbg",
        "timestamp": "2013-11-15 01:59:37.406000",
        "slug": "5b9c",
        "subject": "#2 Second ticket"
      }
    ],
    "discussion_id": "52857ef7b9363c0db81f5bc5",
    "subject": ""
  },
  "mod_date": "2013-11-15 01:59:37.882000",
  "votes_down": 0,
  "votes_up": 0,
  "_id": "52857fcb3e5e8370b9e983b9",
  "discussion_thread_url": "http://sourceforge.net/rest/p/ticketexport/tickets/_discuss/thread/3a7bb9c2/",
  "ticket_num": 2,
  "custom_fields": {
    "_milestone": "1.0"
  },
  "created_date": "2013-11-15 01:58:35.229000"
}


I'm using Notepad++ and JStool plugin (can be easily installed by plugin manager in Notepad++) to view the json object.

[attachment deleted by admin]

[Jenna]

It should not be too hard to translate one json object in another.
PHP can work with json quiet good, Python also as you can see in my dump (not my script, I don't know much about python)  and (of course) javascript can.

Of these three I am most familiar with PHP and javascript, so I could write a tool to convert this.

Did you notice, that features and patches are also dumped (with a slightly different format of course) ?

Besides of this, I have no good feeling with sf.net.

[ollydbg]

It should not be too hard to translate one json object in another.
PHP can work with json quiet good, Python also as you can see in my dump (not my script, I don't know much about python)  and (of course) javascript can.

Of these three I am most familiar with PHP and javascript, so I could write a tool to convert this.

Thanks

Did you notice, that features and patches are also dumped (with a slightly different format of course) ?

In my testing site: https://sourceforge.net/p/ticketexport/tickets/
I only have "tickets" (no bugs/feature-requests/patches)
But I do see some site has those three categories, like the one below:
https://sourceforge.net/p/notepad-plus/_list/tickets
So, I will try to create some sub categories on my testing site, and try to see whether bugs/feature-requests/patches are exported in different json format.

Besides of this, I have no good feeling with sf.net.

I have no idea, SF's ticket system looks good(at least better than berlios).
why not we create our own on codeblocks.org? (since no devs have suggested to do that).

[oBFusCATed]

I don't think switching to the bug tracker of sf.net will be big improvement.
Also I don't think we need to transfer anything.
Most of the problems require user input and we'll loose it.

[dmoore]

I don't think switching to the bug tracker of sf.net will be big improvement.
Also I don't think we need to transfer anything.
Most of the problems require user input and we'll loose it.

Can't we automate a comment on every open bug  in berlios linking to sourceforge/whatever solution we devise telling the reporter to subscribe to/update the bug on the new tracker? That should send an email notification to every reporter.

Noone thinks github issues is a good idea? It's a more user friendly and less rigid system, which I think works well with smaller open source projects.

[oBFusCATed]

Noone thinks github issues is a good idea? It's a more user friendly and less rigid system, which I think works well with smaller open source projects.

We won't switch to git anytime soon, so there is little point in switching to github for bugs/features.

[dmoore]

Noone thinks github issues is a good idea? It's a more user friendly and less rigid system, which I think works well with smaller open source projects.

We won't switch to git anytime soon, so there is little point in switching to github for bugs/features.

Just thinking of the users... Even if we aren't using the pull request system, their issues UI is still a shitload nicer to use than alternatives. Are there any other bug tracking system like it?

To give an idea, here are issues for node.js, one of the bigger github projects:

https://github.com/joyent/node/issues

[oBFusCATed]

... than alternatives. ..

What do you target as alternatives?

[dmoore]

... than alternatives. ..

What do you target as alternatives?

berlios, trac, launchpad. anything else?

[oBFusCATed]

www.jetbrains.com/youtrack/ (a friend is using it, they have free plans for opensource projects)
Jira (I've heard it is awful, but I've not used yet)
Mantis
Bugzilla

[hinojosad]

I see bad news about Sourceforge:
http://www.gluster.org/2013/08/how-far-the-once-mighty-sourceforge-has-fallen/
and
http://www.gimp.org/ (the title is: GIMP Windows Installers move from Sourceforge to ftp.gimp.org)

EDIT: I just check our exe installer on SF was affected, the result is: it still download the correct exe (with the name: codeblocks-12.11mingw-setup.exe), but people may concern whether SF may do some modification to this exe, so we need check-sum numbers.

Say, I'm the Community Manager at SourceForge; I wanted to assure you, as you have seen, we do not impose the DevShare program on projects; DevShare is fully opt in for projects and users.

You can reference our blog post where we announced the program (http://sourceforge.net/blog/today-we-offer-devshare-beta-a-sustainable-way-to-fund-open-source-software/), as well as where we addressed the concerns raised by our community (http://sourceforge.net/blog/advertising-bundling-community-and-criticism/).

If you all need or desire any further clarification, I am available to help.

Best regards,

Daniel.
--

[Alpha]

As another idea, I recently came across Phabricator.  (I am not sure what to make of it though.)