Johannes Sixt <johannes.sixt@telecom.at> wrote:Don't need the `exec-path': git knows the right one already. So this is entirely equivalent to my suggestion, except that the user has to jump through this stupid hoop. Besides, I don't want to run `git' from inetd -- it makes distinguishing things in /etc/hosts.deny harder. Unless I write wrapper scripts or make symlinks for everything, I suppose, but doesn't it seem mad to invent wrapper scripts to distinguish between things which are already distinct but glommed together for some strange reason. Yes, of course. Silly me. It's much better that the service fail to work at all than that it have something strange like /usr/local/bin on its PATH. Besides, the only things git-daemon will actually run are git-upload-pack, upload-archive and receive-pack. * git-upload-pack in turn runs git-pack-objects, which is a builtin and therefore runs setup_path in main. Anything that does will therefore certainly have the same evil on its PATH as would have been inserted by my patch -- though I don't think it actually execs anything else. * git-upload-archive is another builtin, so the same applies; again, I don't think it actually execs anything else. * git-receive-pack is not something you enable if you care about security anyway. Besides, if there /are/ scripts and so on run by git-daemon, then they'll be hook scripts, and will also fail if the Git tools aren't on the PATH. Your argument, if I might stoop to caricature, seems to be `No, we mustn't have git-daemon set up the path itself -- it might actually /work/.' -- [mdw] - To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
| Benjamin Herrenschmidt | Re: [PATCH] Remove process freezer from suspend to RAM pathway |
| Greg KH | [GIT PATCH] driver core patches against 2.6.24 |
| Mariusz Kozlowski | [PATCH 03] drivers/sbus/char/bbc_envctrl.c: kmalloc + memset conversion to kzalloc |
| Yinghai Lu | [PATCH 02/16] x86: introduce nr_irqs for 64bit v3 |
git: | |
| Gerrit Renker | [PATCH 13/37] dccp: Deprecate Ack Ratio sysctl |
| James Morris | Re: [GIT]: Networking |
| Jeff Garzik | Re: [bug?] tg3: Failed to load firmware "tigon/tg3_tso.bin" |
