From: Thomas Braun Date: Mon, 19 May 2014 19:17:58 GMT Subject: Re: Re: [PATCH v2] config: preserve config file permissions on edits Message-ID: <1400527078.32635.5.camel@thomas-debian-x64> In-Reply-To: Am Montag, den 19.05.2014, 11:59 +0200 schrieb Erik Faye-Lund: > On Mon, May 19, 2014 at 9:44 AM, Erik Faye-Lund wrote: > > On Mon, May 19, 2014 at 9:13 AM, Johannes Sixt wrote: > >> Am 5/6/2014 2:17, schrieb Eric Wong: > >>> Users may already store sensitive data such as imap.pass in > >>> ..git/config; making the file world-readable when "git config" > >>> is called to edit means their password would be compromised > >>> on a shared system. > >>> > >>> [v2: updated for section renames, as noted by Junio] > >>> > >>> Signed-off-by: Eric Wong > >>> --- > >>> config.c | 16 ++++++++++++++++ > >>> t/t1300-repo-config.sh | 10 ++++++++++ > >>> 2 files changed, 26 insertions(+) > >>> > >>> diff --git a/config.c b/config.c > >>> index a30cb5c..c227aa8 100644 > >>> --- a/config.c > >>> +++ b/config.c > >>> @@ -1636,6 +1636,13 @@ int git_config_set_multivar_in_file(const char *config_filename, > >>> MAP_PRIVATE, in_fd, 0); > >>> close(in_fd); > >>> > >>> + if (fchmod(fd, st.st_mode & 07777) < 0) { > >>> + error("fchmod on %s failed: %s", > >>> + lock->filename, strerror(errno)); > >>> + ret = CONFIG_NO_WRITE; > >>> + goto out_free; > >>> + } > >> > >> This introduces a severe failure in the Windows port because we do not > >> implement fchmod. Existing fchmod invocations do not check the return > >> value, and they are only interested in removing the write bits, and we > >> generally don't care a lot if files remain writable. > >> > >> I'm not proficient enough to add any ACL fiddling to fchmod that would be > >> required by the above change, whose purpose is to be strict about > >> permissions. Nor am I interested (who the heck is sharing a Windows box > >> anyway? ;-) > >> > >> Therefore, here's just a work-around patch to keep things going on > >> Windows. Any opinions from the Windows corner? > >> > > > > Since we use MSVCRT's chmod implementation directly which only checks > > for S_IWRITE,and Get/SetFileAttributes to simply set or clear the > > FILE_ATTRIBUTE_READONLY-flag, perhaps we could do the same except > > using Get/SetFileInformationByHandle instead? That takes us in a > > better direction, IMO. Adding full ACL support seems like a bigger > > project. > > > > If there's no objection, I'll sketch up a patch for that... > > OK, this turned out a bit more yucky than I assumed, because > SetFileInformationByHandle is only available on Windows Vista and up. > There's a static library (FileExtd.lib) that ships with MSVC that > emulates it for legacy support, I guess I should have a look at what > that does. Do we still want to fully support Windows XP? Adding a work around here for a EOL operating system sounds like a lot of effort. -- -- *** Please reply-to-all at all times *** *** (do not pretend to know who is subscribed and who is not) *** *** Please avoid top-posting. *** The msysGit Wiki is here: https://github.com/msysgit/msysgit/wiki - Github accounts are free. You received this message because you are subscribed to the Google Groups "msysGit" group. To post to this group, send email to msysgit@googlegroups.com To unsubscribe from this group, send email to msysgit+unsubscribe@googlegroups.com For more options, and view previous threads, visit this group at http://groups.google.com/group/msysgit?hl=en_US?hl=en --- You received this message because you are subscribed to the Google Groups "msysGit" group. To unsubscribe from this group and stop receiving emails from it, send an email to msysgit+unsubscribe@googlegroups.com. For more options, visit https://groups.google.com/d/optout.