[Mono-bugs] [Bug 508896] New: simple commands

[bugzilla_noreply at novell.com]

http://bugzilla.novell.com/show_bug.cgi?id=508896


           Summary: simple commands
    Classification: Mono
           Product: Mono: Tools
           Version: unspecified
          Platform: Other
        OS/Version: Other
            Status: NEW
          Severity: Enhancement
          Priority: P5 - None
         Component: MonkeyWrench
        AssignedTo: rkvinge at novell.com
        ReportedBy: ajorgensen at novell.com
         QAContact: mono-bugs at lists.ximian.com
          Found By: ---


I'm finding that a lot of the scripts I am writing for running mono unit tests
are of the form:

cd [some directory]
[simple command]

Examples:

cd mono
make dist

cd mono/mcs/class/System.Core
make run-test PROFILE=net_2_0

It seems wrong to be writing a script for each action of this kind.  It would
be nicer if each step had a parameter for the directory to run in and the
simple command to run.

I can work around this by having a script that takes two parameters, the
directory to run in and the command to run, but it seems to me that should be
pushed up into the system instead of hacked in that way since the script would
be essentially empty.

Some other commands (like my svn-update script) are not nearly so simple so we
need to be able to do things the current way as well.

Maybe we could have the best of both worlds by letting the user mark a file
(script) as executable (we'd want non-executable files for things like the osc
config file) and instead of having an "executable" parameter and an "arguments"
parameter we could have just a "command" parameter and have the default command
be just "{0}".  Then I could change the above examples to:

directory: mono
command: make dist

directory: mono/mcs/class/System.Core
command: make run-test PROFILE=net_2_0

Another nice thing about this would be that it would put what's actually being
run right there in my face without having to dig down to the script.  It's also
more obvious to the untrained mind.

Some percentage of commands will be complex enough to merit a script but I
think enough of them will be simple that this change would be worth while.  One
motive for using simple commands like this is to split up the logs so that it's
easier to see what failed

The directory parameter has it's own merits aside from this and I will open
another enhancement req for that.  In fact if you only gave me that one feature
I could use the current design to do much the same thing but like this:

directory: mono/mcs/class/System.Core
executable: make
argument: run-test PROFILE=net_2_0

I still think having just a command and not an executable and an argument is
more obvious to the user though (and less clicks to get what I need).  Unless
you wanted to go the other way and have an array of arguments, but that makes
for complicated UI.

-- 
Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.