[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Timestamp weirdness

From: Ben Collins-Sussman <sussman_at_collab.net>
Date: 2001-06-27 22:43:19 CEST

Branko =?ISO-8859-2?Q?=C8ibej?= <brane@xbc.nu> writes:

> - text-time="Thu 28 Jun 2001 01:48:46.-1073753252 (day 179, dst 1, gmt_off 007200)"
> + text-time="Wed 27 Jun 2001 23:59:43.-1073744308 (day 178, dst 1, gmt_off 007200)"
>
> Everything looks fine, except the differences in the text-time fields in
> SVN/entries. Neither of these values is correct; I'm typing this at
> 22:21 GMT+2, about half an hour after I ran the test.

You know, I've noticed this oddness for a *long* time and have always
wondered. It seems like the timestamps are always bizarre.

Of course, the bug is (currently) harmless; the timestamps exist
merely as a shortcut to determine if a file has been locally
modified. But as it stands, having a bogus timestamp only removes
this shortcut. Here's the logic behind svn_wc_text_modified_p():

     if (text-base-size != working-file-size)
          return MODIFIED;
     else if (working-file-timestamp == entries-timestamp)
          return NOT_MODIFIED;
     else
          do byte-for-byte comparison of working-file and text-base

So, as you can see, this bug is simply causing a lot of needless
"brute force" testing of file's contents.

Now -- *why* are these timestamps bogus? I have no idea. I wonder if
the APR time interfaces have changed since last October, and we just
didn't notice the breakage.

Either way, please file a bug in Issuezilla!

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:36:32 2006

This is an archived mail posted to the Subversion Dev mailing list.