Rewrite reference handling to be abstract and accurate
This commit actually does three major changes to the way references
are handled within JGit. Unfortunately they were easier to do as
a single massive commit than to break them up into smaller units.
Disambiguate symbolic references:
---------------------------------
Reporting a symbolic reference such as HEAD as though it were
any other normal reference like refs/heads/master causes subtle
programming errors. We have been bitten by this error on several
occasions, as have some downstream applications written by myself.
Instead of reporting HEAD as a reference whose name differs from
its "original name", report it as an actual SymbolicRef object
that the application can test the type and examine the target of.
With this change, Ref is now an abstract type with different
subclasses for the different types.
In the classical example of "HEAD" being a symbolic reference to
branch "refs/heads/master", the Repository.getAllRefs() method
will now return:
Map<String, Ref> all = repository.getAllRefs();
SymbolicRef HEAD = (SymbolicRef) all.get("HEAD");
ObjectIdRef master = (ObjectIdRef) all.get("refs/heads/master");
assertSame(master, HEAD.getTarget());
assertSame(master.getObjectId(), HEAD.getObjectId());
assertEquals("HEAD", HEAD.getName());
assertEquals("refs/heads/master", master.getName());
A nice side-effect of this change is the storage type of the
symbolic reference is no longer ambiguous with the storge type
of the underlying reference it targets. In the above example,
if master was only available in the packed-refs file, then the
following is also true:
assertSame(Ref.Storage.LOOSE, HEAD.getStorage());
assertSame(Ref.Storage.PACKED, master.getStorage());
(Prior to this change we returned the ambiguous storage of
LOOSE_PACKED for HEAD, which was confusing since it wasn't
actually true on disk).
Another nice side-effect of this change is all intermediate
symbolic references are preserved, and are therefore visible
to the application when they walk the target chain. We can
now correctly inspect chains of symbolic references.
As a result of this change the Ref.getOrigName() method has been
removed from the API. Applications should identify a symbolic
reference by testing for isSymbolic() and not by using an arcane
string comparsion between properties.
Abstract the RefDatabase storage:
---------------------------------
RefDatabase is now abstract, similar to ObjectDatabase, and a
new concrete implementation called RefDirectory is used for the
traditional on-disk storage layout. In the future we plan to
support additional implementations, such as a pure in-memory
RefDatabase for unit testing purposes.
Optimize RefDirectory:
----------------------
The implementation of the in-memory reference cache, reading, and
update routines has been completely rewritten. Much of the code
was heavily borrowed or cribbed from the prior implementation,
so copyright notices have been left intact as much as possible.
The RefDirectory cache no longer confuses symbolic references
with normal references. This permits the cache to resolve the
value of a symbolic reference as late as possible, ensuring it
is always current, without needing to maintain reverse pointers.
The cache is now 2 sorted RefLists, rather than 3 HashMaps.
Using sorted lists allows the implementation to reduce the
in-memory footprint when storing many refs. Using specialized
types for the elements allows the code to avoid additional map
lookups for auxiliary stat information.
To improve scan time during getRefs(), the lists are returned via
a copy-on-write contract. Most callers of getRefs() do not modify
the returned collections, so the copy-on-write semantics improves
access on repositories with a large number of packed references.
Iterator traversals of the returned Map<String,Ref> are performed
using a simple merge-join of the two cache lists, ensuring we can
perform the entire traversal in linear time as a function of the
number of references: O(PackedRefs + LooseRefs).
Scans of the loose reference space to update the cache run in
O(LooseRefs log LooseRefs) time, as the directory contents
are sorted before being merged against the in-memory cache.
Since the majority of stable references are kept packed, there
typically are only a handful of reference names to be sorted,
so the sorting cost should not be very high.
Locking is reduced during getRefs() by taking advantage of the
copy-on-write semantics of the improved cache data structure.
This permits concurrent readers to pull back references without
blocking each other. If there is contention updating the cache
during a scan, one or more updates are simply skipped and will
get picked up again in a future scan.
Writing to the $GIT_DIR/packed-refs during reference delete is
now fully atomic. The file is locked, reparsed fresh, and written
back out if a change is necessary. This avoids all race conditions
with concurrent external updates of the packed-refs file.
The RefLogWriter class has been fully folded into RefDirectory
and is therefore deleted. Maintaining the reference's log is
the responsiblity of the database implementation, and not all
implementations will use java.io for access.
Future work still remains to be done to abstract the ReflogReader
class away from local disk IO.
Change-Id: I26b9287c45a4b2d2be35ba2849daa316f5eec85d
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
14 年前 Rewrite reference handling to be abstract and accurate
This commit actually does three major changes to the way references
are handled within JGit. Unfortunately they were easier to do as
a single massive commit than to break them up into smaller units.
Disambiguate symbolic references:
---------------------------------
Reporting a symbolic reference such as HEAD as though it were
any other normal reference like refs/heads/master causes subtle
programming errors. We have been bitten by this error on several
occasions, as have some downstream applications written by myself.
Instead of reporting HEAD as a reference whose name differs from
its "original name", report it as an actual SymbolicRef object
that the application can test the type and examine the target of.
With this change, Ref is now an abstract type with different
subclasses for the different types.
In the classical example of "HEAD" being a symbolic reference to
branch "refs/heads/master", the Repository.getAllRefs() method
will now return:
Map<String, Ref> all = repository.getAllRefs();
SymbolicRef HEAD = (SymbolicRef) all.get("HEAD");
ObjectIdRef master = (ObjectIdRef) all.get("refs/heads/master");
assertSame(master, HEAD.getTarget());
assertSame(master.getObjectId(), HEAD.getObjectId());
assertEquals("HEAD", HEAD.getName());
assertEquals("refs/heads/master", master.getName());
A nice side-effect of this change is the storage type of the
symbolic reference is no longer ambiguous with the storge type
of the underlying reference it targets. In the above example,
if master was only available in the packed-refs file, then the
following is also true:
assertSame(Ref.Storage.LOOSE, HEAD.getStorage());
assertSame(Ref.Storage.PACKED, master.getStorage());
(Prior to this change we returned the ambiguous storage of
LOOSE_PACKED for HEAD, which was confusing since it wasn't
actually true on disk).
Another nice side-effect of this change is all intermediate
symbolic references are preserved, and are therefore visible
to the application when they walk the target chain. We can
now correctly inspect chains of symbolic references.
As a result of this change the Ref.getOrigName() method has been
removed from the API. Applications should identify a symbolic
reference by testing for isSymbolic() and not by using an arcane
string comparsion between properties.
Abstract the RefDatabase storage:
---------------------------------
RefDatabase is now abstract, similar to ObjectDatabase, and a
new concrete implementation called RefDirectory is used for the
traditional on-disk storage layout. In the future we plan to
support additional implementations, such as a pure in-memory
RefDatabase for unit testing purposes.
Optimize RefDirectory:
----------------------
The implementation of the in-memory reference cache, reading, and
update routines has been completely rewritten. Much of the code
was heavily borrowed or cribbed from the prior implementation,
so copyright notices have been left intact as much as possible.
The RefDirectory cache no longer confuses symbolic references
with normal references. This permits the cache to resolve the
value of a symbolic reference as late as possible, ensuring it
is always current, without needing to maintain reverse pointers.
The cache is now 2 sorted RefLists, rather than 3 HashMaps.
Using sorted lists allows the implementation to reduce the
in-memory footprint when storing many refs. Using specialized
types for the elements allows the code to avoid additional map
lookups for auxiliary stat information.
To improve scan time during getRefs(), the lists are returned via
a copy-on-write contract. Most callers of getRefs() do not modify
the returned collections, so the copy-on-write semantics improves
access on repositories with a large number of packed references.
Iterator traversals of the returned Map<String,Ref> are performed
using a simple merge-join of the two cache lists, ensuring we can
perform the entire traversal in linear time as a function of the
number of references: O(PackedRefs + LooseRefs).
Scans of the loose reference space to update the cache run in
O(LooseRefs log LooseRefs) time, as the directory contents
are sorted before being merged against the in-memory cache.
Since the majority of stable references are kept packed, there
typically are only a handful of reference names to be sorted,
so the sorting cost should not be very high.
Locking is reduced during getRefs() by taking advantage of the
copy-on-write semantics of the improved cache data structure.
This permits concurrent readers to pull back references without
blocking each other. If there is contention updating the cache
during a scan, one or more updates are simply skipped and will
get picked up again in a future scan.
Writing to the $GIT_DIR/packed-refs during reference delete is
now fully atomic. The file is locked, reparsed fresh, and written
back out if a change is necessary. This avoids all race conditions
with concurrent external updates of the packed-refs file.
The RefLogWriter class has been fully folded into RefDirectory
and is therefore deleted. Maintaining the reference's log is
the responsiblity of the database implementation, and not all
implementations will use java.io for access.
Future work still remains to be done to abstract the ReflogReader
class away from local disk IO.
Change-Id: I26b9287c45a4b2d2be35ba2849daa316f5eec85d
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
14 年前 Rewrite reference handling to be abstract and accurate
This commit actually does three major changes to the way references
are handled within JGit. Unfortunately they were easier to do as
a single massive commit than to break them up into smaller units.
Disambiguate symbolic references:
---------------------------------
Reporting a symbolic reference such as HEAD as though it were
any other normal reference like refs/heads/master causes subtle
programming errors. We have been bitten by this error on several
occasions, as have some downstream applications written by myself.
Instead of reporting HEAD as a reference whose name differs from
its "original name", report it as an actual SymbolicRef object
that the application can test the type and examine the target of.
With this change, Ref is now an abstract type with different
subclasses for the different types.
In the classical example of "HEAD" being a symbolic reference to
branch "refs/heads/master", the Repository.getAllRefs() method
will now return:
Map<String, Ref> all = repository.getAllRefs();
SymbolicRef HEAD = (SymbolicRef) all.get("HEAD");
ObjectIdRef master = (ObjectIdRef) all.get("refs/heads/master");
assertSame(master, HEAD.getTarget());
assertSame(master.getObjectId(), HEAD.getObjectId());
assertEquals("HEAD", HEAD.getName());
assertEquals("refs/heads/master", master.getName());
A nice side-effect of this change is the storage type of the
symbolic reference is no longer ambiguous with the storge type
of the underlying reference it targets. In the above example,
if master was only available in the packed-refs file, then the
following is also true:
assertSame(Ref.Storage.LOOSE, HEAD.getStorage());
assertSame(Ref.Storage.PACKED, master.getStorage());
(Prior to this change we returned the ambiguous storage of
LOOSE_PACKED for HEAD, which was confusing since it wasn't
actually true on disk).
Another nice side-effect of this change is all intermediate
symbolic references are preserved, and are therefore visible
to the application when they walk the target chain. We can
now correctly inspect chains of symbolic references.
As a result of this change the Ref.getOrigName() method has been
removed from the API. Applications should identify a symbolic
reference by testing for isSymbolic() and not by using an arcane
string comparsion between properties.
Abstract the RefDatabase storage:
---------------------------------
RefDatabase is now abstract, similar to ObjectDatabase, and a
new concrete implementation called RefDirectory is used for the
traditional on-disk storage layout. In the future we plan to
support additional implementations, such as a pure in-memory
RefDatabase for unit testing purposes.
Optimize RefDirectory:
----------------------
The implementation of the in-memory reference cache, reading, and
update routines has been completely rewritten. Much of the code
was heavily borrowed or cribbed from the prior implementation,
so copyright notices have been left intact as much as possible.
The RefDirectory cache no longer confuses symbolic references
with normal references. This permits the cache to resolve the
value of a symbolic reference as late as possible, ensuring it
is always current, without needing to maintain reverse pointers.
The cache is now 2 sorted RefLists, rather than 3 HashMaps.
Using sorted lists allows the implementation to reduce the
in-memory footprint when storing many refs. Using specialized
types for the elements allows the code to avoid additional map
lookups for auxiliary stat information.
To improve scan time during getRefs(), the lists are returned via
a copy-on-write contract. Most callers of getRefs() do not modify
the returned collections, so the copy-on-write semantics improves
access on repositories with a large number of packed references.
Iterator traversals of the returned Map<String,Ref> are performed
using a simple merge-join of the two cache lists, ensuring we can
perform the entire traversal in linear time as a function of the
number of references: O(PackedRefs + LooseRefs).
Scans of the loose reference space to update the cache run in
O(LooseRefs log LooseRefs) time, as the directory contents
are sorted before being merged against the in-memory cache.
Since the majority of stable references are kept packed, there
typically are only a handful of reference names to be sorted,
so the sorting cost should not be very high.
Locking is reduced during getRefs() by taking advantage of the
copy-on-write semantics of the improved cache data structure.
This permits concurrent readers to pull back references without
blocking each other. If there is contention updating the cache
during a scan, one or more updates are simply skipped and will
get picked up again in a future scan.
Writing to the $GIT_DIR/packed-refs during reference delete is
now fully atomic. The file is locked, reparsed fresh, and written
back out if a change is necessary. This avoids all race conditions
with concurrent external updates of the packed-refs file.
The RefLogWriter class has been fully folded into RefDirectory
and is therefore deleted. Maintaining the reference's log is
the responsiblity of the database implementation, and not all
implementations will use java.io for access.
Future work still remains to be done to abstract the ReflogReader
class away from local disk IO.
Change-Id: I26b9287c45a4b2d2be35ba2849daa316f5eec85d
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
14 年前 Rewrite reference handling to be abstract and accurate
This commit actually does three major changes to the way references
are handled within JGit. Unfortunately they were easier to do as
a single massive commit than to break them up into smaller units.
Disambiguate symbolic references:
---------------------------------
Reporting a symbolic reference such as HEAD as though it were
any other normal reference like refs/heads/master causes subtle
programming errors. We have been bitten by this error on several
occasions, as have some downstream applications written by myself.
Instead of reporting HEAD as a reference whose name differs from
its "original name", report it as an actual SymbolicRef object
that the application can test the type and examine the target of.
With this change, Ref is now an abstract type with different
subclasses for the different types.
In the classical example of "HEAD" being a symbolic reference to
branch "refs/heads/master", the Repository.getAllRefs() method
will now return:
Map<String, Ref> all = repository.getAllRefs();
SymbolicRef HEAD = (SymbolicRef) all.get("HEAD");
ObjectIdRef master = (ObjectIdRef) all.get("refs/heads/master");
assertSame(master, HEAD.getTarget());
assertSame(master.getObjectId(), HEAD.getObjectId());
assertEquals("HEAD", HEAD.getName());
assertEquals("refs/heads/master", master.getName());
A nice side-effect of this change is the storage type of the
symbolic reference is no longer ambiguous with the storge type
of the underlying reference it targets. In the above example,
if master was only available in the packed-refs file, then the
following is also true:
assertSame(Ref.Storage.LOOSE, HEAD.getStorage());
assertSame(Ref.Storage.PACKED, master.getStorage());
(Prior to this change we returned the ambiguous storage of
LOOSE_PACKED for HEAD, which was confusing since it wasn't
actually true on disk).
Another nice side-effect of this change is all intermediate
symbolic references are preserved, and are therefore visible
to the application when they walk the target chain. We can
now correctly inspect chains of symbolic references.
As a result of this change the Ref.getOrigName() method has been
removed from the API. Applications should identify a symbolic
reference by testing for isSymbolic() and not by using an arcane
string comparsion between properties.
Abstract the RefDatabase storage:
---------------------------------
RefDatabase is now abstract, similar to ObjectDatabase, and a
new concrete implementation called RefDirectory is used for the
traditional on-disk storage layout. In the future we plan to
support additional implementations, such as a pure in-memory
RefDatabase for unit testing purposes.
Optimize RefDirectory:
----------------------
The implementation of the in-memory reference cache, reading, and
update routines has been completely rewritten. Much of the code
was heavily borrowed or cribbed from the prior implementation,
so copyright notices have been left intact as much as possible.
The RefDirectory cache no longer confuses symbolic references
with normal references. This permits the cache to resolve the
value of a symbolic reference as late as possible, ensuring it
is always current, without needing to maintain reverse pointers.
The cache is now 2 sorted RefLists, rather than 3 HashMaps.
Using sorted lists allows the implementation to reduce the
in-memory footprint when storing many refs. Using specialized
types for the elements allows the code to avoid additional map
lookups for auxiliary stat information.
To improve scan time during getRefs(), the lists are returned via
a copy-on-write contract. Most callers of getRefs() do not modify
the returned collections, so the copy-on-write semantics improves
access on repositories with a large number of packed references.
Iterator traversals of the returned Map<String,Ref> are performed
using a simple merge-join of the two cache lists, ensuring we can
perform the entire traversal in linear time as a function of the
number of references: O(PackedRefs + LooseRefs).
Scans of the loose reference space to update the cache run in
O(LooseRefs log LooseRefs) time, as the directory contents
are sorted before being merged against the in-memory cache.
Since the majority of stable references are kept packed, there
typically are only a handful of reference names to be sorted,
so the sorting cost should not be very high.
Locking is reduced during getRefs() by taking advantage of the
copy-on-write semantics of the improved cache data structure.
This permits concurrent readers to pull back references without
blocking each other. If there is contention updating the cache
during a scan, one or more updates are simply skipped and will
get picked up again in a future scan.
Writing to the $GIT_DIR/packed-refs during reference delete is
now fully atomic. The file is locked, reparsed fresh, and written
back out if a change is necessary. This avoids all race conditions
with concurrent external updates of the packed-refs file.
The RefLogWriter class has been fully folded into RefDirectory
and is therefore deleted. Maintaining the reference's log is
the responsiblity of the database implementation, and not all
implementations will use java.io for access.
Future work still remains to be done to abstract the ReflogReader
class away from local disk IO.
Change-Id: I26b9287c45a4b2d2be35ba2849daa316f5eec85d
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
14 年前 Rewrite reference handling to be abstract and accurate
This commit actually does three major changes to the way references
are handled within JGit. Unfortunately they were easier to do as
a single massive commit than to break them up into smaller units.
Disambiguate symbolic references:
---------------------------------
Reporting a symbolic reference such as HEAD as though it were
any other normal reference like refs/heads/master causes subtle
programming errors. We have been bitten by this error on several
occasions, as have some downstream applications written by myself.
Instead of reporting HEAD as a reference whose name differs from
its "original name", report it as an actual SymbolicRef object
that the application can test the type and examine the target of.
With this change, Ref is now an abstract type with different
subclasses for the different types.
In the classical example of "HEAD" being a symbolic reference to
branch "refs/heads/master", the Repository.getAllRefs() method
will now return:
Map<String, Ref> all = repository.getAllRefs();
SymbolicRef HEAD = (SymbolicRef) all.get("HEAD");
ObjectIdRef master = (ObjectIdRef) all.get("refs/heads/master");
assertSame(master, HEAD.getTarget());
assertSame(master.getObjectId(), HEAD.getObjectId());
assertEquals("HEAD", HEAD.getName());
assertEquals("refs/heads/master", master.getName());
A nice side-effect of this change is the storage type of the
symbolic reference is no longer ambiguous with the storge type
of the underlying reference it targets. In the above example,
if master was only available in the packed-refs file, then the
following is also true:
assertSame(Ref.Storage.LOOSE, HEAD.getStorage());
assertSame(Ref.Storage.PACKED, master.getStorage());
(Prior to this change we returned the ambiguous storage of
LOOSE_PACKED for HEAD, which was confusing since it wasn't
actually true on disk).
Another nice side-effect of this change is all intermediate
symbolic references are preserved, and are therefore visible
to the application when they walk the target chain. We can
now correctly inspect chains of symbolic references.
As a result of this change the Ref.getOrigName() method has been
removed from the API. Applications should identify a symbolic
reference by testing for isSymbolic() and not by using an arcane
string comparsion between properties.
Abstract the RefDatabase storage:
---------------------------------
RefDatabase is now abstract, similar to ObjectDatabase, and a
new concrete implementation called RefDirectory is used for the
traditional on-disk storage layout. In the future we plan to
support additional implementations, such as a pure in-memory
RefDatabase for unit testing purposes.
Optimize RefDirectory:
----------------------
The implementation of the in-memory reference cache, reading, and
update routines has been completely rewritten. Much of the code
was heavily borrowed or cribbed from the prior implementation,
so copyright notices have been left intact as much as possible.
The RefDirectory cache no longer confuses symbolic references
with normal references. This permits the cache to resolve the
value of a symbolic reference as late as possible, ensuring it
is always current, without needing to maintain reverse pointers.
The cache is now 2 sorted RefLists, rather than 3 HashMaps.
Using sorted lists allows the implementation to reduce the
in-memory footprint when storing many refs. Using specialized
types for the elements allows the code to avoid additional map
lookups for auxiliary stat information.
To improve scan time during getRefs(), the lists are returned via
a copy-on-write contract. Most callers of getRefs() do not modify
the returned collections, so the copy-on-write semantics improves
access on repositories with a large number of packed references.
Iterator traversals of the returned Map<String,Ref> are performed
using a simple merge-join of the two cache lists, ensuring we can
perform the entire traversal in linear time as a function of the
number of references: O(PackedRefs + LooseRefs).
Scans of the loose reference space to update the cache run in
O(LooseRefs log LooseRefs) time, as the directory contents
are sorted before being merged against the in-memory cache.
Since the majority of stable references are kept packed, there
typically are only a handful of reference names to be sorted,
so the sorting cost should not be very high.
Locking is reduced during getRefs() by taking advantage of the
copy-on-write semantics of the improved cache data structure.
This permits concurrent readers to pull back references without
blocking each other. If there is contention updating the cache
during a scan, one or more updates are simply skipped and will
get picked up again in a future scan.
Writing to the $GIT_DIR/packed-refs during reference delete is
now fully atomic. The file is locked, reparsed fresh, and written
back out if a change is necessary. This avoids all race conditions
with concurrent external updates of the packed-refs file.
The RefLogWriter class has been fully folded into RefDirectory
and is therefore deleted. Maintaining the reference's log is
the responsiblity of the database implementation, and not all
implementations will use java.io for access.
Future work still remains to be done to abstract the ReflogReader
class away from local disk IO.
Change-Id: I26b9287c45a4b2d2be35ba2849daa316f5eec85d
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
14 年前 Rewrite reference handling to be abstract and accurate
This commit actually does three major changes to the way references
are handled within JGit. Unfortunately they were easier to do as
a single massive commit than to break them up into smaller units.
Disambiguate symbolic references:
---------------------------------
Reporting a symbolic reference such as HEAD as though it were
any other normal reference like refs/heads/master causes subtle
programming errors. We have been bitten by this error on several
occasions, as have some downstream applications written by myself.
Instead of reporting HEAD as a reference whose name differs from
its "original name", report it as an actual SymbolicRef object
that the application can test the type and examine the target of.
With this change, Ref is now an abstract type with different
subclasses for the different types.
In the classical example of "HEAD" being a symbolic reference to
branch "refs/heads/master", the Repository.getAllRefs() method
will now return:
Map<String, Ref> all = repository.getAllRefs();
SymbolicRef HEAD = (SymbolicRef) all.get("HEAD");
ObjectIdRef master = (ObjectIdRef) all.get("refs/heads/master");
assertSame(master, HEAD.getTarget());
assertSame(master.getObjectId(), HEAD.getObjectId());
assertEquals("HEAD", HEAD.getName());
assertEquals("refs/heads/master", master.getName());
A nice side-effect of this change is the storage type of the
symbolic reference is no longer ambiguous with the storge type
of the underlying reference it targets. In the above example,
if master was only available in the packed-refs file, then the
following is also true:
assertSame(Ref.Storage.LOOSE, HEAD.getStorage());
assertSame(Ref.Storage.PACKED, master.getStorage());
(Prior to this change we returned the ambiguous storage of
LOOSE_PACKED for HEAD, which was confusing since it wasn't
actually true on disk).
Another nice side-effect of this change is all intermediate
symbolic references are preserved, and are therefore visible
to the application when they walk the target chain. We can
now correctly inspect chains of symbolic references.
As a result of this change the Ref.getOrigName() method has been
removed from the API. Applications should identify a symbolic
reference by testing for isSymbolic() and not by using an arcane
string comparsion between properties.
Abstract the RefDatabase storage:
---------------------------------
RefDatabase is now abstract, similar to ObjectDatabase, and a
new concrete implementation called RefDirectory is used for the
traditional on-disk storage layout. In the future we plan to
support additional implementations, such as a pure in-memory
RefDatabase for unit testing purposes.
Optimize RefDirectory:
----------------------
The implementation of the in-memory reference cache, reading, and
update routines has been completely rewritten. Much of the code
was heavily borrowed or cribbed from the prior implementation,
so copyright notices have been left intact as much as possible.
The RefDirectory cache no longer confuses symbolic references
with normal references. This permits the cache to resolve the
value of a symbolic reference as late as possible, ensuring it
is always current, without needing to maintain reverse pointers.
The cache is now 2 sorted RefLists, rather than 3 HashMaps.
Using sorted lists allows the implementation to reduce the
in-memory footprint when storing many refs. Using specialized
types for the elements allows the code to avoid additional map
lookups for auxiliary stat information.
To improve scan time during getRefs(), the lists are returned via
a copy-on-write contract. Most callers of getRefs() do not modify
the returned collections, so the copy-on-write semantics improves
access on repositories with a large number of packed references.
Iterator traversals of the returned Map<String,Ref> are performed
using a simple merge-join of the two cache lists, ensuring we can
perform the entire traversal in linear time as a function of the
number of references: O(PackedRefs + LooseRefs).
Scans of the loose reference space to update the cache run in
O(LooseRefs log LooseRefs) time, as the directory contents
are sorted before being merged against the in-memory cache.
Since the majority of stable references are kept packed, there
typically are only a handful of reference names to be sorted,
so the sorting cost should not be very high.
Locking is reduced during getRefs() by taking advantage of the
copy-on-write semantics of the improved cache data structure.
This permits concurrent readers to pull back references without
blocking each other. If there is contention updating the cache
during a scan, one or more updates are simply skipped and will
get picked up again in a future scan.
Writing to the $GIT_DIR/packed-refs during reference delete is
now fully atomic. The file is locked, reparsed fresh, and written
back out if a change is necessary. This avoids all race conditions
with concurrent external updates of the packed-refs file.
The RefLogWriter class has been fully folded into RefDirectory
and is therefore deleted. Maintaining the reference's log is
the responsiblity of the database implementation, and not all
implementations will use java.io for access.
Future work still remains to be done to abstract the ReflogReader
class away from local disk IO.
Change-Id: I26b9287c45a4b2d2be35ba2849daa316f5eec85d
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
14 年前 Rewrite reference handling to be abstract and accurate
This commit actually does three major changes to the way references
are handled within JGit. Unfortunately they were easier to do as
a single massive commit than to break them up into smaller units.
Disambiguate symbolic references:
---------------------------------
Reporting a symbolic reference such as HEAD as though it were
any other normal reference like refs/heads/master causes subtle
programming errors. We have been bitten by this error on several
occasions, as have some downstream applications written by myself.
Instead of reporting HEAD as a reference whose name differs from
its "original name", report it as an actual SymbolicRef object
that the application can test the type and examine the target of.
With this change, Ref is now an abstract type with different
subclasses for the different types.
In the classical example of "HEAD" being a symbolic reference to
branch "refs/heads/master", the Repository.getAllRefs() method
will now return:
Map<String, Ref> all = repository.getAllRefs();
SymbolicRef HEAD = (SymbolicRef) all.get("HEAD");
ObjectIdRef master = (ObjectIdRef) all.get("refs/heads/master");
assertSame(master, HEAD.getTarget());
assertSame(master.getObjectId(), HEAD.getObjectId());
assertEquals("HEAD", HEAD.getName());
assertEquals("refs/heads/master", master.getName());
A nice side-effect of this change is the storage type of the
symbolic reference is no longer ambiguous with the storge type
of the underlying reference it targets. In the above example,
if master was only available in the packed-refs file, then the
following is also true:
assertSame(Ref.Storage.LOOSE, HEAD.getStorage());
assertSame(Ref.Storage.PACKED, master.getStorage());
(Prior to this change we returned the ambiguous storage of
LOOSE_PACKED for HEAD, which was confusing since it wasn't
actually true on disk).
Another nice side-effect of this change is all intermediate
symbolic references are preserved, and are therefore visible
to the application when they walk the target chain. We can
now correctly inspect chains of symbolic references.
As a result of this change the Ref.getOrigName() method has been
removed from the API. Applications should identify a symbolic
reference by testing for isSymbolic() and not by using an arcane
string comparsion between properties.
Abstract the RefDatabase storage:
---------------------------------
RefDatabase is now abstract, similar to ObjectDatabase, and a
new concrete implementation called RefDirectory is used for the
traditional on-disk storage layout. In the future we plan to
support additional implementations, such as a pure in-memory
RefDatabase for unit testing purposes.
Optimize RefDirectory:
----------------------
The implementation of the in-memory reference cache, reading, and
update routines has been completely rewritten. Much of the code
was heavily borrowed or cribbed from the prior implementation,
so copyright notices have been left intact as much as possible.
The RefDirectory cache no longer confuses symbolic references
with normal references. This permits the cache to resolve the
value of a symbolic reference as late as possible, ensuring it
is always current, without needing to maintain reverse pointers.
The cache is now 2 sorted RefLists, rather than 3 HashMaps.
Using sorted lists allows the implementation to reduce the
in-memory footprint when storing many refs. Using specialized
types for the elements allows the code to avoid additional map
lookups for auxiliary stat information.
To improve scan time during getRefs(), the lists are returned via
a copy-on-write contract. Most callers of getRefs() do not modify
the returned collections, so the copy-on-write semantics improves
access on repositories with a large number of packed references.
Iterator traversals of the returned Map<String,Ref> are performed
using a simple merge-join of the two cache lists, ensuring we can
perform the entire traversal in linear time as a function of the
number of references: O(PackedRefs + LooseRefs).
Scans of the loose reference space to update the cache run in
O(LooseRefs log LooseRefs) time, as the directory contents
are sorted before being merged against the in-memory cache.
Since the majority of stable references are kept packed, there
typically are only a handful of reference names to be sorted,
so the sorting cost should not be very high.
Locking is reduced during getRefs() by taking advantage of the
copy-on-write semantics of the improved cache data structure.
This permits concurrent readers to pull back references without
blocking each other. If there is contention updating the cache
during a scan, one or more updates are simply skipped and will
get picked up again in a future scan.
Writing to the $GIT_DIR/packed-refs during reference delete is
now fully atomic. The file is locked, reparsed fresh, and written
back out if a change is necessary. This avoids all race conditions
with concurrent external updates of the packed-refs file.
The RefLogWriter class has been fully folded into RefDirectory
and is therefore deleted. Maintaining the reference's log is
the responsiblity of the database implementation, and not all
implementations will use java.io for access.
Future work still remains to be done to abstract the ReflogReader
class away from local disk IO.
Change-Id: I26b9287c45a4b2d2be35ba2849daa316f5eec85d
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
14 年前 Rewrite reference handling to be abstract and accurate
This commit actually does three major changes to the way references
are handled within JGit. Unfortunately they were easier to do as
a single massive commit than to break them up into smaller units.
Disambiguate symbolic references:
---------------------------------
Reporting a symbolic reference such as HEAD as though it were
any other normal reference like refs/heads/master causes subtle
programming errors. We have been bitten by this error on several
occasions, as have some downstream applications written by myself.
Instead of reporting HEAD as a reference whose name differs from
its "original name", report it as an actual SymbolicRef object
that the application can test the type and examine the target of.
With this change, Ref is now an abstract type with different
subclasses for the different types.
In the classical example of "HEAD" being a symbolic reference to
branch "refs/heads/master", the Repository.getAllRefs() method
will now return:
Map<String, Ref> all = repository.getAllRefs();
SymbolicRef HEAD = (SymbolicRef) all.get("HEAD");
ObjectIdRef master = (ObjectIdRef) all.get("refs/heads/master");
assertSame(master, HEAD.getTarget());
assertSame(master.getObjectId(), HEAD.getObjectId());
assertEquals("HEAD", HEAD.getName());
assertEquals("refs/heads/master", master.getName());
A nice side-effect of this change is the storage type of the
symbolic reference is no longer ambiguous with the storge type
of the underlying reference it targets. In the above example,
if master was only available in the packed-refs file, then the
following is also true:
assertSame(Ref.Storage.LOOSE, HEAD.getStorage());
assertSame(Ref.Storage.PACKED, master.getStorage());
(Prior to this change we returned the ambiguous storage of
LOOSE_PACKED for HEAD, which was confusing since it wasn't
actually true on disk).
Another nice side-effect of this change is all intermediate
symbolic references are preserved, and are therefore visible
to the application when they walk the target chain. We can
now correctly inspect chains of symbolic references.
As a result of this change the Ref.getOrigName() method has been
removed from the API. Applications should identify a symbolic
reference by testing for isSymbolic() and not by using an arcane
string comparsion between properties.
Abstract the RefDatabase storage:
---------------------------------
RefDatabase is now abstract, similar to ObjectDatabase, and a
new concrete implementation called RefDirectory is used for the
traditional on-disk storage layout. In the future we plan to
support additional implementations, such as a pure in-memory
RefDatabase for unit testing purposes.
Optimize RefDirectory:
----------------------
The implementation of the in-memory reference cache, reading, and
update routines has been completely rewritten. Much of the code
was heavily borrowed or cribbed from the prior implementation,
so copyright notices have been left intact as much as possible.
The RefDirectory cache no longer confuses symbolic references
with normal references. This permits the cache to resolve the
value of a symbolic reference as late as possible, ensuring it
is always current, without needing to maintain reverse pointers.
The cache is now 2 sorted RefLists, rather than 3 HashMaps.
Using sorted lists allows the implementation to reduce the
in-memory footprint when storing many refs. Using specialized
types for the elements allows the code to avoid additional map
lookups for auxiliary stat information.
To improve scan time during getRefs(), the lists are returned via
a copy-on-write contract. Most callers of getRefs() do not modify
the returned collections, so the copy-on-write semantics improves
access on repositories with a large number of packed references.
Iterator traversals of the returned Map<String,Ref> are performed
using a simple merge-join of the two cache lists, ensuring we can
perform the entire traversal in linear time as a function of the
number of references: O(PackedRefs + LooseRefs).
Scans of the loose reference space to update the cache run in
O(LooseRefs log LooseRefs) time, as the directory contents
are sorted before being merged against the in-memory cache.
Since the majority of stable references are kept packed, there
typically are only a handful of reference names to be sorted,
so the sorting cost should not be very high.
Locking is reduced during getRefs() by taking advantage of the
copy-on-write semantics of the improved cache data structure.
This permits concurrent readers to pull back references without
blocking each other. If there is contention updating the cache
during a scan, one or more updates are simply skipped and will
get picked up again in a future scan.
Writing to the $GIT_DIR/packed-refs during reference delete is
now fully atomic. The file is locked, reparsed fresh, and written
back out if a change is necessary. This avoids all race conditions
with concurrent external updates of the packed-refs file.
The RefLogWriter class has been fully folded into RefDirectory
and is therefore deleted. Maintaining the reference's log is
the responsiblity of the database implementation, and not all
implementations will use java.io for access.
Future work still remains to be done to abstract the ReflogReader
class away from local disk IO.
Change-Id: I26b9287c45a4b2d2be35ba2849daa316f5eec85d
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
14 年前 |
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563 |
- /*
- * Copyright (C) 2007, Robin Rosenberg <robin.rosenberg@dewire.com>
- * Copyright (C) 2006-2021, Shawn O. Pearce <spearce@spearce.org> and others
- *
- * This program and the accompanying materials are made available under the
- * terms of the Eclipse Distribution License v. 1.0 which is available at
- * https://www.eclipse.org/org/documents/edl-v10.php.
- *
- * SPDX-License-Identifier: BSD-3-Clause
- */
-
- package org.eclipse.jgit.internal.storage.file;
-
- import static org.eclipse.jgit.lib.Constants.LOCK_SUFFIX;
-
- import java.io.File;
- import java.io.FileInputStream;
- import java.io.FileNotFoundException;
- import java.io.FileOutputStream;
- import java.io.FilenameFilter;
- import java.io.IOException;
- import java.io.OutputStream;
- import java.nio.ByteBuffer;
- import java.nio.channels.Channels;
- import java.nio.channels.FileChannel;
- import java.nio.file.Files;
- import java.nio.file.StandardCopyOption;
- import java.nio.file.attribute.FileTime;
- import java.text.MessageFormat;
- import java.time.Instant;
- import java.util.concurrent.TimeUnit;
-
- import org.eclipse.jgit.internal.JGitText;
- import org.eclipse.jgit.lib.Constants;
- import org.eclipse.jgit.lib.ObjectId;
- import org.eclipse.jgit.util.FS;
- import org.eclipse.jgit.util.FS.LockToken;
- import org.eclipse.jgit.util.FileUtils;
- import org.slf4j.Logger;
- import org.slf4j.LoggerFactory;
-
- /**
- * Git style file locking and replacement.
- * <p>
- * To modify a ref file Git tries to use an atomic update approach: we write the
- * new data into a brand new file, then rename it in place over the old name.
- * This way we can just delete the temporary file if anything goes wrong, and
- * nothing has been damaged. To coordinate access from multiple processes at
- * once Git tries to atomically create the new temporary file under a well-known
- * name.
- */
- public class LockFile {
- private static final Logger LOG = LoggerFactory.getLogger(LockFile.class);
-
- /**
- * Unlock the given file.
- * <p>
- * This method can be used for recovering from a thrown
- * {@link org.eclipse.jgit.errors.LockFailedException} . This method does
- * not validate that the lock is or is not currently held before attempting
- * to unlock it.
- *
- * @param file
- * a {@link java.io.File} object.
- * @return true if unlocked, false if unlocking failed
- */
- public static boolean unlock(File file) {
- final File lockFile = getLockFile(file);
- final int flags = FileUtils.RETRY | FileUtils.SKIP_MISSING;
- try {
- FileUtils.delete(lockFile, flags);
- } catch (IOException ignored) {
- // Ignore and return whether lock file still exists
- }
- return !lockFile.exists();
- }
-
- /**
- * Get the lock file corresponding to the given file.
- *
- * @param file
- * @return lock file
- */
- static File getLockFile(File file) {
- return new File(file.getParentFile(),
- file.getName() + LOCK_SUFFIX);
- }
-
- /** Filter to skip over active lock files when listing a directory. */
- static final FilenameFilter FILTER = (File dir,
- String name) -> !name.endsWith(LOCK_SUFFIX);
-
- private final File ref;
-
- private final File lck;
-
- private boolean haveLck;
-
- private FileOutputStream os;
-
- private boolean needSnapshot;
-
- private boolean fsync;
-
- private boolean isAppend;
-
- private boolean written;
-
- private FileSnapshot commitSnapshot;
-
- private LockToken token;
-
- /**
- * Create a new lock for any file.
- *
- * @param f
- * the file that will be locked.
- */
- public LockFile(File f) {
- ref = f;
- lck = getLockFile(ref);
- }
-
- /**
- * Try to establish the lock.
- *
- * @return true if the lock is now held by the caller; false if it is held
- * by someone else.
- * @throws java.io.IOException
- * the temporary output file could not be created. The caller
- * does not hold the lock.
- */
- public boolean lock() throws IOException {
- if (haveLck) {
- throw new IllegalStateException(
- MessageFormat.format(JGitText.get().lockAlreadyHeld, ref));
- }
- FileUtils.mkdirs(lck.getParentFile(), true);
- try {
- token = FS.DETECTED.createNewFileAtomic(lck);
- } catch (IOException e) {
- LOG.error(JGitText.get().failedCreateLockFile, lck, e);
- throw e;
- }
- boolean obtainedLock = token.isCreated();
- if (obtainedLock) {
- haveLck = true;
- isAppend = false;
- written = false;
- } else {
- closeToken();
- }
- return obtainedLock;
- }
-
- /**
- * Try to establish the lock for appending.
- *
- * @return true if the lock is now held by the caller; false if it is held
- * by someone else.
- * @throws java.io.IOException
- * the temporary output file could not be created. The caller
- * does not hold the lock.
- */
- public boolean lockForAppend() throws IOException {
- if (!lock()) {
- return false;
- }
- copyCurrentContent();
- isAppend = true;
- written = false;
- return true;
- }
-
- // For tests only
- boolean isLocked() {
- return haveLck;
- }
-
- private FileOutputStream getStream() throws IOException {
- return new FileOutputStream(lck, isAppend);
- }
-
- /**
- * Copy the current file content into the temporary file.
- * <p>
- * This method saves the current file content by inserting it into the
- * temporary file, so that the caller can safely append rather than replace
- * the primary file.
- * <p>
- * This method does nothing if the current file does not exist, or exists
- * but is empty.
- *
- * @throws java.io.IOException
- * the temporary file could not be written, or a read error
- * occurred while reading from the current file. The lock is
- * released before throwing the underlying IO exception to the
- * caller.
- * @throws java.lang.RuntimeException
- * the temporary file could not be written. The lock is released
- * before throwing the underlying exception to the caller.
- */
- public void copyCurrentContent() throws IOException {
- requireLock();
- try (FileOutputStream out = getStream()) {
- try (FileInputStream fis = new FileInputStream(ref)) {
- if (fsync) {
- FileChannel in = fis.getChannel();
- long pos = 0;
- long cnt = in.size();
- while (0 < cnt) {
- long r = out.getChannel().transferFrom(in, pos, cnt);
- pos += r;
- cnt -= r;
- }
- } else {
- final byte[] buf = new byte[2048];
- int r;
- while ((r = fis.read(buf)) >= 0) {
- out.write(buf, 0, r);
- }
- }
- } catch (FileNotFoundException fnfe) {
- if (ref.exists()) {
- throw fnfe;
- }
- // Don't worry about a file that doesn't exist yet, it
- // conceptually has no current content to copy.
- }
- } catch (IOException | RuntimeException | Error ioe) {
- unlock();
- throw ioe;
- }
- }
-
- /**
- * Write an ObjectId and LF to the temporary file.
- *
- * @param id
- * the id to store in the file. The id will be written in hex,
- * followed by a sole LF.
- * @throws java.io.IOException
- * the temporary file could not be written. The lock is released
- * before throwing the underlying IO exception to the caller.
- * @throws java.lang.RuntimeException
- * the temporary file could not be written. The lock is released
- * before throwing the underlying exception to the caller.
- */
- public void write(ObjectId id) throws IOException {
- byte[] buf = new byte[Constants.OBJECT_ID_STRING_LENGTH + 1];
- id.copyTo(buf, 0);
- buf[Constants.OBJECT_ID_STRING_LENGTH] = '\n';
- write(buf);
- }
-
- /**
- * Write arbitrary data to the temporary file.
- *
- * @param content
- * the bytes to store in the temporary file. No additional bytes
- * are added, so if the file must end with an LF it must appear
- * at the end of the byte array.
- * @throws java.io.IOException
- * the temporary file could not be written. The lock is released
- * before throwing the underlying IO exception to the caller.
- * @throws java.lang.RuntimeException
- * the temporary file could not be written. The lock is released
- * before throwing the underlying exception to the caller.
- */
- public void write(byte[] content) throws IOException {
- requireLock();
- try (FileOutputStream out = getStream()) {
- if (written) {
- throw new IOException(MessageFormat
- .format(JGitText.get().lockStreamClosed, ref));
- }
- if (fsync) {
- FileChannel fc = out.getChannel();
- ByteBuffer buf = ByteBuffer.wrap(content);
- while (0 < buf.remaining()) {
- fc.write(buf);
- }
- fc.force(true);
- } else {
- out.write(content);
- }
- written = true;
- } catch (IOException | RuntimeException | Error ioe) {
- unlock();
- throw ioe;
- }
- }
-
- /**
- * Obtain the direct output stream for this lock.
- * <p>
- * The stream may only be accessed once, and only after {@link #lock()} has
- * been successfully invoked and returned true. Callers must close the
- * stream prior to calling {@link #commit()} to commit the change.
- *
- * @return a stream to write to the new file. The stream is unbuffered.
- */
- public OutputStream getOutputStream() {
- requireLock();
-
- if (written || os != null) {
- throw new IllegalStateException(MessageFormat
- .format(JGitText.get().lockStreamMultiple, ref));
- }
-
- return new OutputStream() {
-
- private OutputStream out;
-
- private boolean closed;
-
- private OutputStream get() throws IOException {
- if (written) {
- throw new IOException(MessageFormat
- .format(JGitText.get().lockStreamMultiple, ref));
- }
- if (out == null) {
- os = getStream();
- if (fsync) {
- out = Channels.newOutputStream(os.getChannel());
- } else {
- out = os;
- }
- }
- return out;
- }
-
- @Override
- public void write(byte[] b, int o, int n) throws IOException {
- get().write(b, o, n);
- }
-
- @Override
- public void write(byte[] b) throws IOException {
- get().write(b);
- }
-
- @Override
- public void write(int b) throws IOException {
- get().write(b);
- }
-
- @Override
- public void close() throws IOException {
- if (closed) {
- return;
- }
- closed = true;
- try {
- if (written) {
- throw new IOException(MessageFormat
- .format(JGitText.get().lockStreamClosed, ref));
- }
- if (out != null) {
- if (fsync) {
- os.getChannel().force(true);
- }
- out.close();
- os = null;
- }
- written = true;
- } catch (IOException | RuntimeException | Error ioe) {
- unlock();
- throw ioe;
- }
- }
- };
- }
-
- void requireLock() {
- if (!haveLck) {
- unlock();
- throw new IllegalStateException(MessageFormat.format(JGitText.get().lockOnNotHeld, ref));
- }
- }
-
- /**
- * Request that {@link #commit()} remember modification time.
- * <p>
- * This is an alias for {@code setNeedSnapshot(true)}.
- *
- * @param on
- * true if the commit method must remember the modification time.
- */
- public void setNeedStatInformation(boolean on) {
- setNeedSnapshot(on);
- }
-
- /**
- * Request that {@link #commit()} remember the
- * {@link org.eclipse.jgit.internal.storage.file.FileSnapshot}.
- *
- * @param on
- * true if the commit method must remember the FileSnapshot.
- */
- public void setNeedSnapshot(boolean on) {
- needSnapshot = on;
- }
-
- /**
- * Request that {@link #commit()} force dirty data to the drive.
- *
- * @param on
- * true if dirty data should be forced to the drive.
- */
- public void setFSync(boolean on) {
- fsync = on;
- }
-
- /**
- * Wait until the lock file information differs from the old file.
- * <p>
- * This method tests the last modification date. If both are the same, this
- * method sleeps until it can force the new lock file's modification date to
- * be later than the target file.
- *
- * @throws java.lang.InterruptedException
- * the thread was interrupted before the last modified date of
- * the lock file was different from the last modified date of
- * the target file.
- */
- public void waitForStatChange() throws InterruptedException {
- FileSnapshot o = FileSnapshot.save(ref);
- FileSnapshot n = FileSnapshot.save(lck);
- long fsTimeResolution = FS.getFileStoreAttributes(lck.toPath())
- .getFsTimestampResolution().toNanos();
- while (o.equals(n)) {
- TimeUnit.NANOSECONDS.sleep(fsTimeResolution);
- try {
- Files.setLastModifiedTime(lck.toPath(),
- FileTime.from(Instant.now()));
- } catch (IOException e) {
- n.waitUntilNotRacy();
- }
- n = FileSnapshot.save(lck);
- }
- }
-
- /**
- * Commit this change and release the lock.
- * <p>
- * If this method fails (returns false) the lock is still released.
- *
- * @return true if the commit was successful and the file contains the new
- * data; false if the commit failed and the file remains with the
- * old data.
- * @throws java.lang.IllegalStateException
- * the lock is not held.
- */
- public boolean commit() {
- if (os != null) {
- unlock();
- throw new IllegalStateException(MessageFormat.format(JGitText.get().lockOnNotClosed, ref));
- }
-
- saveStatInformation();
- try {
- FileUtils.rename(lck, ref, StandardCopyOption.ATOMIC_MOVE);
- haveLck = false;
- isAppend = false;
- written = false;
- closeToken();
- return true;
- } catch (IOException e) {
- unlock();
- return false;
- }
- }
-
- private void closeToken() {
- if (token != null) {
- token.close();
- token = null;
- }
- }
-
- private void saveStatInformation() {
- if (needSnapshot)
- commitSnapshot = FileSnapshot.save(lck);
- }
-
- /**
- * Get the modification time of the output file when it was committed.
- *
- * @return modification time of the lock file right before we committed it.
- * @deprecated use {@link #getCommitLastModifiedInstant()} instead
- */
- @Deprecated
- public long getCommitLastModified() {
- return commitSnapshot.lastModified();
- }
-
- /**
- * Get the modification time of the output file when it was committed.
- *
- * @return modification time of the lock file right before we committed it.
- */
- public Instant getCommitLastModifiedInstant() {
- return commitSnapshot.lastModifiedInstant();
- }
-
- /**
- * Get the {@link FileSnapshot} just before commit.
- *
- * @return get the {@link FileSnapshot} just before commit.
- */
- public FileSnapshot getCommitSnapshot() {
- return commitSnapshot;
- }
-
- /**
- * Update the commit snapshot {@link #getCommitSnapshot()} before commit.
- * <p>
- * This may be necessary if you need time stamp before commit occurs, e.g
- * while writing the index.
- */
- public void createCommitSnapshot() {
- saveStatInformation();
- }
-
- /**
- * Unlock this file and abort this change.
- * <p>
- * The temporary file (if created) is deleted before returning.
- */
- public void unlock() {
- if (os != null) {
- try {
- os.close();
- } catch (IOException e) {
- LOG.error(MessageFormat
- .format(JGitText.get().unlockLockFileFailed, lck), e);
- }
- os = null;
- }
-
- if (haveLck) {
- haveLck = false;
- try {
- FileUtils.delete(lck, FileUtils.RETRY);
- } catch (IOException e) {
- LOG.error(MessageFormat
- .format(JGitText.get().unlockLockFileFailed, lck), e);
- } finally {
- closeToken();
- }
- }
- isAppend = false;
- written = false;
- }
-
- /** {@inheritDoc} */
- @SuppressWarnings("nls")
- @Override
- public String toString() {
- return "LockFile[" + lck + ", haveLck=" + haveLck + "]";
- }
- }
|