Hi Phil,

Philipp Marek wrote:
> Hello Gunnar!
> 
>> thanks for this new release. I am just evaluating the upgrade path from 
>> 1.1.16
>> to 1.2.1. Some questions about the handling:
>>
>> On playing with the examples of grouping patterns i noticed a small typo:
>> http://doc.fsvs-software.org/doxygen-gif/index.html
> Yes, thank you. I'll fix that.
> 
>> Fix is to put a trailing comma after each modifier. Seems this gets default 
>> now.
> There's a conversion script for the ignore/grouping patterns; you don't have 
> to do that
> by hand.
> 
> 
>> Regarding introduction of groups and modifiers. Is it now possible to 
>> explicitly
>> discard check of the mtime of directories? I use fsvs in a cron job for
>> monitoring and I get notifications when someone with owner privileges open 
>> files
>> within directories, because this causes an update of directories mtime 
>> timestamp.

> Well, the open itself shouldn't modify the mtime of the directory ... but the 
> creation
> of a temporary file will.

Ah, so this happens when I use vim and a swapfile is created. I missed that 
fact.

> 
> You could try using the filter option "text,owner,mode,group,new,deleted" ... 
> that's
> simply everything but mtime.
> You won't see files with *only* mtime changed, too - but that might be ok, as 
> they
> should be given as soon as the data changes.

So ignoring mtime globally is my only option, at least I know its possible. Do
you think this can become a match pattern so that it can be used together with
"dironly" in a global scope affecting all dirs?

> 
> 
>> Regarding the last example:
>>
>>>      m:04:0
>>>      t,./etc/
>>>      ./**
>> Do I get the same behaviour with this?:
>>
>>>      t,m:004:004,./etc/
>>>      ./**
>>From a quick look, yes. But I didn't test.
> 
> 
>> I like the idea to keep the old WAA file hierarchy for now and leave the new 
>> for
>> advanced users with the configure option.
>> Regarding the transition of files to sqlite do you plan to also integrate the
>> _base symlinks?
> Maybe ...
> The symlinks are there just for informational purposes, and so they might 
> stay.
> If there's only a sqlite database, it's not that easy to see which working 
> copy this is.
> Maybe I should put a "README" or something like that in there? But that's not 
> so easily
> used by scripts ...

Just asking because we had problems with the _base symlink when debian packages
take use of the "find" command. This still happens with "base-files" causing an
endless loop.
> 
>> I get a segfault with "fsvs unversion", bugreport is on its way...
> Fine ;-/, please tell me.

This happens when using the "fsvs add" and performing unversion. I have to
figure out the exact scenario.

> 
>> I will then continue now with checking the functionality against ssl secured
>> repository...
> I'll be happy to hear any results!
> 
> 
> Regards,
> 
> Phil
> 
> 

Rgds,
Gunnar

------------------------------------------------------
http://fsvs.tigris.org/ds/viewMessage.do?dsForumId=3928&dsMessageId=2414354

To unsubscribe from this discussion, e-mail: 
[[email protected]].

Attachment: signature.asc
Description: PGP signature

Reply via email to