From: Stephen D. <sd...@gm...> - 2008-11-19 17:48:16
|
On Wed, Aug 20, 2008 at 7:16 AM, Vasiljevic Zoran <zv...@ar...> wrote: > > On 20.08.2008, at 00:19, Stephen Deasey wrote: > >> I think it's probably better to stay with the familiar cvs. It doesn't >> scare anyone, SF handles everything, and patches are probably too >> infrequent to for the vcs to make much difference anyway. >> >> Agree? > > Yes. Lets keep it simple for now. > I've changed my mind about this again. It's just too much of a pain in the ass to do anything more than minor fixes without trampling on other peoples stuff, using CVS. Here's something new though: http://www.bitbucket.org/naviserver/naviserver/overview/ Fancy hosting for mercurial repos. It's like github (if you're familiar with that, very popular), but for mercurial. It's up to date for both branches of AOLserver, and naviserver with the exception of Zoran's commit the other day (I'll update that now). It also models the fact that naviserver is a fork AOLserver 4.0.10. This would have all the advantages we talked about previously, plus a couple of new ones which would be particularly helpful when developing larger changes which need a bit of back and forth development before being committed: branches a patch queues. If you look at the interface you'll see a 'fork' link. The idea is any random stranger can fork any project to make changes. They then send a pull request to you and merge the changes back, and they delete their fork. Bitbucket also supports mercurial queues, which is almost the same but a bit easier when reworking a set of patches. So the ideas is that we'd use bitbucket for hosting the mercurial repos and keep everything else as is. How does this sound? (Tell me you username on bitbucket and I'll add you to the project). |
From: Andrew P. <at...@pi...> - 2008-11-19 19:21:23
|
On Wed, Nov 19, 2008 at 05:48:11PM +0000, Stephen Deasey wrote: > http://www.bitbucket.org/naviserver/naviserver/overview/ > It's up to date for both branches of AOLserver, and naviserver with > the exception of Zoran's commit the other day (I'll update that > now). It also models the fact that naviserver is a fork AOLserver > 4.0.10. I've yet to really use Mercurial at all, but that all sounds very cool. I take it that one way CVS to Mercurial importing is very robust by now? -- Andrew Piskorski <at...@pi...> http://www.piskorski.com/ |
From: Stephen D. <sd...@gm...> - 2008-11-19 20:03:55
|
On Wed, Nov 19, 2008 at 6:54 PM, Andrew Piskorski <at...@pi...> wrote: > On Wed, Nov 19, 2008 at 05:48:11PM +0000, Stephen Deasey wrote: > >> http://www.bitbucket.org/naviserver/naviserver/overview/ > >> It's up to date for both branches of AOLserver, and naviserver with >> the exception of Zoran's commit the other day (I'll update that >> now). It also models the fact that naviserver is a fork AOLserver >> 4.0.10. > > I've yet to really use Mercurial at all, but that all sounds very > cool. > > I take it that one way CVS to Mercurial importing is very robust by > now? It is. Here's the command I use: $ hg convert -A ~/authors.map \ --config convert.hg.usebranchnames=0 --config convert.hg.clonebranches=1 \ ~/in/naviserver-HEAD ns-HEAD-hg Where authors.map is a file which maps sf username like 'sdeasey' to mercurial users like 'Stephen Deasey <sd...@gm...>', and the branch stuff is saying I prefer to represent branches as separate repos. However, I have done a bunch of clean ups of the commit messages, some squashing of commits which should have been one commit, or where a followup commit fixed a typo or some other trivial mistake with a log message of 'oops...'. So the conversion is 'accurate', but better than the original :-) Even if you're not interested in mercurial, you might want to browse the aolserver repos here for this reason alone: http://www.bitbucket.org/aolserver/aolserver/overview/ http://www.bitbucket.org/aolserver/aolserver-40x/overview/ For example, according to the AOLserver ChangeLog the last change was made in June, and it's one of only two changes made this year: http://aolserver.cvs.sourceforge.net/viewvc/aolserver/aolserver/ChangeLog?revision=1.386&view=markup Browsing the mercurial shortlog on bitbucket you can see there's actually been 9 changes by 4 different authors. This would be useful for these guys: http://openacs.org/forums/message-view?message_id=2438982 to whom it looks, on the outside, like nothing's been touched in either aolserver or naviserver for years... |
From: Vlad S. <vl...@cr...> - 2008-11-20 18:18:24
|
I frankly don't care about any particular system, it just should be only one, fast and accessible all the time. Stephen Deasey wrote: > On Wed, Aug 20, 2008 at 7:16 AM, Vasiljevic Zoran <zv...@ar...> wrote: >> On 20.08.2008, at 00:19, Stephen Deasey wrote: >> >>> I think it's probably better to stay with the familiar cvs. It doesn't >>> scare anyone, SF handles everything, and patches are probably too >>> infrequent to for the vcs to make much difference anyway. >>> >>> Agree? >> Yes. Lets keep it simple for now. >> > > > I've changed my mind about this again. It's just too much of a pain in > the ass to do anything more than minor fixes without trampling on > other peoples stuff, using CVS. > > Here's something new though: > > http://www.bitbucket.org/naviserver/naviserver/overview/ > > Fancy hosting for mercurial repos. It's like github (if you're > familiar with that, very popular), but for mercurial. It's up to date > for both branches of AOLserver, and naviserver with the exception of > Zoran's commit the other day (I'll update that now). It also models > the fact that naviserver is a fork AOLserver 4.0.10. > > This would have all the advantages we talked about previously, plus a > couple of new ones which would be particularly helpful when developing > larger changes which need a bit of back and forth development before > being committed: branches a patch queues. > > If you look at the interface you'll see a 'fork' link. The idea is any > random stranger can fork any project to make changes. They then send a > pull request to you and merge the changes back, and they delete their > fork. Bitbucket also supports mercurial queues, which is almost the > same but a bit easier when reworking a set of patches. > > So the ideas is that we'd use bitbucket for hosting the mercurial > repos and keep everything else as is. > > How does this sound? > > > (Tell me you username on bitbucket and I'll add you to the project). > > ------------------------------------------------------------------------- > This SF.Net email is sponsored by the Moblin Your Move Developer's challenge > Build the coolest Linux based applications with Moblin SDK & win great prizes > Grand prize is a trip for two to an Open Source event anywhere in the world > http://moblin-contest.org/redirect.php?banner_id=100&url=/ > _______________________________________________ > naviserver-devel mailing list > nav...@li... > https://lists.sourceforge.net/lists/listinfo/naviserver-devel > |
From: Vlad S. <vl...@cr...> - 2008-11-20 18:25:11
|
vseryakov is my username, looks oike adding openssh pub key does not work, it keeps saying SSH key is not valid Stephen Deasey wrote: > On Wed, Aug 20, 2008 at 7:16 AM, Vasiljevic Zoran <zv...@ar...> wrote: >> On 20.08.2008, at 00:19, Stephen Deasey wrote: >> >>> I think it's probably better to stay with the familiar cvs. It doesn't >>> scare anyone, SF handles everything, and patches are probably too >>> infrequent to for the vcs to make much difference anyway. >>> >>> Agree? >> Yes. Lets keep it simple for now. >> > > > I've changed my mind about this again. It's just too much of a pain in > the ass to do anything more than minor fixes without trampling on > other peoples stuff, using CVS. > > Here's something new though: > > http://www.bitbucket.org/naviserver/naviserver/overview/ > > Fancy hosting for mercurial repos. It's like github (if you're > familiar with that, very popular), but for mercurial. It's up to date > for both branches of AOLserver, and naviserver with the exception of > Zoran's commit the other day (I'll update that now). It also models > the fact that naviserver is a fork AOLserver 4.0.10. > > This would have all the advantages we talked about previously, plus a > couple of new ones which would be particularly helpful when developing > larger changes which need a bit of back and forth development before > being committed: branches a patch queues. > > If you look at the interface you'll see a 'fork' link. The idea is any > random stranger can fork any project to make changes. They then send a > pull request to you and merge the changes back, and they delete their > fork. Bitbucket also supports mercurial queues, which is almost the > same but a bit easier when reworking a set of patches. > > So the ideas is that we'd use bitbucket for hosting the mercurial > repos and keep everything else as is. > > How does this sound? > > > (Tell me you username on bitbucket and I'll add you to the project). > > ------------------------------------------------------------------------- > This SF.Net email is sponsored by the Moblin Your Move Developer's challenge > Build the coolest Linux based applications with Moblin SDK & win great prizes > Grand prize is a trip for two to an Open Source event anywhere in the world > http://moblin-contest.org/redirect.php?banner_id=100&url=/ > _______________________________________________ > naviserver-devel mailing list > nav...@li... > https://lists.sourceforge.net/lists/listinfo/naviserver-devel > |
From: Stephen D. <sd...@gm...> - 2008-11-20 19:33:45
|
On 11/20/08, Vlad Seryakov <vl...@cr...> wrote: > vseryakov is my username, OK. I've given you write access. > looks oike adding openssh pub key does not work, it keeps saying SSH key is not valid Perhaps during cut 'n paste a stray \n crept in? Worked for me. Anyway, you don't need to use ssh, ssl works just as well (http is mercurial's native transport): $ hg clone https://vse...@bi.../naviserver/naviserver/ ~/in/naviserver-hg btw. it seems that the bitbucket site is set up such that all repos belong to a person, so for the naviserver repo I created a 'fake' person 'naviserver' to own it. I'll send you and Zoran the password, in case I'm run over by a walrus or something. You won't need to use it on a day to day basis. It's just for adding more people to the commit list and changing the details on the website. Also, make sure you have this in your ~/.hgrc [ui] username=Vlad Seryakov <vse...@cr...> so that it matches all the other log entries, otherwise you'll be vlad@localhost or whatever. And for log messages note that the convention is to have a single line, like an email subject, then a blank line, then any extra explanation, in normal paragraphs. Word wrap to 70-80 characters. Use active voice, 'fix' rather than 'fixed'. No full stop. if it's a module, prefix with 'nsperm: ' or whatever. Check what's there already, you'll get the idea. And because commit and push are separate, you can check and redo as many times as you like before anything is public. If you find yourself trying to describe two things, they should have been separate commits. Mercurial makes this easy. You can commit multiple times, and then at the end push to the public repo. And clean your finger nails, and brush your teeth, and say your prayers... |
From: Vasiljevic Z. <zv...@ar...> - 2008-11-20 19:40:50
|
On 20.11.2008, at 20:33, Stephen Deasey wrote: > > And clean your finger nails, and brush your teeth, and say your > prayers... Eh... brave new world... So yet another thing to learn... I guess Internet is full of docs about this and I will have fun reading it. Cheers, Zoran |
From: Stephen D. <sd...@gm...> - 2008-11-20 19:49:08
|
On 11/20/08, Vasiljevic Zoran <zv...@ar...> wrote: > > On 20.11.2008, at 20:33, Stephen Deasey wrote: > > > > > And clean your finger nails, and brush your teeth, and say your > > prayers... > > > Eh... brave new world... > > So yet another thing to learn... I guess Internet is full of docs > about this and I will have fun reading it. Hopefully, it will take you about 2 minutes to get the basics, which is the advantage of mercurial over git, an otherwise great system. Clone the repository first. You need the bits on your hard drive: $ hg clone https://zo...@bi.../naviserver/naviserver/ $ cd naviserver ... hack hack hack... $ hg commit $ hg push Now obviously there are bells n' whistles, but the basics are easy enough that it shouldn't prevent you from getting things done. (The built-in help is excellent: hg help, hg pull --help, etc.) |