2011-10-30 03:48:46 +00:00
|
|
|
{- git-annex command data types
|
|
|
|
|
-
|
make CommandStart return a StartMessage
The goal is to be able to run CommandStart in the main thread when -J is
used, rather than unncessarily passing it off to a worker thread, which
incurs overhead that is signficant when the CommandStart is going to
quickly decide to stop.
To do that, the message it displays needs to be displayed in the worker
thread, after the CommandStart has run.
Also, the change will mean that CommandStart will no longer necessarily
run with the same Annex state as CommandPerform. While its docs already
said it should avoid modifying Annex state, I audited all the
CommandStart code as part of the conversion. (Note that CommandSeek
already sometimes runs with a different Annex state, and that has not been
a source of any problems, so I am not too worried that this change will
lead to breakage going forward.)
The only modification of Annex state I found was it calling
allowMessages in some Commands that default to noMessages. Dealt with
that by adding a startCustomOutput and a startingUsualMessages.
This lets a command start with noMessages and then select the output it
wants for each CommandStart.
One bit of breakage: onlyActionOn has been removed from commands that used it.
The plan is that, since a StartMessage contains an ActionItem,
when a Key can be extracted from that, the parallel job runner can
run onlyActionOn' automatically. Then commands won't need to worry about
this detail. Future work.
Otherwise, this was a fairly straightforward process of making each
CommandStart compile again. Hopefully other behavior changes were mostly
avoided.
In a few cases, a command had a CommandStart that called a CommandPerform
that then called showStart multiple times. I have collapsed those
down to a single start action. The main command to perhaps suffer from it
is Command.Direct, which used to show a start for each file, and no
longer does.
Another minor behavior change is that some commands used showStart
before, but had an associated file and a Key available, so were changed
to ShowStart with an ActionItemAssociatedFile. That will not change the
normal output or behavior, but --json output will now include the key.
This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
|
|
|
- Copyright 2010-2019 Joey Hess <id@joeyh.name>
|
2011-10-30 03:48:46 +00:00
|
|
|
-
|
2019-03-13 19:48:14 +00:00
|
|
|
- Licensed under the GNU AGPL version 3 or higher.
|
2011-10-30 03:48:46 +00:00
|
|
|
-}
|
|
|
|
|
|
|
|
|
|
module Types.Command where
|
|
|
|
|
|
2012-02-16 20:53:44 +00:00
|
|
|
import Data.Ord
|
2015-07-08 16:33:27 +00:00
|
|
|
import Options.Applicative.Types (Parser)
|
2012-02-16 20:53:44 +00:00
|
|
|
|
2011-10-30 03:48:46 +00:00
|
|
|
import Types
|
2016-09-05 19:32:59 +00:00
|
|
|
import Types.DeferredParse
|
make CommandStart return a StartMessage
The goal is to be able to run CommandStart in the main thread when -J is
used, rather than unncessarily passing it off to a worker thread, which
incurs overhead that is signficant when the CommandStart is going to
quickly decide to stop.
To do that, the message it displays needs to be displayed in the worker
thread, after the CommandStart has run.
Also, the change will mean that CommandStart will no longer necessarily
run with the same Annex state as CommandPerform. While its docs already
said it should avoid modifying Annex state, I audited all the
CommandStart code as part of the conversion. (Note that CommandSeek
already sometimes runs with a different Annex state, and that has not been
a source of any problems, so I am not too worried that this change will
lead to breakage going forward.)
The only modification of Annex state I found was it calling
allowMessages in some Commands that default to noMessages. Dealt with
that by adding a startCustomOutput and a startingUsualMessages.
This lets a command start with noMessages and then select the output it
wants for each CommandStart.
One bit of breakage: onlyActionOn has been removed from commands that used it.
The plan is that, since a StartMessage contains an ActionItem,
when a Key can be extracted from that, the parallel job runner can
run onlyActionOn' automatically. Then commands won't need to worry about
this detail. Future work.
Otherwise, this was a fairly straightforward process of making each
CommandStart compile again. Hopefully other behavior changes were mostly
avoided.
In a few cases, a command had a CommandStart that called a CommandPerform
that then called showStart multiple times. I have collapsed those
down to a single start action. The main command to perhaps suffer from it
is Command.Direct, which used to show a start for each file, and no
longer does.
Another minor behavior change is that some commands used showStart
before, but had an associated file and a Key available, so were changed
to ShowStart with an ActionItemAssociatedFile. That will not change the
normal output or behavior, but --json output will now include the key.
This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
|
|
|
import Types.ActionItem
|
2011-10-30 03:48:46 +00:00
|
|
|
|
|
|
|
|
{- A command runs in these stages.
|
|
|
|
|
-
|
2015-07-08 16:33:27 +00:00
|
|
|
- a. The parser stage parses the command line and generates a CommandSeek
|
|
|
|
|
- action. -}
|
|
|
|
|
type CommandParser = Parser CommandSeek
|
|
|
|
|
{- b. The check stage runs checks, that error out if
|
2011-10-30 03:48:46 +00:00
|
|
|
- anything prevents the command from running. -}
|
|
|
|
|
data CommandCheck = CommandCheck { idCheck :: Int, runCheck :: Annex () }
|
2015-07-08 16:33:27 +00:00
|
|
|
{- c. The seek stage is passed input from the parser, looks through
|
|
|
|
|
- the repo to find things to act on (ie, new files to add), and
|
|
|
|
|
- runs commandAction to handle all necessary actions. -}
|
|
|
|
|
type CommandSeek = Annex ()
|
make CommandStart return a StartMessage
The goal is to be able to run CommandStart in the main thread when -J is
used, rather than unncessarily passing it off to a worker thread, which
incurs overhead that is signficant when the CommandStart is going to
quickly decide to stop.
To do that, the message it displays needs to be displayed in the worker
thread, after the CommandStart has run.
Also, the change will mean that CommandStart will no longer necessarily
run with the same Annex state as CommandPerform. While its docs already
said it should avoid modifying Annex state, I audited all the
CommandStart code as part of the conversion. (Note that CommandSeek
already sometimes runs with a different Annex state, and that has not been
a source of any problems, so I am not too worried that this change will
lead to breakage going forward.)
The only modification of Annex state I found was it calling
allowMessages in some Commands that default to noMessages. Dealt with
that by adding a startCustomOutput and a startingUsualMessages.
This lets a command start with noMessages and then select the output it
wants for each CommandStart.
One bit of breakage: onlyActionOn has been removed from commands that used it.
The plan is that, since a StartMessage contains an ActionItem,
when a Key can be extracted from that, the parallel job runner can
run onlyActionOn' automatically. Then commands won't need to worry about
this detail. Future work.
Otherwise, this was a fairly straightforward process of making each
CommandStart compile again. Hopefully other behavior changes were mostly
avoided.
In a few cases, a command had a CommandStart that called a CommandPerform
that then called showStart multiple times. I have collapsed those
down to a single start action. The main command to perhaps suffer from it
is Command.Direct, which used to show a start for each file, and no
longer does.
Another minor behavior change is that some commands used showStart
before, but had an associated file and a Key available, so were changed
to ShowStart with an ActionItemAssociatedFile. That will not change the
normal output or behavior, but --json output will now include the key.
This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
|
|
|
{- d. The start stage is run before anything is output, is passed some
|
|
|
|
|
- value from the seek stage, and can check if anything needs to be
|
|
|
|
|
- done, and early abort if not. It should run quickly and should
|
|
|
|
|
- not modify Annex state or output anything. -}
|
|
|
|
|
type CommandStart = Annex (Maybe (StartMessage, CommandPerform))
|
2015-07-08 16:33:27 +00:00
|
|
|
{- e. The perform stage is run after a message is printed about the command
|
2011-10-30 03:48:46 +00:00
|
|
|
- being run, and it should be where the bulk of the work happens. -}
|
|
|
|
|
type CommandPerform = Annex (Maybe CommandCleanup)
|
2015-07-08 16:33:27 +00:00
|
|
|
{- f. The cleanup stage is run only if the perform stage succeeds, and it
|
2011-10-30 03:48:46 +00:00
|
|
|
- returns the overall success/fail of the command. -}
|
|
|
|
|
type CommandCleanup = Annex Bool
|
|
|
|
|
|
make CommandStart return a StartMessage
The goal is to be able to run CommandStart in the main thread when -J is
used, rather than unncessarily passing it off to a worker thread, which
incurs overhead that is signficant when the CommandStart is going to
quickly decide to stop.
To do that, the message it displays needs to be displayed in the worker
thread, after the CommandStart has run.
Also, the change will mean that CommandStart will no longer necessarily
run with the same Annex state as CommandPerform. While its docs already
said it should avoid modifying Annex state, I audited all the
CommandStart code as part of the conversion. (Note that CommandSeek
already sometimes runs with a different Annex state, and that has not been
a source of any problems, so I am not too worried that this change will
lead to breakage going forward.)
The only modification of Annex state I found was it calling
allowMessages in some Commands that default to noMessages. Dealt with
that by adding a startCustomOutput and a startingUsualMessages.
This lets a command start with noMessages and then select the output it
wants for each CommandStart.
One bit of breakage: onlyActionOn has been removed from commands that used it.
The plan is that, since a StartMessage contains an ActionItem,
when a Key can be extracted from that, the parallel job runner can
run onlyActionOn' automatically. Then commands won't need to worry about
this detail. Future work.
Otherwise, this was a fairly straightforward process of making each
CommandStart compile again. Hopefully other behavior changes were mostly
avoided.
In a few cases, a command had a CommandStart that called a CommandPerform
that then called showStart multiple times. I have collapsed those
down to a single start action. The main command to perhaps suffer from it
is Command.Direct, which used to show a start for each file, and no
longer does.
Another minor behavior change is that some commands used showStart
before, but had an associated file and a Key available, so were changed
to ShowStart with an ActionItemAssociatedFile. That will not change the
normal output or behavior, but --json output will now include the key.
This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
|
|
|
{- Message that is displayed when starting to perform an action on
|
|
|
|
|
- something. The String is typically the name of the command or action
|
|
|
|
|
- being performed.
|
|
|
|
|
-}
|
|
|
|
|
data StartMessage
|
|
|
|
|
= StartMessage String ActionItem
|
|
|
|
|
| StartUsualMessages String ActionItem
|
finish CommandStart transition
The hoped for optimisation of CommandStart with -J did not materialize.
In fact, not runnign CommandStart in parallel is slower than -J3.
So, CommandStart are still run in parallel.
(The actual bad performance I've been seeing with -J in my big repo
has to do with building the remoteList.)
But, this is still progress toward making -J faster, because it gets rid
of the onlyActionOn roadblock in the way of making CommandCleanup jobs
run separate from CommandPerform jobs.
Added OnlyActionOn constructor for ActionItem which fixes the
onlyActionOn breakage in the last commit.
Made CustomOutput include an ActionItem, so even things using it can
specify OnlyActionOn.
In Command.Move and Command.Sync, there were CommandStarts that used
includeCommandAction, so output messages, which is no longer allowed.
Fixed by using startingCustomOutput, but that's still not quite right,
since it prevents message display for the includeCommandAction run
inside it too.
2019-06-12 13:23:26 +00:00
|
|
|
-- ^ Like StartMessage, but makes sure to enable usual message
|
|
|
|
|
-- display in case it was disabled by cmdnomessages.
|
2019-06-12 18:11:23 +00:00
|
|
|
| StartNoMessage ActionItem
|
|
|
|
|
-- ^ Starts, without displaying any message but also without
|
|
|
|
|
-- disabling display of any of the usual messages.
|
finish CommandStart transition
The hoped for optimisation of CommandStart with -J did not materialize.
In fact, not runnign CommandStart in parallel is slower than -J3.
So, CommandStart are still run in parallel.
(The actual bad performance I've been seeing with -J in my big repo
has to do with building the remoteList.)
But, this is still progress toward making -J faster, because it gets rid
of the onlyActionOn roadblock in the way of making CommandCleanup jobs
run separate from CommandPerform jobs.
Added OnlyActionOn constructor for ActionItem which fixes the
onlyActionOn breakage in the last commit.
Made CustomOutput include an ActionItem, so even things using it can
specify OnlyActionOn.
In Command.Move and Command.Sync, there were CommandStarts that used
includeCommandAction, so output messages, which is no longer allowed.
Fixed by using startingCustomOutput, but that's still not quite right,
since it prevents message display for the includeCommandAction run
inside it too.
2019-06-12 13:23:26 +00:00
|
|
|
| CustomOutput ActionItem
|
2019-06-12 18:11:23 +00:00
|
|
|
-- ^ Prevents any start, end, or other usual messages from
|
finish CommandStart transition
The hoped for optimisation of CommandStart with -J did not materialize.
In fact, not runnign CommandStart in parallel is slower than -J3.
So, CommandStart are still run in parallel.
(The actual bad performance I've been seeing with -J in my big repo
has to do with building the remoteList.)
But, this is still progress toward making -J faster, because it gets rid
of the onlyActionOn roadblock in the way of making CommandCleanup jobs
run separate from CommandPerform jobs.
Added OnlyActionOn constructor for ActionItem which fixes the
onlyActionOn breakage in the last commit.
Made CustomOutput include an ActionItem, so even things using it can
specify OnlyActionOn.
In Command.Move and Command.Sync, there were CommandStarts that used
includeCommandAction, so output messages, which is no longer allowed.
Fixed by using startingCustomOutput, but that's still not quite right,
since it prevents message display for the includeCommandAction run
inside it too.
2019-06-12 13:23:26 +00:00
|
|
|
-- being displayed, letting a command output its own custom format.
|
|
|
|
|
deriving (Show)
|
|
|
|
|
|
|
|
|
|
instance MkActionItem StartMessage where
|
|
|
|
|
mkActionItem (StartMessage _ ai) = ai
|
|
|
|
|
mkActionItem (StartUsualMessages _ ai) = ai
|
2019-06-12 18:11:23 +00:00
|
|
|
mkActionItem (StartNoMessage ai) = ai
|
finish CommandStart transition
The hoped for optimisation of CommandStart with -J did not materialize.
In fact, not runnign CommandStart in parallel is slower than -J3.
So, CommandStart are still run in parallel.
(The actual bad performance I've been seeing with -J in my big repo
has to do with building the remoteList.)
But, this is still progress toward making -J faster, because it gets rid
of the onlyActionOn roadblock in the way of making CommandCleanup jobs
run separate from CommandPerform jobs.
Added OnlyActionOn constructor for ActionItem which fixes the
onlyActionOn breakage in the last commit.
Made CustomOutput include an ActionItem, so even things using it can
specify OnlyActionOn.
In Command.Move and Command.Sync, there were CommandStarts that used
includeCommandAction, so output messages, which is no longer allowed.
Fixed by using startingCustomOutput, but that's still not quite right,
since it prevents message display for the includeCommandAction run
inside it too.
2019-06-12 13:23:26 +00:00
|
|
|
mkActionItem (CustomOutput ai) = ai
|
make CommandStart return a StartMessage
The goal is to be able to run CommandStart in the main thread when -J is
used, rather than unncessarily passing it off to a worker thread, which
incurs overhead that is signficant when the CommandStart is going to
quickly decide to stop.
To do that, the message it displays needs to be displayed in the worker
thread, after the CommandStart has run.
Also, the change will mean that CommandStart will no longer necessarily
run with the same Annex state as CommandPerform. While its docs already
said it should avoid modifying Annex state, I audited all the
CommandStart code as part of the conversion. (Note that CommandSeek
already sometimes runs with a different Annex state, and that has not been
a source of any problems, so I am not too worried that this change will
lead to breakage going forward.)
The only modification of Annex state I found was it calling
allowMessages in some Commands that default to noMessages. Dealt with
that by adding a startCustomOutput and a startingUsualMessages.
This lets a command start with noMessages and then select the output it
wants for each CommandStart.
One bit of breakage: onlyActionOn has been removed from commands that used it.
The plan is that, since a StartMessage contains an ActionItem,
when a Key can be extracted from that, the parallel job runner can
run onlyActionOn' automatically. Then commands won't need to worry about
this detail. Future work.
Otherwise, this was a fairly straightforward process of making each
CommandStart compile again. Hopefully other behavior changes were mostly
avoided.
In a few cases, a command had a CommandStart that called a CommandPerform
that then called showStart multiple times. I have collapsed those
down to a single start action. The main command to perhaps suffer from it
is Command.Direct, which used to show a start for each file, and no
longer does.
Another minor behavior change is that some commands used showStart
before, but had an associated file and a Key available, so were changed
to ShowStart with an ActionItemAssociatedFile. That will not change the
normal output or behavior, but --json output will now include the key.
This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
|
|
|
|
2011-10-30 03:48:46 +00:00
|
|
|
{- A command is defined by specifying these things. -}
|
2012-01-06 02:48:59 +00:00
|
|
|
data Command = Command
|
2015-07-10 17:49:46 +00:00
|
|
|
{ cmdcheck :: [CommandCheck] -- check stage
|
2012-09-16 00:46:38 +00:00
|
|
|
, cmdnocommit :: Bool -- don't commit journalled state changes
|
2016-01-20 18:07:13 +00:00
|
|
|
, cmdnomessages :: Bool -- don't output normal messages
|
2012-01-06 02:48:59 +00:00
|
|
|
, cmdname :: String
|
2015-07-08 20:58:54 +00:00
|
|
|
, cmdparamdesc :: CmdParamsDesc -- description of params for usage
|
2013-03-24 22:28:21 +00:00
|
|
|
, cmdsection :: CommandSection
|
2012-01-06 02:48:59 +00:00
|
|
|
, cmddesc :: String -- description of command for usage
|
2015-07-08 16:33:27 +00:00
|
|
|
, cmdparser :: CommandParser -- command line parser
|
2016-09-05 19:32:59 +00:00
|
|
|
, cmdglobaloptions :: [GlobalOption] -- additional global options
|
2015-07-08 19:39:05 +00:00
|
|
|
, cmdnorepo :: Maybe (Parser (IO ())) -- used when not in a repo
|
2012-01-06 02:48:59 +00:00
|
|
|
}
|
2011-10-30 03:48:46 +00:00
|
|
|
|
2015-07-08 16:33:27 +00:00
|
|
|
{- Command-line parameters, after the command is selected and options
|
|
|
|
|
- are parsed. -}
|
2013-11-30 19:18:40 +00:00
|
|
|
type CmdParams = [String]
|
|
|
|
|
|
2015-07-08 20:58:54 +00:00
|
|
|
type CmdParamsDesc = String
|
|
|
|
|
|
2011-10-30 03:48:46 +00:00
|
|
|
{- CommandCheck functions can be compared using their unique id. -}
|
|
|
|
|
instance Eq CommandCheck where
|
|
|
|
|
a == b = idCheck a == idCheck b
|
2012-02-16 20:53:44 +00:00
|
|
|
|
|
|
|
|
instance Eq Command where
|
|
|
|
|
a == b = cmdname a == cmdname b
|
|
|
|
|
|
2013-03-25 16:41:57 +00:00
|
|
|
{- Order commands by name. -}
|
2012-02-16 20:53:44 +00:00
|
|
|
instance Ord Command where
|
2013-03-25 16:41:57 +00:00
|
|
|
compare = comparing cmdname
|
2013-03-24 22:28:21 +00:00
|
|
|
|
|
|
|
|
{- The same sections are listed in doc/git-annex.mdwn -}
|
|
|
|
|
data CommandSection
|
|
|
|
|
= SectionCommon
|
|
|
|
|
| SectionSetup
|
|
|
|
|
| SectionMaintenance
|
|
|
|
|
| SectionQuery
|
2014-02-19 18:55:34 +00:00
|
|
|
| SectionMetaData
|
2013-03-24 22:28:21 +00:00
|
|
|
| SectionUtility
|
|
|
|
|
| SectionPlumbing
|
2014-08-01 16:49:26 +00:00
|
|
|
| SectionTesting
|
2013-03-24 22:28:21 +00:00
|
|
|
deriving (Eq, Ord, Enum, Bounded)
|
2013-03-25 14:23:05 +00:00
|
|
|
|
|
|
|
|
descSection :: CommandSection -> String
|
|
|
|
|
descSection SectionCommon = "Commonly used commands"
|
|
|
|
|
descSection SectionSetup = "Repository setup commands"
|
|
|
|
|
descSection SectionMaintenance = "Repository maintenance commands"
|
|
|
|
|
descSection SectionQuery = "Query commands"
|
2014-02-19 18:55:34 +00:00
|
|
|
descSection SectionMetaData = "Metadata commands"
|
2013-03-25 14:23:05 +00:00
|
|
|
descSection SectionUtility = "Utility commands"
|
|
|
|
|
descSection SectionPlumbing = "Plumbing commands"
|
2014-08-01 16:49:26 +00:00
|
|
|
descSection SectionTesting = "Testing commands"
|