Bug#407329: emacs-snapshot: emacs should make a backup for every file by default
On 2007-01-18 08:43:56 +0100, Romain Francoise wrote: Anyway, I think you overestimated the reliability of Emacs's backup system; I can see why you're surprised in this particular situation but the current defaults are tuned for the general case and everything is customizable. If you want the default settings changed, you'd better ask on emacs-devel. I've posted a message to the French newsgroup f.c.a.e to see first what users say. A good strategy to avoid losing information is to organize your backups yourself instead of counting on your tools to do it for you. (Since you seem to be working with SVN, I'd recommend making local branches with svk and commit everything in there.) However it would be easier to forget to commit to the main repository. BTW, when I know that a commit will take a long time, I usually build a patch with svn diff, which I scp to some other machine. This time I didn't. And anyway, this is not a reason to have confusing defaults. -- Vincent Lefèvre [EMAIL PROTECTED] - Web: http://www.vinc17.org/ 100% accessible validated (X)HTML - Blog: http://www.vinc17.org/blog/ Work: CR INRIA - computer arithmetic / Arenaire project (LIP, ENS-Lyon)
Bug#407329: emacs-snapshot: emacs should make a backup for every file by default
Vincent Lefevre [EMAIL PROTECTED] writes: One shouldn't lose a backup until the buffer is modified and saved. Of course the file must be saved again after being reopened for the backup to be overwritten -- Emacs doesn't throw the backup away if you just open the file... Anyway, I think you overestimated the reliability of Emacs's backup system; I can see why you're surprised in this particular situation but the current defaults are tuned for the general case and everything is customizable. If you want the default settings changed, you'd better ask on emacs-devel. A good strategy to avoid losing information is to organize your backups yourself instead of counting on your tools to do it for you. (Since you seem to be working with SVN, I'd recommend making local branches with svk and commit everything in there.) Thanks, -- ,''`. : :' :Romain Francoise [EMAIL PROTECTED] `. `' http://people.debian.org/~rfrancoise/ `- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#407329: emacs-snapshot: emacs should make a backup for every file by default
On 2007-01-17 21:18:53 +0100, Romain Francoise wrote: Backups are disabled in working copies because it's annoying to keep backup versions when the same information is recorded in the version control system... As I explained, this is not the same information, because of intermediate (non-committed) contents. Anyway, there could be an option to remove the backups once the files have been committed. The use case you describe of someone who does significant amounts of work in a working copy but doesn't have write access is pretty unusual in my experience, and in this case users can just set `vc-make-backup-files' to t. In my experience, users don't always do a commit just after editing a file. And backups are precisely important when many changes have been done on a file. I'd also like to note that relying on Emacs to backup important information is unwise -- unless you set `version-control' to t, you'll have only one backup and it gets overwritten every time you visit the file: You missed the point. Emacs does that on *every* file. So, there is no surprise concerning this behavior. This is not the case with the behavior on version-control files; there is not a single warning from Emacs. killing the buffer and reopening the file is enough to lose your only backup. This is suboptimal. One shouldn't lose a backup until the buffer is modified and saved. And before modifying, one generally has the time to notice that the data are incorrect (e.g. after a svn revert in the wrong directory...). P.S.: You might like to know that Emacs doesn't make backups of files in or under /tmp, either. I noticed that in /var/tmp too. All these particular cases are annoying. -- Vincent Lefèvre [EMAIL PROTECTED] - Web: http://www.vinc17.org/ 100% accessible validated (X)HTML - Blog: http://www.vinc17.org/blog/ Work: CR INRIA - computer arithmetic / Arenaire project (LIP, ENS-Lyon)