Comments on: You had me at delete. http://www.donniepinkston.net/2006-05-02/svn-delete.html (there are many more like it, but this one is mine) Tue, 06 Jan 2009 19:41:09 +0000 http://wordpress.org/?v=2.5.1 By: John http://www.donniepinkston.net/2006-05-02/svn-delete.html#comment-13431 John Mon, 25 Feb 2008 06:31:18 +0000 http://www.donniepinkston.net/?p=68#comment-13431 The lack of globbing for delete sure seems like a nuisance, but xargs to the rescue... Eg: to remove all the (uppercase) jpegs... ls *.JPG | xargs svn rm The lack of globbing for delete sure seems like a nuisance, but xargs to the rescue…

Eg: to remove all the (uppercase) jpegs…

ls *.JPG | xargs svn rm

]]> By: Ben http://www.donniepinkston.net/2006-05-02/svn-delete.html#comment-2196 Ben Sat, 16 Sep 2006 03:56:15 +0000 http://www.donniepinkston.net/?p=68#comment-2196 I'm glad you got this "straightened out" in your head, because cheap copies are pretty much as good as it gets unless you want to have the SCM start to understand the content it maintains. That, or you change the data metaphor to something besides files and directories... or, I guess, both. You can look at Arch and BitKeeper and a few of the alternatives to see the realm of possibility, but essentially the ultimate pain -- merge -- does not magically go away. Maybe some information theorist can give us something analagous to a Shannon limit for maximal merge simplicity. :) I’m glad you got this “straightened out” in your head, because cheap copies are pretty much as good as it gets unless you want to have the SCM start to understand the content it maintains. That, or you change the data metaphor to something besides files and directories… or, I guess, both.

You can look at Arch and BitKeeper and a few of the alternatives to see the realm of possibility, but essentially the ultimate pain — merge — does not magically go away. Maybe some information theorist can give us something analagous to a Shannon limit for maximal merge simplicity. :)

]]> By: donnie http://www.donniepinkston.net/2006-05-02/svn-delete.html#comment-173 donnie Thu, 11 May 2006 18:03:31 +0000 http://www.donniepinkston.net/?p=68#comment-173 No I completely agree with you - I am used to the CVS way of doing things, and I'm sure that biases me to some extent. In fact, Subversion is at a disadvantage because somebody just <em>told</em> me these things about CVS, but I had to look up the Subversion stuff myself. (Already "more work.") I feel like Subversion might have missed the <em>real</em> opportunity to contribute something revolutionary to the whole space, by really presenting a clean, powerful abstraction for branching, tagging, and merging between branches. That is always one of the most difficult problems in SCM, but most people don't realize it because only a relatively few people grapple with it, and only on projects with an actual "maintenance phase" to the cycle. I want to rant about that too, but I want to think about how it "should be done" before I go off half-cocked... ;-) No I completely agree with you - I am used to the CVS way of doing things, and I’m sure that biases me to some extent. In fact, Subversion is at a disadvantage because somebody just told me these things about CVS, but I had to look up the Subversion stuff myself. (Already “more work.”)

I feel like Subversion might have missed the real opportunity to contribute something revolutionary to the whole space, by really presenting a clean, powerful abstraction for branching, tagging, and merging between branches. That is always one of the most difficult problems in SCM, but most people don’t realize it because only a relatively few people grapple with it, and only on projects with an actual “maintenance phase” to the cycle. I want to rant about that too, but I want to think about how it “should be done” before I go off half-cocked… ;-)

]]> By: Doug http://www.donniepinkston.net/2006-05-02/svn-delete.html#comment-161 Doug Sat, 06 May 2006 06:49:29 +0000 http://www.donniepinkston.net/?p=68#comment-161 Donnie, if you think about it, there is nothing really "simple" about all the CVSisms you are used to -- it's just that you are used to them :) For example, there is nothing simple about .cvsignore. It probably takes the same effort to track down and use .cvsignore as it does to use svn propset. The end result is the same -- as soon as you develop new habits, they are equally easy. As for tagging being a copy of the repository in subversion, that is kind of goofy, I have to admit. And the imposition of the directory structure sucked too. But after getting used to it, it is transparent. To do work on a branch or tag, you have to check out using special flags or switch your sandbox using a special command I'd have to look up to remember it. In subversion, I change to a different directory. It's all half a dozen of one, 6 of the other in the end. Donnie, if you think about it, there is nothing really “simple” about all the CVSisms you are used to — it’s just that you are used to them :) For example, there is nothing simple about .cvsignore. It probably takes the same effort to track down and use .cvsignore as it does to use svn propset. The end result is the same — as soon as you develop new habits, they are equally easy.

As for tagging being a copy of the repository in subversion, that is kind of goofy, I have to admit. And the imposition of the directory structure sucked too. But after getting used to it, it is transparent. To do work on a branch or tag, you have to check out using special flags or switch your sandbox using a special command I’d have to look up to remember it. In subversion, I change to a different directory. It’s all half a dozen of one, 6 of the other in the end.

]]>