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

Re: Suggestions

From: Stefan Küng <tortoisesvn_at_gmail.com>
Date: Wed, 17 Mar 2010 21:43:46 +0100

On 17.03.2010 16:02, Damian Powell wrote:

>>> please make Overlays work consistently on Win 7 x64?
>> We're trying.
> I noticed! That's why I installed the nightly, I wanted to see if any of the
> recent changes being discussed had been implemented and if so, whether they
> made any difference for me. The 16 image limit seems to be a problem,

It's a limit of 15, not 16.

> especially when you consider that TortoiseXxx isn't the only set of apps to
> use icon overlays - lots of offsite backup tools use them also (DropBox,
> Mozy, etc...). Have you considered raising a case on the Microsoft Connect
> site to get them to increase the limit? I am sure that TortoiseXxx
> (especially TSVN) has a large enough developer base to put some real
> pressure on MS if enough people vote for it.

That won't happen. The limit is actually not in Explorer but in the list
control and image list.

>>> Or maybe have a *single* instance of TSVNCache rather than one for x86
> and another for x64.
>> Could *maybe* be done, but the x64 and win32 versions use the same struct,
> but due to the different word lengths those structs actually have different
> sizes.
> Is the communication between Explorer and TSVNCache in-proc, then? If it's
> out of proc, it would be fairly straight forward to have a fixed-width
> serialization format, no? Or perhaps it would it be feasible to use a
> Windows API-style approach of having a byte or two at the front of the
> struct to indicate what size it is?

I know how to do it. But it requires a lot of work and data conversion.

>>> How about making TSVN (or more likely TSVNCache) recognise that a WC on a
> path containing a junction is actually the same as another WC on a path
> *without* a junction. Just sayin'...
>> Tell that the svn guys (or better: the apr guys).
> I should point out that the thing I'm trying to avoid here is having the
> same subtree traversed twice for caching purposes (or four times on a 64bit
> OS). I don't understand why the windows-specific part of the tool (i.e.
> TSVN) shouldn't be the piece to do that given that it is the thing that
> makes the call to the SVN/APR API in the first place. Presumably TSVN
> watches for file changes then, if a file change is in a WC, makes
> appropriate calls to cache the status for Explorer? If that's the case, then
> surely TSVN could crawl up the tree (or remember) to see if a file is on a
> junction or not - if it is, then throw the call away given that the same

Checking every path whether it is a junction or not is completely out of
the question here. That would cause a major performance decrease (and I
mean *major*).

Also, that wouldn't help with SUBSTed drives since there's no junction.

If you use junctions or SUBSTed drives, just put one of the paths into
the ignore list.


   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 2010-03-17 21:43:49 CET

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

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.