kvm.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* Re: [KVM_AUTOTEST][RFC] pre_command chaining
       [not found] <1691222823.286001247239165845.JavaMail.root@zmail05.collab.prod.int.phx2.redhat.com>
@ 2009-07-10 15:27 ` Michael Goldish
  2009-07-13  7:40   ` Lukáš Doktor
  0 siblings, 1 reply; 3+ messages in thread
From: Michael Goldish @ 2009-07-10 15:27 UTC (permalink / raw)
  To: Lukáš Doktor; +Cc: KVM list


----- "Lukáš Doktor" <ldoktor@redhat.com> wrote:

> Hi,
> 
> the way how kvm_autotest currently handle pre_command/post_command it
> don't allow to specify more than one command. BASH can handle this 
> itself with a small change in the framework , as shown in the
> attachment.

Why do you say the framework doesn't allow chaining pre_commands?
What's wrong with:
pre_command = "command0"
pre_command += " && command1"
pre_command += " && command2"

> In .cfg file we just change variable from:
>   pre_command = "command"
> to:
>   pre_commane += "command &&"
> produce:
>   $(command && true)
> 
> Framework adds the last command true, which enclose whole command.
> This 
> way we can chain infinite pre/post_commands without losing the return
> 
> value (if something go wrong, other commands are not executed and
> return 
> value is preserve.
> 
> example:
> in cfg:
>   pre_command += "echo A &&"
>   pre_command += "echo B &&"
>   pre_command += "echo C &&"
> framework params.get("pre_command"):
>   "echo A && echo B && echo C &&"
> framework process_command execute on the host:
>   "echo A && echo B && echo C && true"
> 
> regards,
> Lukáš Doktor

In any case, the proposed solution does not allow the user to use
pre_command in the most straightforward way:
pre_command = "command"
because that would get translated into:
"command true"
So the user must append && to the command, which makes little sense.

There could be other solutions, like

1. Specifying "pre_command = true" at the top of the config file, and
then using:
pre_command += " && command0"
pre_command += " && command1"

"pre_command = command" will also work fine in this case.

2. Removing the final "&&" from the command, if any, so that if the
user enters:
pre_command = "command0 &&"
pre_command += "command1 &&"
the framework will run:
"command0 && command1" instead of "command0 && command1 &&".

In any case, can you provide an example where it's impossible or
difficult to do command chaining without changing the framework?

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [KVM_AUTOTEST][RFC] pre_command chaining
  2009-07-10 15:27 ` [KVM_AUTOTEST][RFC] pre_command chaining Michael Goldish
@ 2009-07-13  7:40   ` Lukáš Doktor
  0 siblings, 0 replies; 3+ messages in thread
From: Lukáš Doktor @ 2009-07-13  7:40 UTC (permalink / raw)
  To: Michael Goldish; +Cc: KVM list

Hi Michael,

you are right, it is possible. But if I specify "pre_command = true" at 
the top of my config file, this command will be executed even if no 
additional command is added into the queue. (tests with pre_commands are 
not selected)

That is my reason why I'd like to see this two lines change into the 
framework.

Still you are right that it's basically a "cosmetic" modification  for 
simplification the config file.

Dne 10.7.2009 17:27, Michael Goldish napsal(a):
> ----- "Lukáš Doktor"<ldoktor@redhat.com>  wrote:
>
>> Hi,
>>
>> the way how kvm_autotest currently handle pre_command/post_command it
>> don't allow to specify more than one command. BASH can handle this
>> itself with a small change in the framework , as shown in the
>> attachment.
>
> Why do you say the framework doesn't allow chaining pre_commands?
> What's wrong with:
> pre_command = "command0"
> pre_command += "&&  command1"
> pre_command += "&&  command2"
>
>> In .cfg file we just change variable from:
>>    pre_command = "command"
>> to:
>>    pre_commane += "command&&"
>> produce:
>>    $(command&&  true)
>>
>> Framework adds the last command true, which enclose whole command.
>> This
>> way we can chain infinite pre/post_commands without losing the return
>>
>> value (if something go wrong, other commands are not executed and
>> return
>> value is preserve.
>>
>> example:
>> in cfg:
>>    pre_command += "echo A&&"
>>    pre_command += "echo B&&"
>>    pre_command += "echo C&&"
>> framework params.get("pre_command"):
>>    "echo A&&  echo B&&  echo C&&"
>> framework process_command execute on the host:
>>    "echo A&&  echo B&&  echo C&&  true"
>>
>> regards,
>> Lukáš Doktor
>
> In any case, the proposed solution does not allow the user to use
> pre_command in the most straightforward way:
> pre_command = "command"
> because that would get translated into:
> "command true"
> So the user must append&&  to the command, which makes little sense.
>
> There could be other solutions, like
>
> 1. Specifying "pre_command = true" at the top of the config file, and
> then using:
> pre_command += "&&  command0"
> pre_command += "&&  command1"
>
> "pre_command = command" will also work fine in this case.
>
> 2. Removing the final "&&" from the command, if any, so that if the
> user enters:
> pre_command = "command0&&"
> pre_command += "command1&&"
> the framework will run:
> "command0&&  command1" instead of "command0&&  command1&&".
>
> In any case, can you provide an example where it's impossible or
> difficult to do command chaining without changing the framework?
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html


^ permalink raw reply	[flat|nested] 3+ messages in thread

* [KVM_AUTOTEST][RFC] pre_command chaining
@ 2009-07-10 11:35 Lukáš Doktor
  0 siblings, 0 replies; 3+ messages in thread
From: Lukáš Doktor @ 2009-07-10 11:35 UTC (permalink / raw)
  To: KVM list

[-- Attachment #1: Type: text/plain, Size: 875 bytes --]

Hi,

the way how kvm_autotest currently handle pre_command/post_command it 
don't allow to specify more than one command. BASH can handle this 
itself with a small change in the framework , as shown in the attachment.

In .cfg file we just change variable from:
  pre_command = "command"
to:
  pre_commane += "command &&"
produce:
  $(command && true)

Framework adds the last command true, which enclose whole command. This 
way we can chain infinite pre/post_commands without losing the return 
value (if something go wrong, other commands are not executed and return 
value is preserve.

example:
in cfg:
  pre_command += "echo A &&"
  pre_command += "echo B &&"
  pre_command += "echo C &&"
framework params.get("pre_command"):
  "echo A && echo B && echo C &&"
framework process_command execute on the host:
  "echo A && echo B && echo C && true"

regards,
Lukáš Doktor

[-- Attachment #2: command_chaining.patch --]
[-- Type: text/plain, Size: 1168 bytes --]

diff -Narup kvm-autotest/client/tests/kvm/kvm_preprocessing.py kvm-autotest-new/client/tests/kvm/kvm_preprocessing.py
--- kvm-autotest/client/tests/kvm/kvm_preprocessing.py	2009-07-08 08:31:01.492284501 +0200
+++ kvm-autotest-new/client/tests/kvm/kvm_preprocessing.py	2009-07-10 13:18:35.407285172 +0200
@@ -229,7 +229,8 @@ def preprocess(test, params, env):
 
     #execute any pre_commands
     if params.get("pre_command"):
-        process_command(test, params, env, params.get("pre_command"),
+        process_command(test, params, env, 
+                        (params.get("pre_command") + " true"),
                         params.get("pre_command_timeout"),
                         params.get("pre_command_noncritical"))
 
@@ -287,7 +288,8 @@ def postprocess(test, params, env):
 
     #execute any post_commands
     if params.get("post_command"):
-        process_command(test, params, env, params.get("post_command"),
+        process_command(test, params, env,
+                        (params.get("post_command") + " true"),
                         params.get("post_command_timeout"),
                         params.get("post_command_noncritical"))
 

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2009-07-13  7:41 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <1691222823.286001247239165845.JavaMail.root@zmail05.collab.prod.int.phx2.redhat.com>
2009-07-10 15:27 ` [KVM_AUTOTEST][RFC] pre_command chaining Michael Goldish
2009-07-13  7:40   ` Lukáš Doktor
2009-07-10 11:35 Lukáš Doktor

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).