"standard" can now be used as a first-class keyword in preferred content expressions.

For example "standard or (include=otherdir/*)" or even "not standard"

Note that the implementation avoids any potential for loops (if a
standard preferred content expression itself mentioned standard).

This commit was sponsored by Jochen Bartl.
This commit is contained in:
Joey Hess 2014-03-14 15:04:33 -04:00
parent 7a1faf76ef
commit 3551d40b05
5 changed files with 166 additions and 149 deletions

View file

@ -56,23 +56,26 @@ parsedToMatcher parsed = case partitionEithers parsed of
([], vs) -> Right $ generate vs
(es, _) -> Left $ unwords $ map ("Parse failure: " ++) es
exprParser :: GroupMap -> M.Map UUID RemoteConfig -> Maybe UUID -> String -> [Either String (Token MatchFiles)]
exprParser groupmap configmap mu expr =
exprParser :: FileMatcher -> GroupMap -> M.Map UUID RemoteConfig -> Maybe UUID -> String -> [Either String (Token MatchFiles)]
exprParser matchstandard groupmap configmap mu expr =
map parse $ tokenizeMatcher expr
where
parse = parseToken
parse = parseToken
matchstandard
(limitPresent mu)
(limitInDir preferreddir)
groupmap
preferreddir = fromMaybe "public" $
M.lookup "preferreddir" =<< (`M.lookup` configmap) =<< mu
parseToken :: MkLimit -> MkLimit -> GroupMap -> String -> Either String (Token MatchFiles)
parseToken checkpresent checkpreferreddir groupmap t
parseToken :: FileMatcher -> MkLimit -> MkLimit -> GroupMap -> String -> Either String (Token MatchFiles)
parseToken matchstandard checkpresent checkpreferreddir groupmap t
| t `elem` tokens = Right $ token t
| t == "standard" = Right $ Operation $ \notpresent mi ->
matchMrun matchstandard $ \a -> a notpresent mi
| t == "present" = use checkpresent
| t == "inpreferreddir" = use checkpreferreddir
| t == "unused" = Right (Operation limitUnused)
| t == "unused" = Right $ Operation limitUnused
| otherwise = maybe (Left $ "near " ++ show t) use $ M.lookup k $
M.fromList
[ ("include", limitInclude)
@ -109,5 +112,5 @@ largeFilesMatcher = go =<< annexLargeFiles <$> Annex.getGitConfig
rc <- readRemoteLog
u <- getUUID
either badexpr return $
parsedToMatcher $ exprParser gm rc (Just u) expr
parsedToMatcher $ exprParser matchAll gm rc (Just u) expr
badexpr e = error $ "bad annex.largefiles configuration: " ++ e

View file

@ -67,29 +67,25 @@ preferredContentMapLoad = do
- versions of git-annex may add new features. Instead, parse errors
- result in a Matcher that will always succeed. -}
makeMatcher :: GroupMap -> M.Map UUID RemoteConfig -> UUID -> PreferredContentExpression -> FileMatcher
makeMatcher groupmap configmap u expr
| expr == "standard" = standardMatcher groupmap configmap u
| null (lefts tokens) = Utility.Matcher.generate $ rights tokens
| otherwise = matchAll
makeMatcher groupmap configmap u = go True
where
tokens = exprParser groupmap configmap (Just u) expr
{- Standard matchers are pre-defined for some groups. If none is defined,
- or a repository is in multiple groups with standard matchers, match all. -}
standardMatcher :: GroupMap -> M.Map UUID RemoteConfig -> UUID -> FileMatcher
standardMatcher groupmap configmap u =
maybe matchAll (makeMatcher groupmap configmap u . preferredContent) $
getStandardGroup =<< u `M.lookup` groupsByUUID groupmap
go expandstandard expr
| null (lefts tokens) = Utility.Matcher.generate $ rights tokens
| otherwise = matchAll
where
tokens = exprParser matchstandard groupmap configmap (Just u) expr
matchstandard
| expandstandard = maybe matchAll (go False . preferredContent) $
getStandardGroup =<< u `M.lookup` groupsByUUID groupmap
| otherwise = matchAll
{- Checks if an expression can be parsed, if not returns Just error -}
checkPreferredContentExpression :: PreferredContentExpression -> Maybe String
checkPreferredContentExpression expr
| expr == "standard" = Nothing
| otherwise = case parsedToMatcher tokens of
Left e -> Just e
Right _ -> Nothing
checkPreferredContentExpression expr = case parsedToMatcher tokens of
Left e -> Just e
Right _ -> Nothing
where
tokens = exprParser emptyGroupMap M.empty Nothing expr
tokens = exprParser matchAll emptyGroupMap M.empty Nothing expr
{- Puts a UUID in a standard group, and sets its preferred content to use
- the standard expression for that group, unless something is already set. -}

2
debian/changelog vendored
View file

@ -17,6 +17,8 @@ git-annex (5.20140307) UNRELEASED; urgency=medium
* Fix ssh connection caching stop method to work with openssh 6.5p1,
which broke the old method.
* Better workaround for problem umasks when eg, setting up ssh keys.
* "standard" can now be used as a first-class keyword in preferred content
expressions. For example "standard or (include=otherdir/*)"
-- Joey Hess <joeyh@debian.org> Thu, 06 Mar 2014 16:17:01 -0400

View file

@ -18,6 +18,20 @@ If a file matches, it's preferred to have its content stored in the
repository. If it doesn't, it's preferred to drop its content from
the repository (if there are enough copies elsewhere).
Rather than writing your own preferred content expression, you can use
several canned ones included in git-annex that are tuned to cover different
common use cases. You do this by putting a repository in a group,
and simply setting its preferred content to "standard" to match whatever
is standard for that group. See [[standard_groups]].
To check at the command line which files are matched by preferred content
settings, you can use the --want-get and --want-drop options.
For example, "git annex find --want-get --not --in ." will find all the
files that "git annex get --auto" will want to get, and "git annex find
--want-drop --in ." will find all the files that "git annex drop --auto"
will want to drop.
The expressions are very similar to the matching options documented
on the [[git-annex]] man page. At the command line, you can use those
options in commands like this:
@ -86,130 +100,17 @@ The name of the directory can be configured using
(If no directory name is configured, it uses "public" by default.)
## testing preferred content settings
### difference: "standard"
To check at the command line which files are matched by preferred content
settings, you can use the --want-get and --want-drop options.
git-annex comes with some standard preferred content expressions, that
can be used with repositories that are in some pre-defined groups,
as listed in [[standard_groups]].
For example, "git annex find --want-get --not --in ." will find all the
files that "git annex get --auto" will want to get, and "git annex find
--want-drop --in ." will find all the files that "git annex drop --auto"
will want to drop.
When a repository is in exactly one such group, you can use the "standard"
keyword in its preferred content expression, to match whatever content
the group preferrs to have. (If a repository is put into multiple standard
groups, "standard" will match anything.. so don't do that!)
## standard expressions
git-annex comes with some standard preferred content expressions, that can
be used with repositories that are in some pre-defined groups. To make a
repository use one of these, just set its preferred content expression
to "standard", and put it in one of these groups.
(Note that most of these standard expressions also make the repository
prefer any content that is only currently available on untrusted and
dead repositories. So if an untrusted repository gets connected,
any repository that can will back it up.)
### client
All content is preferred, unless it's for a file in a "archive" directory,
which has reached an archive repository, or is unused.
`(((exclude=*/archive/* and exclude=archive/*) or (not (copies=archive:1 or copies=smallarchive:1))) and not unused) or roughlylackingcopies=1`
### transfer
Use for repositories that are used to transfer data between other
repositories, but do not need to retain data themselves. For
example, a repository on a server, or in the cloud, or a small
USB drive used in a sneakernet.
The preferred content expression for these causes them to get and retain
data until all clients have a copy.
`not (inallgroup=client and copies=client:2) and ($client)`
(Where $client is a copy of the preferred content expression used for
clients.)
The "copies=client:2" part of the above handles the case where
there is only one client repository. It makes a transfer repository
speculatively prefer content in this case, even though it as of yet
has nowhere to transfer it to. Presumably, another client repository
will be added later.
### backup
All content is preferred.
`include=* or unused`
### incremental backup
Only prefers content that's not already backed up to another backup
or incremental backup repository.
`((include=* or unused) and (not copies=backup:1) and (not copies=incrementalbackup:1)) or approxlackingcopies=1`
### small archive
Only prefers content that's located in an "archive" directory, and
only if it's not already been archived somewhere else.
`((include=*/archive/* or include=archive/*) and not (copies=archive:1 or copies=smallarchive:1)) or approxlackingcopies=1`
### full archive
All content is preferred, unless it's already been archived somewhere else.
`(not (copies=archive:1 or copies=smallarchive:1)) or approxlackingcopies=1`
Note that if you want to archive multiple copies (not a bad idea!),
you should instead configure all your archive repositories with a
version of the above preferred content expression with a larger
number of copies.
### source
Use for repositories where files are often added, but that do not need to
retain files for local use. For example, a repository on a camera, where
it's desirable to remove photos as soon as they're transferred elsewhere.
The preferred content expression for these causes them to only retain
data until a copy has been sent to some other repository.
`not (copies=1)`
### manual
This gives you nearly full manual control over what content is stored in the
repository. This allows using the [[assistant]] without it trying to keep a
local copy of every file. Instead, you can manually run `git annex get`,
`git annex drop`, etc to manage content. Only content that is present
is preferred.
The exception to this manual control is that content that a client
repository would not want is not preferred. So, files in archive
directories are not preferred once their content has
reached an archive repository.
`present and ($client)`
(Where $client is a copy of the preferred content expression used for
clients.)
### public
This is used for publishing information to a repository that can be
publically accessed. Only files in a directory with a particular name
will be published. (The directory can be located anywhere in the
repository.)
The name of the directory can be configured using
`git annex enableremote $remote preferreddir=$dirname`
### unwanted
Use for repositories that you don't want to exist. This will result
in any content on them being moved away to other repositories. (Works
best when the unwanted repository is also marked as untrusted or dead.)
`exclude=*`
Most often, the whole preferred content expression is simply "standard".
But, you can do more complicated things, for example:
"`standard or include=otherdir/*`"

View file

@ -0,0 +1,115 @@
git-annex comes with some pre-defined [[preferred_content]] settings, that can
be used with repositories that are in special groups. To make a
repository use one of these, just set its preferred content expression
to "standard", and put it in one of these groups.
(Note that most of these standard expressions also make the repository
prefer any content that is only currently available on untrusted and
dead repositories. So if an untrusted repository gets connected,
any repository that can will back it up.)
### client
All content is preferred, unless it's for a file in a "archive" directory,
which has reached an archive repository, or is unused.
`(((exclude=*/archive/* and exclude=archive/*) or (not (copies=archive:1 or copies=smallarchive:1))) and not unused) or roughlylackingcopies=1`
### transfer
Use for repositories that are used to transfer data between other
repositories, but do not need to retain data themselves. For
example, a repository on a server, or in the cloud, or a small
USB drive used in a sneakernet.
The preferred content expression for these causes them to get and retain
data until all clients have a copy.
`not (inallgroup=client and copies=client:2) and ($client)`
(Where $client is a copy of the preferred content expression used for
clients.)
The "copies=client:2" part of the above handles the case where
there is only one client repository. It makes a transfer repository
speculatively prefer content in this case, even though it as of yet
has nowhere to transfer it to. Presumably, another client repository
will be added later.
### backup
All content is preferred.
`include=* or unused`
### incremental backup
Only prefers content that's not already backed up to another backup
or incremental backup repository.
`((include=* or unused) and (not copies=backup:1) and (not copies=incrementalbackup:1)) or approxlackingcopies=1`
### small archive
Only prefers content that's located in an "archive" directory, and
only if it's not already been archived somewhere else.
`((include=*/archive/* or include=archive/*) and not (copies=archive:1 or copies=smallarchive:1)) or approxlackingcopies=1`
### full archive
All content is preferred, unless it's already been archived somewhere else.
`(not (copies=archive:1 or copies=smallarchive:1)) or approxlackingcopies=1`
Note that if you want to archive multiple copies (not a bad idea!),
you should instead configure all your archive repositories with a
version of the above preferred content expression with a larger
number of copies.
### source
Use for repositories where files are often added, but that do not need to
retain files for local use. For example, a repository on a camera, where
it's desirable to remove photos as soon as they're transferred elsewhere.
The preferred content expression for these causes them to only retain
data until a copy has been sent to some other repository.
`not (copies=1)`
### manual
This gives you nearly full manual control over what content is stored in the
repository. This allows using the [[assistant]] without it trying to keep a
local copy of every file. Instead, you can manually run `git annex get`,
`git annex drop`, etc to manage content. Only content that is present
is preferred.
The exception to this manual control is that content that a client
repository would not want is not preferred. So, files in archive
directories are not preferred once their content has
reached an archive repository.
`present and ($client)`
(Where $client is a copy of the preferred content expression used for
clients.)
### public
This is used for publishing information to a repository that can be
publically accessed. Only files in a directory with a particular name
will be published. (The directory can be located anywhere in the
repository.)
The name of the directory can be configured using
`git annex enableremote $remote preferreddir=$dirname`
### unwanted
Use for repositories that you don't want to exist. This will result
in any content on them being moved away to other repositories. (Works
best when the unwanted repository is also marked as untrusted or dead.)
`exclude=*`