From tom at eborcom.com Thu Dec 1 15:40:02 2005 From: tom at eborcom.com (Tom Hukins) Date: Thu, 1 Dec 2005 23:40:02 +0000 Subject: Managing Projects with Perl Message-ID: <20051201234002.GA56357@eborcom.com> Imagine this: you find yourself in a new job with no version control, minimal documentation (mostly outdated), and no formal bug tracking system. In this hypothetical world, you might introduce Subversion, with the support of a colleague who also values version control. You might find you also have an internal bug tracking system that nobody uses because it's a little clunky. But you persuade them to use it more. You think that a Wiki would help with documentation. You acknowledge you'll have to persuade people to use it, but recognise this as a non-technical problem. And you're happy with that. A few months later you might find yourself discussing this with various people who happen to subscribe to a certain mailing list. Just maybe. And later you might decide to post to that list. And you might happily run Subversion; you might like Mediawiki's interface; and you might appreciate that some strange bug tracker is better than nothing. But you still miss RT. But you feel a little uncomfortable that the tools you like weren't designed to integrate. You look at trac and envy the way it integrates issue tracking, Wikis and version control with the same Python envy that causes us all to use Mailman. trac looks really nice, but it's not Perl. Imagine you could write rt#1234 in your Subversion commit log or your Wiki, and your tools recognised this as a ticket number and automatically linked to it. Imagine your bug tracker recognised WikiWords and automatically linked to ThingsLikeThis in your Wiki. What would you do? Would you install trac? Would you try to tie together RT with SVN::Web (hoping you can bribe its maintainer with oriental buffets when required) and some sort of Wiki that doesn't exist (I hate Kwiki: it has horrible edge cases) or would you accept that this particular holy trinity was never meant to integrate? Or would you do something else? Not that I actually care and this is a real problem I'm trying to tackle or anything. Of course. Obviously. Tom From peterdragon at users.sourceforge.net Thu Dec 1 16:03:47 2005 From: peterdragon at users.sourceforge.net (Peter Edwards) Date: Fri, 2 Dec 2005 00:03:47 -0000 Subject: Managing Projects with Perl Message-ID: <005a01c5f6d3$e043cfc0$6501a8c0@metis> Just today I sent an email extolling the joys of Trac and Subversion to the software company I do most of my work for. The situation is almost the same as yours and I am trying to persuade them to use it. Trac is great, even though it's written in python and uses SQlite (or postgres, we use Mysql). It's not as fully featured as RT, for example no email interface to raise issues, though you could easily write one. Its strength is in the integration. Said customer has an issue tracking system, a support system, a change control system, a project plan, and *none of them are hooked up together*. Agh. Regards, Peter -----Original Message----- From: miltonkeynes-pm-bounces at pm.org [mailto:miltonkeynes-pm-bounces at pm.org] On Behalf Of Tom Hukins Sent: 01 December 2005 23:40 To: Milton Keynes Perl Mongers Subject: Managing Projects with Perl Imagine this: you find yourself in a new job with no version control, minimal documentation (mostly outdated), and no formal bug tracking system. In this hypothetical world, you might introduce Subversion, with the support of a colleague who also values version control. You might find you also have an internal bug tracking system that nobody uses because it's a little clunky. But you persuade them to use it more. You think that a Wiki would help with documentation. You acknowledge you'll have to persuade people to use it, but recognise this as a non-technical problem. And you're happy with that. A few months later you might find yourself discussing this with various people who happen to subscribe to a certain mailing list. Just maybe. And later you might decide to post to that list. And you might happily run Subversion; you might like Mediawiki's interface; and you might appreciate that some strange bug tracker is better than nothing. But you still miss RT. But you feel a little uncomfortable that the tools you like weren't designed to integrate. You look at trac and envy the way it integrates issue tracking, Wikis and version control with the same Python envy that causes us all to use Mailman. trac looks really nice, but it's not Perl. Imagine you could write rt#1234 in your Subversion commit log or your Wiki, and your tools recognised this as a ticket number and automatically linked to it. Imagine your bug tracker recognised WikiWords and automatically linked to ThingsLikeThis in your Wiki. What would you do? Would you install trac? Would you try to tie together RT with SVN::Web (hoping you can bribe its maintainer with oriental buffets when required) and some sort of Wiki that doesn't exist (I hate Kwiki: it has horrible edge cases) or would you accept that this particular holy trinity was never meant to integrate? Or would you do something else? Not that I actually care and this is a real problem I'm trying to tackle or anything. Of course. Obviously. Tom _______________________________________________ MiltonKeynes-pm mailing list MiltonKeynes-pm at pm.org http://mail.pm.org/mailman/listinfo/miltonkeynes-pm From nik at ngo.org.uk Mon Dec 5 09:35:06 2005 From: nik at ngo.org.uk (Nik Clayton) Date: Mon, 5 Dec 2005 17:35:06 +0000 Subject: Managing Projects with Perl In-Reply-To: <20051201234002.GA56357@eborcom.com> References: <20051201234002.GA56357@eborcom.com> Message-ID: On 1 Dec 2005, at 23:40, Tom Hukins wrote: > Imagine this: you find yourself in a new job with no version control, > minimal documentation (mostly outdated), and no formal bug tracking > system. [...] > What would you do? Would you install trac? Would you try to tie > together RT with SVN::Web (hoping you can bribe its maintainer with > oriental buffets when required) and some sort of Wiki that doesn't > exist (I hate Kwiki: it has horrible edge cases) or would you accept > that this particular holy trinity was never meant to integrate? > > Or would you do something else? If you've got a single project, and trac does everything you want, then trac is probably the best way to go. That also assumes that either you don't want to extend it, or, if you do, you're familiar with Python. trac gets a lot of things right, but it's not designed to support multiple projects out of a single Subversion repo (although newer releases get close), and it's (deliberately) not intended to be an all-singing all-dancing system. And, of course, because everything's so tightly integrated, it's difficult to pull one piece out and replace it with another... Being of the opinion that it's better to have loosely coupled components that work well together than a single monolithic whole, SVN::Web tries to play nicely with others. That's why newer versions make it much easier for whoever's setting it up to have keywords in commit messages be turned in to links. See here, for an example: http://jc.ngo.org.uk/~nik/cgi-bin/svnweb/index.cgi/jc/revision/?rev=534 There's no reason why (for example) you couldn't have the log messages pass through a WikiText recogniser, so that WikiWords become links to your local wiki implementation too. I imagine it wouldn't be too much work to take any open source wiki and change the parser so that (to use trac syntax) source:/some/path/in/your/repo or source:/some/path/in/your/repo at 123 became links back to an SVN::Web installation. I share your distaste for Kwiki, and haven't yet found a wiki that I really like. Sitting down and writing an articulate description of the shortcomings in wikis (so I can use it as a jumping off point for writing something better, or enhancing an existing one) has been on my todo list for some time. Something that looks interesting, although I've not done anything with it beyond peruse the manual, is scmbug, http://www.mkgnu.net/?q=scmbug. The manual is at http://files.mkgnu.net/files/scmbug/doc/latest_manual/html/index.html. N -------------- next part -------------- A non-text attachment was scrubbed... Name: PGP.sig Type: application/pgp-signature Size: 186 bytes Desc: This is a digitally signed message part Url : http://mail.pm.org/pipermail/miltonkeynes-pm/attachments/20051205/a8efb223/PGP.bin From tom at eborcom.com Mon Dec 12 14:02:09 2005 From: tom at eborcom.com (Tom Hukins) Date: Mon, 12 Dec 2005 22:02:09 +0000 Subject: Meetings Message-ID: <20051212220209.GA37894@eborcom.com> Hi all, At our November meeting, we decided not to meet at the end of December as we'll all find ourselves occupied by other things. So, we will have our next scheduled meeting on Thursday 26th January. I've updated the Web page accordingly. Four of us showed up in November: we had a few interesting discussions and managed to go home at a sensible time despite the extended opening hours. Also, three of us showed up for the London Perl Workshop, which provided great value for money. Not only did we hear lots of interesting talks for free, but Fotango put some money behind the bar afterwards. In the mean time, if anyone feels like having lunch or holding an unscheduled meeting for any other reason, let us know. From past experience, I suspect a few of us will show up. Also, feel free to use this mailing list and our IRC channel for anything else that seems vaguely appropriate. I hope you all have a pleasant break over Christmas and the New Year, and I hope the lack of daylight doesn't make you feel as sleepy as I do. Tom From rtcn2 at herman.open.ac.uk Tue Dec 20 06:57:30 2005 From: rtcn2 at herman.open.ac.uk (Rod Norfor) Date: Tue, 20 Dec 2005 14:57:30 +0000 (GMT) Subject: Christmas Meetings In-Reply-To: <20051212220209.GA37894@eborcom.com> References: <20051212220209.GA37894@eborcom.com> Message-ID: Ian and I and intending to be in wetherspoons (usual one) on the 23rd from about noon onwards, If anyone would like to join us, you are most welcome. Merry Christmas and Happy New Year. Rod