gdb/testsuite/sim: Remove redundant setting of timeout

  In the config/sim.exp file two functions are defined.  Both of these
functions define local timeout variables and then call gdb_expect,
which (through a call to get_largest_timeout) will find the local
definition of timeout.

However, both of these functions set the local timeout to some
arbitrary value and print a log message for this "new" timeout just
before returning.

As in both cases, the timeout is a local variable, this final setting
of the timeout has no effect and can be removed.

gdb/testsuite/ChangeLog:

	* config/sim.exp (gdb_target_sim): Remove redundant adjustment of
	local timeout variable before return.
	(gdb_load): Likewise.
---
 gdb/testsuite/ChangeLog      | 6 ++++++
 gdb/testsuite/config/sim.exp | 4 ----
 2 files changed, 6 insertions(+), 4 deletions(-)
  

On 2018-12-04 06:33, Andrew Burgess wrote:
> In the config/sim.exp file two functions are defined.  Both of these
> functions define local timeout variables and then call gdb_expect,
> which (through a call to get_largest_timeout) will find the local
> definition of timeout.
> 
> However, both of these functions set the local timeout to some
> arbitrary value and print a log message for this "new" timeout just
> before returning.
> 
> As in both cases, the timeout is a local variable, this final setting
> of the timeout has no effect and can be removed.

Hi Andrew,

Can you verify whether the remaining "set timeout" in those functions 
have any effect at all?  As you said, they are just local variables, so 
I don't expect them to influence the behavior of gdb_expect.  Either we 
need "global timeout", or we pass the timeout directly as an argument to 
gdb_expect (the latter sounds better).

Simon
  

On 12/04/2018 03:43 PM, Simon Marchi wrote:
> On 2018-12-04 06:33, Andrew Burgess wrote:
>> In the config/sim.exp file two functions are defined.  Both of these
>> functions define local timeout variables and then call gdb_expect,
>> which (through a call to get_largest_timeout) will find the local
>> definition of timeout.
>>
>> However, both of these functions set the local timeout to some
>> arbitrary value and print a log message for this "new" timeout just
>> before returning.
>>
>> As in both cases, the timeout is a local variable, this final setting
>> of the timeout has no effect and can be removed.
> 
> Hi Andrew,
> 
> Can you verify whether the remaining "set timeout" in those functions have any effect at all?  As you said, they are just local variables, so I don't expect them to influence the behavior of gdb_expect.  Either we need "global timeout", or we pass the timeout directly as an argument to gdb_expect (the latter sounds better).

Keep this in mind, from man expect:

       Expect  takes  a  rather  liberal view of scoping.  In particular, 
       variables read by commands specific to the Expect program will be sought
       first from the local scope, and if not found, in the global scope.  For
       example, this obviates the need to place "global timeout" in every procedure
       you write that uses expect.   On the  other hand, variables written are always
       in the local scope (unless a "global" command has been issued).  The most
       common problem this causes is when spawn is executed in a procedure.  Outside
       the procedure, spawn_id no longer exists, so the spawned process is no longer
       accessible simply because of scoping.  Add a "global spawn_id" to such a procedure.


Mimicking that behavior, gdb_test, gdb_test_multiple and gdb_expect pick the
local timeout variable in the caller via upvar.  E.g.:

proc gdb_test { args } {
    global gdb_prompt
    upvar timeout timeout

gdb_expect is a little more disguised, but it does the same, here,
in the get_largest_timeout path:

proc gdb_expect { args } {
...
    # A timeout argument takes precedence, otherwise of all the timeouts
    # select the largest.
    if [info exists atimeout] {
	set tmt $atimeout
    } else {
	set tmt [get_largest_timeout]
    }
...
}

and then get_largest_timeout does:

proc get_largest_timeout {} {
    upvar #0 timeout gtimeout
    upvar 2 timeout timeout
    ^^^^^^^^^^^^^^^^^^^^^^^
    ... 

Thanks,
Pedro Alves
  

On 2018-12-04 10:54, Pedro Alves wrote:
> On 12/04/2018 03:43 PM, Simon Marchi wrote:
>> On 2018-12-04 06:33, Andrew Burgess wrote:
>>> In the config/sim.exp file two functions are defined.  Both of these
>>> functions define local timeout variables and then call gdb_expect,
>>> which (through a call to get_largest_timeout) will find the local
>>> definition of timeout.
>>> 
>>> However, both of these functions set the local timeout to some
>>> arbitrary value and print a log message for this "new" timeout just
>>> before returning.
>>> 
>>> As in both cases, the timeout is a local variable, this final setting
>>> of the timeout has no effect and can be removed.
>> 
>> Hi Andrew,
>> 
>> Can you verify whether the remaining "set timeout" in those functions 
>> have any effect at all?  As you said, they are just local variables, 
>> so I don't expect them to influence the behavior of gdb_expect.  
>> Either we need "global timeout", or we pass the timeout directly as an 
>> argument to gdb_expect (the latter sounds better).
> 
> Keep this in mind, from man expect:
> 
>        Expect  takes  a  rather  liberal view of scoping.  In 
> particular,
>        variables read by commands specific to the Expect program will 
> be sought
>        first from the local scope, and if not found, in the global 
> scope.  For
>        example, this obviates the need to place "global timeout" in
> every procedure
>        you write that uses expect.   On the  other hand, variables
> written are always
>        in the local scope (unless a "global" command has been issued).  
> The most
>        common problem this causes is when spawn is executed in a
> procedure.  Outside
>        the procedure, spawn_id no longer exists, so the spawned
> process is no longer
>        accessible simply because of scoping.  Add a "global spawn_id"
> to such a procedure.
> 
> 
> Mimicking that behavior, gdb_test, gdb_test_multiple and gdb_expect 
> pick the
> local timeout variable in the caller via upvar.  E.g.:
> 
> proc gdb_test { args } {
>     global gdb_prompt
>     upvar timeout timeout
> 
> gdb_expect is a little more disguised, but it does the same, here,
> in the get_largest_timeout path:
> 
> proc gdb_expect { args } {
> ...
>     # A timeout argument takes precedence, otherwise of all the 
> timeouts
>     # select the largest.
>     if [info exists atimeout] {
> 	set tmt $atimeout
>     } else {
> 	set tmt [get_largest_timeout]
>     }
> ...
> }
> 
> and then get_largest_timeout does:
> 
> proc get_largest_timeout {} {
>     upvar #0 timeout gtimeout
>     upvar 2 timeout timeout
>     ^^^^^^^^^^^^^^^^^^^^^^^
>     ...

That's very confusing, to say the least.

Simon
  

On 12/04/2018 04:08 PM, Simon Marchi wrote:
> On 2018-12-04 10:54, Pedro Alves wrote:
>> On 12/04/2018 03:43 PM, Simon Marchi wrote:
>>> On 2018-12-04 06:33, Andrew Burgess wrote:
>>>> In the config/sim.exp file two functions are defined.  Both of these
>>>> functions define local timeout variables and then call gdb_expect,
>>>> which (through a call to get_largest_timeout) will find the local
>>>> definition of timeout.
>>>>
>>>> However, both of these functions set the local timeout to some
>>>> arbitrary value and print a log message for this "new" timeout just
>>>> before returning.
>>>>
>>>> As in both cases, the timeout is a local variable, this final setting
>>>> of the timeout has no effect and can be removed.
>>>
>>> Hi Andrew,
>>>
>>> Can you verify whether the remaining "set timeout" in those functions have any effect at all?  As you said, they are just local variables, so I don't expect them to influence the behavior of gdb_expect.  Either we need "global timeout", or we pass the timeout directly as an argument to gdb_expect (the latter sounds better).
>>
>> Keep this in mind, from man expect:
>>
>>        Expect  takes  a  rather  liberal view of scoping.  In particular,
>>        variables read by commands specific to the Expect program will be sought
>>        first from the local scope, and if not found, in the global scope.  For
>>        example, this obviates the need to place "global timeout" in
>> every procedure
>>        you write that uses expect.   On the  other hand, variables
>> written are always
>>        in the local scope (unless a "global" command has been issued).  The most
>>        common problem this causes is when spawn is executed in a
>> procedure.  Outside
>>        the procedure, spawn_id no longer exists, so the spawned
>> process is no longer
>>        accessible simply because of scoping.  Add a "global spawn_id"
>> to such a procedure.
>>
>>
>> Mimicking that behavior, gdb_test, gdb_test_multiple and gdb_expect pick the
>> local timeout variable in the caller via upvar.  E.g.:
>>
>> proc gdb_test { args } {
>>     global gdb_prompt
>>     upvar timeout timeout
>>
>> gdb_expect is a little more disguised, but it does the same, here,
>> in the get_largest_timeout path:
>>
>> proc gdb_expect { args } {
>> ...
>>     # A timeout argument takes precedence, otherwise of all the timeouts
>>     # select the largest.
>>     if [info exists atimeout] {
>>     set tmt $atimeout
>>     } else {
>>     set tmt [get_largest_timeout]
>>     }
>> ...
>> }
>>
>> and then get_largest_timeout does:
>>
>> proc get_largest_timeout {} {
>>     upvar #0 timeout gtimeout
>>     upvar 2 timeout timeout
>>     ^^^^^^^^^^^^^^^^^^^^^^^
>>     ...
> 
> That's very confusing, to say the least.

Don't shoot the messenger.  :-)

Thanks,
Pedro Alves
  

On 2018-12-04 11:11, Pedro Alves wrote:
> On 12/04/2018 04:08 PM, Simon Marchi wrote:
>> That's very confusing, to say the least.
> 
> Don't shoot the messenger.  :-)

Hehe, of course.

In light of this information, I think Andrew's patch is fine.  Do you?

Simon
  

On 12/04/2018 04:15 PM, Simon Marchi wrote:
> On 2018-12-04 11:11, Pedro Alves wrote:
>> On 12/04/2018 04:08 PM, Simon Marchi wrote:
>>> That's very confusing, to say the least.
>>
>> Don't shoot the messenger.  :-)
> 
> Hehe, of course.
> 
> In light of this information, I think Andrew's patch is fine.  Do you?
Sort of.  At least with the removing the tail "set timeout" part,
I agree it's not doing anything.

As for the verbose call, we print "Timeout is now ..." messages
in a lot of places, and if you're looking at the log, I think
seeing a "Timeout is now ..." indication without seeing it changed
again reads like the timeout was never restored...

That's a preexisting problem, of course, since currently
we give the impression that we actually changed the timeout
at the end of the function but we actually didn't...

Still, IMHO, one of these would be a better change:

 a) - remove the initial verbose call too, or,
 b) - add "global timeout" at the start of the function, and restore
      the on-entry value on exit.  That way both "Timeout is now ..."
      messages will be truthful.  This is what e.g.,
      testsuiteconfig/sid.exp does.

In either case, there will be no imbalance in the verbose output.

Thanks,
Pedro Alves