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

Re: patch: if ReadDirectoryChanges buffer was over-flooded, crawl the watched dir

From: Stefan Küng <tortoisesvn_at_gmail.com>
Date: Sun, 30 Aug 2009 07:58:32 +0200

On 30.08.2009 01:39, Gunnar Dalsnes wrote:
> patch: If there are many changes in the filesystem, the
> ReadDirectoryChanges 4KB buffer may be flooded and cleared (numBytes
> = 0). Failover to crawl the watched folder.
> My goal is to get the filesystem watcher reliable enough to make
> most\all other explicit crawls redundant:-)

Thanks for the patch. Committed in r17048.

Your goal is very nice, but I doubt that it's possible. One reason
being: there are a lot of people who share working copies (even though
that's not recommended at all, but still, they do that). For those, the
watcher simply can not work reliably: one computer is down, while on
another computer the working copy gets modified. First computer starts
up again - watcher missed all modifications done while this computer was
down -> completely wrong overlays.
And another area where the watcher isn't reliable: if the working copy
is on a network share (again, not recommended but often done by many
users), the watcher misses a lot of changes due to network load (those
data packets can get dropped under heavy net load) or if the share is
not properly configured (e.g., SAMBA has problems with those notifications).


   oo  // \\      "De Chelonian Mobile"
  (_,\/ \_/ \     TortoiseSVN
    \ \_/_\_/>    The coolest Interface to (Sub)Version Control
    /_/   \_\     http://tortoisesvn.net
To unsubscribe from this discussion, e-mail: [dev-unsubscribe_at_tortoisesvn.tigris.org].
Received on 2009-08-30 07:58:44 CEST

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