From: Junio C Hamano <junkio@cox.net>
Date: Sat, 03 Mar 2007 00:12:01 GMT
Subject: Re: [PATCH] Add core.symlinks to mark filesystems that do not support symbolic links.
Message-ID: <7vr6s7chim.fsf@assigned-by-dhcp.cox.net>
In-Reply-To: <7vvehjchsp.fsf@assigned-by-dhcp.cox.net>

Junio C Hamano <junkio@cox.net> writes:

> This code (I am the guilty one before your change above) always
> confused me.  How about doing something like this instead?
>
>     static inline unsigned int ce_mode_from_stat(struct cache_entry *ce,...
>     {
>         /*
>          * A regular file that appears on the filesystem can have
>          * a "wrong" st_mode information.  A few repository config
>          * variables can tell us to trust the mode recorded in the
>          * index more than what we get from the filesystem.
>          */
>         if (ce && S_ISREG(mode)) {
>             extern int trust_executable_bit, has_symlinks;
>
>             if (!has_symlinks && S_ISLNK(ntohl(ce->ce_mode)))
>                 return ce->ce_mode;

Oops, these three lines

>             if (!trust_executable_bit && S_ISREG(ntohl(ce->ce_mode)))
>                 return ce->ce_mode;
>             return create_ce_mode(0666);

should be:

		if (!trust...) {
                	if (S_ISREG(...))
                        	return ce->ce_mode;
			return create_ce_mode(0666);
		}

Still confused I am ;-(.