All the Perl that's Practical to Extract and Report

Use the damned request tracker!

schwern on 2005-02-18T07:02:29

I think I found a bug in Test::More, but before I file it in the request tracker, I wanted to confirm with you.

ARGH! I hate getting mail like this. People are trying to be nice but they're actually making my life harder. When the bug comes to me directly I have to deal with it immediately rather than run the risk of it sinking to the murky, fossilized depths of my inbox. Or I have to enter the bug into RT myself taking up more time. Any discussion we have is lost to the ether.

Why do people do this? The only thing I can figure is that there's some perceived stigma associated with "officially" reporting a bug for all to see. Let me make it clear, there is no stigma associated with reporting a bug via rt.cpan.org! You're not accusing me of writing shoddy software. I don't care what my public bug count is. And if its not a bug I'll just reply and resolve it. If it is a bug its in the system, it won't get lost, and we have a nice log of the conversation. It makes my life easier.

Use the request tracker. That's what its there for, to track requests so I don't have to.

Bounce

davorg on 2005-02-18T08:14:48

That's what the 'b' key is for in mutt. I just bounce it to RT and it sits there appearing to have come from the original sender.

Re:Bounce

nicholas on 2005-02-18T11:34:40

I've been bouncing direct mail to perl5-porters for a while. Sod ettiquette - if someone is not able to read the fine manual on where to report problems, I'm going to take direct action. I've also changed my public e-mail address on PAUSE to nwc10+please+use+perlbug+for+perl+queries@colon.colondot.net (and yes, that is a valid address that works, although MBM's MTA is deliberately very very picky and reasons it gives for bouncing mail are probably valid). acme said that he used to get a lot of private mail about perl 5 when 5.005_04 was the most recent official release - I can only infer that people saw my previous public PAUSE e-mail address (possibly via search.cpan.org) and thought that talking.bollo.cx wasn't a real domain, whereas astray.com conforms to their expectations.

$ host -t mx talking.bollo.cx
talking.bollo.cx mail is handled (pri=30) by mango.news.easynet.net
talking.bollo.cx mail is handled (pri=10) by plum.flirble.org
talking.bollo.cx mail is handled (pri=20) by peach.flirble.org
talking.bollo.cx mail is handled (pri=30) by lemon.news.easynet.net

Looks real enough to me :-)

Re:Bounce

schwern on 2005-02-18T12:23:27

So I can just bounce it to bug-Module-Name@rt.cpan.org? Brilliant!

Re:Bounce

davorg on 2005-02-18T12:40:13

Yep. Isn't science wonderful :)

seen rgs

rafael on 2005-02-18T08:29:50

Enhancement requests via IRC are worse.

Tell them how to report bugs in the docs

brian_d_foy on 2005-02-18T17:33:06

I don't think most people know about RT. Hardly anyone advertises it or tells people that it even exists. There are more people who don't follow the mailing lists, discussion fora, and blogs than those who do. Some long time Perl users don't even know about CPAN (and I'm always shocked to see people feverishly copying down the address when I mention it in an advanced Perl class.)

The Test::More documentation that I have doesn't tell the user to report bugs to RT. Indeed, it uses your pobox address instead of your cpan.org address. :)

Re:Tell them how to report bugs in the docs

schwern on 2005-02-23T14:24:49

Test::Harness does this. Note that the rt.cpan.org ticket numbers went from about 3000 to 11000 in about a month. Spammers absolutely clobbered the queue. Seems to be under control now though.

But yes, I should probably mention it in some form in the docs. However, in this case, the reporter deliberately side-stepped RT though it turns out there were ulterior motives.

A-feared of being wrong?

MarkyMoon on 2005-02-19T14:11:03

Although I've never tried reporting a suspected bug - In fact if I can't get something to work, my first thought is that it's a PEBKAC error, not a problem in the software;) - I imagine some people are afraid to have their own misunderstanding officialy logged for eveyone to see and giggle at.

Re:A-feared of being wrong?

mako132 on 2005-02-21T23:52:52

>
> I imagine some people are afraid to have their own misunderstanding officialy logged for eveyone to see and giggle at.
>

I think this is significant. In my own department I introduced RT for sys admin support requests. Another closely associated department liked it, but also wanted to have all comments and replies cc-ed to the mailing list that everyone was on.

The result was that staff stopped using it on their own unless directed by management to submit a ticket. Management continued to use it to submit requests, but staff stopped replying to tickets and instead either spoke directly to the system administrators or replied by sending email about the tickets directly to sys admins or management.

I initially thought people got frustrated with the RT interface (a recent update slowed RT's responsiveness), but now I see something else - people are reluctant to submit a ticket for fear of "officially" logging what could be their own misunderstanding - or - fear of looking bad when the resolution the ticket was some minor config change or RTFM that they could have done themselves.

Plus, anytime management stepped into a ticket to stop a resolution (e.g. they didn't want the sys admin staff to spend time on a given issue, or didn't want to pay for the solution, upgrade, update, etc.) staff got scared.

My own feeling (the reason why I introduced RT in the first place) was that I felt this could keep everyone informed of what was going on and we could all learn from our mistakes.

But now I see just how strong the desire to not look bad is.