But it doesn’t work, the filter looks really weird when imported, I think it’s made for an older version of Duplicati.
I’m not saying these are good filters, but it’s a start, there are definetely some of them I wouldn’t like to use.
So anybody that has some good filters for windows?
Weird how? Imported how? It’s probably meant to be put in Filters three-dot menu Edit as text.
Filters in User Manual also mentions the built-in filter groups. They’re a bit old, but have you looked?
Because they’re user-sensitive, a good way to get a personalized view is GUI Commandline and run help, with no Target URL in that box, and with filter-groups in the Commandline arguments box.
Your output is pretty strange. For the 63 input lines, I get 63 entries in Edit as list, basically one for one. Your first entry with a label is the Videos one that should be at line 57, but you got it as entry 3…
In case there’s some weirdness with invisible characters, such as line endings, maybe just type a few.
are my first 4 entries which look about right for the first 4 lines what we both are showing for raw input. Indtast udtryk her seems to be Danish for Enter expression here, so did that pick up nothing?
Again, trying a bit of it at a time, manually, might be worth doing to see if the resulting view gets better.
Some of the filter builder outputs have always looked a little suspect (not the way I’d hand-code them). What’s going on here is in the opposite direction (always a bit of guesswork), and it seems way wrong.
That’s why I’m suggesting testing a little at a time to see if the reversal can do better than being empty.
Note that “System Files” here means something different than under Exclude. It is path based instead of file attribute based. These filters are adapted to the OS, but I don’t think they were specifically updated for windows 11 (the only distinction is mac/win/linux).
Since these filters are not very well described, here is the internal list for my system:
What you are looking for is probably Default excludes (everything combined), but check that there is nothing else missing from that list, as it is not very up to date.
Right now, it is not displayed in the UI which files will be excluded by these filters, but I made an update for the future. You can also just look at that list to see what you need and add it manually, then it should be shown correctly.
Hi @ts678
Thankyou for your replies, they at least highlight a big problem for me, I’m not the only one that doesn’t understand the filters
And thankyou for showing the GUI in danish, it’s not necessary, but it’s a very kind thought!
I’m on the beta (which is the fist line in the original post), and I’m not sure canary is a good idea for me
I will try and build the filters, and see what happens, thankyou again!
Thankyou for the list, that is very interesting.
Even more interesting is it, that if I copy your list, it shows the same weird formatting as the gist I found originally.
Hmm, there is really something wrong with the filter interpreter.
Entering
“Exclude file extension” → tmp or .tmp or *.tmp
Just blanks out the ‘method’
“Exclude files whose names contains” → *.tmp is accepted, but I have no idea if that works yet.
I’ve also tried entering some of the filters by hand (buest guess in method):
Below original post is a link to another topic on the weird look. It looks like Firefox 123.0 does it.
Is that what you’re using now? Did it used to look better, even on the immediately prior Firefox?
I tried this very syntax with Duplicati 2.0.7.1 beta on a Windows server and I got:
analyse de ".*\Internet Explorer\[^\\]*" - Séquence d'échappement \I non reconnue
2 interesting nuggets of information:
this is NOT a new problem of last Canary - which the ‘invisible filter’ may be.
this is NOT a Duplicati error message - I had set my install to the English language, this is a ‘native’ message generated at a lower level, the system being set to my native language (something that I can’t change as easily as for Duplicati)
I don’t have much time to investigate this further just now.
Edit: just add this:
System.ArgumentException: analyse de ".*\Internet Explorer\[^\\]*" - Séquence d'échappement \I non reconnue.
à System.Text.RegularExpressions.RegexParser.ScanCharEscape()
à System.Text.RegularExpressions.RegexParser.ScanBasicBackslash()
à System.Text.RegularExpressions.RegexParser.ScanRegex()
à System.Text.RegularExpressions.RegexParser.Parse(String re, RegexOptions op)
à System.Text.RegularExpressions.Regex..ctor(String pattern, RegexOptions options, TimeSpan matchTimeout, Boolean useCache)
à System.Text.RegularExpressions.Regex..ctor(String pattern, RegexOptions options)
à Duplicati.Library.Utility.FilterExpression.FilterEntry..ctor(String filter)
is unclear about the Firefox testing level, but 2.0.6.3_beta_2021-06-17 is mostly blank for me. 2.0.7.100_canary_2023-12-27 is too, so lack of displays seems to be related to the latest Firefox.
regexstorm.net claims to aim specifically at .NET regex (they do vary a bit) and shows the oddity.
This problem probably arises in the user’s best guess of usage not matching filter builder’s guess.
Filter builder was probably not expecting backslashes in Exclude files whose names contain because on Windows this can’t happen. There will be backslashes before, but none after a name.
That last point is probably what the right hand side of the regex is enforcing. No backslash to right.
Testing Exclude directories whose names contain and Internet Explorer looks plausible:
-*Internet Explorer*\
and seems pretty stable on different Duplicati versions and different browsers, so a different issue?
The problem with “exclude file whose name contains” is, it is only supposed to match the filename part. So it shouldn’t contain slashes in the first place, and internally it uses a regexp without escaping.