Re: [RFA/Testsuite] help.exp

Re: [RFA/Testsuite] help.exp

[Michael Elizabeth Chastain]

Elena Zannoni writes:
> Oops, I committed the wildcard change already. 

My bad, I didn't read to the end of the thread before replying.

Any way that Fernando likes it, I like it.

Michael C

Target specific tests [Was: [RFA/Testsuite] help.exp]

[Fernando Nasser]

Michael Elizabeth Chastain wrote:
> 
> Elena Zannoni writes:
> > Oops, I committed the wildcard change already.
> 
> My bad, I didn't read to the end of the thread before replying.
> 
> Any way that Fernando likes it, I like it.
> 


Thanks for reviewing it Michael. 

I do not strictly prefer one way or the other 
(and I guess so is Michael Snyder).  I believe the
decision can be made based on the importance of testing
the specific target output.  That is why I've asked Elena
if she wanted to test the powerpc target output or just
prevent the test to fail (she is in better position to
decide).

I wish we had a cleaner way of adding target specific tests.
Maybe we can establish that a new test has to be created so
that the specific target output is tested and it would only
run for that target and produce unsupported for any other.
This brings a setup_supported (?) proc idea to mind.  Oh My!!!

Regards to all,
Fernando

-- 
Fernando Nasser
Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
2323 Yonge Street, Suite #300
Toronto, Ontario   M4P 2C9

Re: [RFA/Testsuite] help.exp

[Michael Elizabeth Chastain]

It looks good to me.  I prefer the target-sensitive test to a wildcard
expression, because the target-sensitive test expresses the success
conditions more precisely.

I proofread it, but I did not test it.

I recommend this patch for approval.

Michael C

2002-04-23  Elena Zannoni  <ezannoni@redhat.com>

	* gdb.base/help.exp: Change 'help status' to allow for PowerPC
	differences.

Re: [RFA/Testsuite] help.exp

[Elena Zannoni]

Michael Elizabeth Chastain writes:
 > It looks good to me.  I prefer the target-sensitive test to a wildcard
 > expression, because the target-sensitive test expresses the success
 > conditions more precisely.
 > 
 > I proofread it, but I did not test it.
 > 
 > I recommend this patch for approval.
 > 
 > Michael C
 > 
 > 2002-04-23  Elena Zannoni  <ezannoni@redhat.com>
 > 
 > 	* gdb.base/help.exp: Change 'help status' to allow for PowerPC
 > 	differences.

Oops, I committed the wildcard change already. 
Ok, I can change it again. Fernando?

Elena

[RFA/Testsuite] help.exp

[Elena Zannoni]

The recent changes to help.exp eliminating the wildcards introduced a few
failures on the powerpc.

How about this to deal with the difference in output for the powerpc
vs. other targets?

Elena


2002-04-23  Elena Zannoni  <ezannoni@redhat.com>

	* gdb.base/help.exp: Change 'help status' to allow for PowerPC
	differences.

Index: help.exp
===================================================================
RCS file: /cvs/uberbaum/gdb/testsuite/gdb.base/help.exp,v
retrieving revision 1.10
diff -u -p -r1.10 help.exp
--- help.exp	29 Mar 2002 01:19:11 -0000	1.10
+++ help.exp	23 Apr 2002 23:04:23 -0000
@@ -535,7 +535,13 @@ gdb_test "help source" "Read commands fr
 # test help stack
 gdb_test "help stack" "Examining the stack\..*\[\r\n\]+When the program being debugged stops, gdb selects the innermost frame\.\[\r\n\]+The commands below can be used to select other frames by number or address\.\[\r\n\]+List of commands:\[\r\n\]+backtrace -- Print backtrace of all stack frames\[\r\n\]+bt -- Print backtrace of all stack frames\[\r\n\]+down -- Select and print stack frame called by this one\[\r\n\]+frame -- Select and print a stack frame\[\r\n\]+return -- Make selected stack frame return to its caller\[\r\n\]+select-frame -- Select a stack frame without printing anything\[\r\n\]+up -- Select and print stack frame that called this one\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help stack"
 # test help status
-gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
+
+if [istarget "powerpc-*-*"] then {
+    gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+info powerpc -- Various POWERPC info specific commands\[\r\n\]+info powerpc altivec -- Display the contents of the AltiVec registers\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
+} else {
+    gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
+}
+
 # test help support
 # FIXME -- Ultrix hangs randomly on this very long output from gdb and
 # continues with its output only if something is sent to gdb.

Re: [RFA/Testsuite] help.exp

[Michael Snyder]

Elena Zannoni wrote:
> 
> The recent changes to help.exp eliminating the wildcards introduced a few
> failures on the powerpc.

Ooops, didn't mean to do  that.  ;-)

> How about this to deal with the difference in output for the powerpc
> vs. other targets?

OK with me (the author of the breakage).
Or you could just use a parenthesis to make the
powerpc command output optional...

Michael

> 2002-04-23  Elena Zannoni  <ezannoni@redhat.com>
> 
>         * gdb.base/help.exp: Change 'help status' to allow for PowerPC
>         differences.
> 
> Index: help.exp
> ===================================================================
> RCS file: /cvs/uberbaum/gdb/testsuite/gdb.base/help.exp,v
> retrieving revision 1.10
> diff -u -p -r1.10 help.exp
> --- help.exp    29 Mar 2002 01:19:11 -0000      1.10
> +++ help.exp    23 Apr 2002 23:04:23 -0000
> @@ -535,7 +535,13 @@ gdb_test "help source" "Read commands fr
>  # test help stack
>  gdb_test "help stack" "Examining the stack\..*\[\r\n\]+When the program being debugged stops, gdb selects the innermost frame\.\[\r\n\]+The commands below can be used to select other frames by number or address\.\[\r\n\]+List of commands:\[\r\n\]+backtrace -- Print backtrace of all stack frames\[\r\n\]+bt -- Print backtrace of all stack frames\[\r\n\]+down -- Select and print stack frame called by this one\[\r\n\]+frame -- Select and print a stack frame\[\r\n\]+return -- Make selected stack frame return to its caller\[\r\n\]+select-frame -- Select a stack frame without printing anything\[\r\n\]+up -- Select and print stack frame that called this one\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help stack"
>  # test help status
> -gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
> +
> +if [istarget "powerpc-*-*"] then {
> +    gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+info powerpc -- Various POWERPC info specific commands\[\r\n\]+info powerpc altivec -- Display the contents of the AltiVec registers\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
> +} else {
> +    gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
> +}
> +
>  # test help support
>  # FIXME -- Ultrix hangs randomly on this very long output from gdb and
>  # continues with its output only if something is sent to gdb.

Re: [RFA/Testsuite] help.exp

[Fernando Nasser]

Michael Snyder wrote:
> 
> Or you could just use a parenthesis to make the
> powerpc command output optional...
> 

Good idea, more specific than my suggested '.*'.  But we don't
need to be so stringent in this case.  As long as we see the 
generic ones the help is working.  

Unless Elena wants to test if these commands show up for the powerpc
targets help.  In this case, neither my suggestion or yours are 
checking for it. :-(

If that is the case, we will have to go with Elena's original patch.

Fernando

-- 
Fernando Nasser
Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
2323 Yonge Street, Suite #300
Toronto, Ontario   M4P 2C9

Re: [RFA/Testsuite] help.exp

[Elena Zannoni]

Fernando Nasser writes:
 > Michael Snyder wrote:
 > > 
 > > Or you could just use a parenthesis to make the
 > > powerpc command output optional...
 > > 
 > 
 > Good idea, more specific than my suggested '.*'.  But we don't
 > need to be so stringent in this case.  As long as we see the 
 > generic ones the help is working.  
 > 
 > Unless Elena wants to test if these commands show up for the powerpc
 > targets help.  In this case, neither my suggestion or yours are 
 > checking for it. :-(
 > 

Nah, I don't really want to be that specific.

 > If that is the case, we will have to go with Elena's original patch.
 > 
 
But you guys decide, if we got rid of the wildcards do we want them back?

Elena



> Fernando
 > 
 > -- 
 > Fernando Nasser
 > Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
 > 2323 Yonge Street, Suite #300
 > Toronto, Ontario   M4P 2C9

Re: [RFA/Testsuite] help.exp

[Michael Snyder]

Elena Zannoni wrote:
> 
> Fernando Nasser writes:
>  > Michael Snyder wrote:
>  > >
>  > > Or you could just use a parenthesis to make the
>  > > powerpc command output optional...
>  > >
>  >
>  > Good idea, more specific than my suggested '.*'.  But we don't
>  > need to be so stringent in this case.  As long as we see the
>  > generic ones the help is working.
>  >
>  > Unless Elena wants to test if these commands show up for the powerpc
>  > targets help.  In this case, neither my suggestion or yours are
>  > checking for it. :-(
>  >
> 
> Nah, I don't really want to be that specific.
> 
>  > If that is the case, we will have to go with Elena's original patch.
>  >
> 
> But you guys decide, if we got rid of the wildcards do we want them back?

I was only deleting them where I thought they were unnecessary.
In this case, I was wrong, 'cause I didn't know PPC had its own
variants.  So yes, it would be appropriate to put that one back.

Michael

Re: [RFA/Testsuite] help.exp

[Elena Zannoni]

Michael Snyder writes:
 > Elena Zannoni wrote:
 > > 
 > > Fernando Nasser writes:
 > >  > Michael Snyder wrote:
 > >  > >
 > >  > > Or you could just use a parenthesis to make the
 > >  > > powerpc command output optional...
 > >  > >
 > >  >
 > >  > Good idea, more specific than my suggested '.*'.  But we don't
 > >  > need to be so stringent in this case.  As long as we see the
 > >  > generic ones the help is working.
 > >  >
 > >  > Unless Elena wants to test if these commands show up for the powerpc
 > >  > targets help.  In this case, neither my suggestion or yours are
 > >  > checking for it. :-(
 > >  >
 > > 
 > > Nah, I don't really want to be that specific.
 > > 
 > >  > If that is the case, we will have to go with Elena's original patch.
 > >  >
 > > 
 > > But you guys decide, if we got rid of the wildcards do we want them back?
 > 
 > I was only deleting them where I thought they were unnecessary.
 > In this case, I was wrong, 'cause I didn't know PPC had its own
 > variants.  So yes, it would be appropriate to put that one back.
 > 
 > Michael

OK, I committed the following:

2002-04-23  Elena Zannoni  <ezannoni@redhat.com>

	* gdb.base/help.exp: Change 'help status' to allow for target
	dependent output differences.

Index: help.exp
===================================================================
RCS file: /cvs/uberbaum/gdb/testsuite/gdb.base/help.exp,v
retrieving revision 1.10
diff -u -p -r1.10 help.exp
--- help.exp	29 Mar 2002 01:19:11 -0000	1.10
+++ help.exp	24 Apr 2002 03:05:07 -0000
@@ -535,7 +535,9 @@ gdb_test "help source" "Read commands fr
 # test help stack
 gdb_test "help stack" "Examining the stack\..*\[\r\n\]+When the program being debugged stops, gdb selects the innermost frame\.\[\r\n\]+The commands below can be used to select other frames by number or address\.\[\r\n\]+List of commands:\[\r\n\]+backtrace -- Print backtrace of all stack frames\[\r\n\]+bt -- Print backtrace of all stack frames\[\r\n\]+down -- Select and print stack frame called by this one\[\r\n\]+frame -- Select and print a stack frame\[\r\n\]+return -- Make selected stack frame return to its caller\[\r\n\]+select-frame -- Select a stack frame without printing anything\[\r\n\]+up -- Select and print stack frame that called this one\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help stack"
 # test help status
-gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
+gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged.*\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
+
 # test help support
 # FIXME -- Ultrix hangs randomly on this very long output from gdb and
 # continues with its output only if something is sent to gdb.

Re: [RFA/Testsuite] help.exp

[Fernando Nasser]

Elena Zannoni wrote:
> 
> The recent changes to help.exp eliminating the wildcards introduced a few
> failures on the powerpc.
> 
> How about this to deal with the difference in output for the powerpc
> vs. other targets?
> 
> Elena
> 
Hi Elena,

I would rather have a '.*' in there instead of the test for a specific
target.

You don't need to resubmit: just add the '.*' and post the final
version.

Regards,
Fernando


> 2002-04-23  Elena Zannoni  <ezannoni@redhat.com>
> 
>         * gdb.base/help.exp: Change 'help status' to allow for PowerPC
>         differences.
> 
> Index: help.exp
> ===================================================================
> RCS file: /cvs/uberbaum/gdb/testsuite/gdb.base/help.exp,v
> retrieving revision 1.10
> diff -u -p -r1.10 help.exp
> --- help.exp    29 Mar 2002 01:19:11 -0000      1.10
> +++ help.exp    23 Apr 2002 23:04:23 -0000
> @@ -535,7 +535,13 @@ gdb_test "help source" "Read commands fr
>  # test help stack
>  gdb_test "help stack" "Examining the stack\..*\[\r\n\]+When the program being debugged stops, gdb selects the innermost frame\.\[\r\n\]+The commands below can be used to select other frames by number or address\.\[\r\n\]+List of commands:\[\r\n\]+backtrace -- Print backtrace of all stack frames\[\r\n\]+bt -- Print backtrace of all stack frames\[\r\n\]+down -- Select and print stack frame called by this one\[\r\n\]+frame -- Select and print a stack frame\[\r\n\]+return -- Make selected stack frame return to its caller\[\r\n\]+select-frame -- Select a stack frame without printing anything\[\r\n\]+up -- Select and print stack frame that called this one\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help stack"
>  # test help status
> -gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
> +
> +if [istarget "powerpc-*-*"] then {
> +    gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+info powerpc -- Various POWERPC info specific commands\[\r\n\]+info powerpc altivec -- Display the contents of the AltiVec registers\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
> +} else {
> +    gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
> +}
> +
>  # test help support
>  # FIXME -- Ultrix hangs randomly on this very long output from gdb and
>  # continues with its output only if something is sent to gdb.

-- 
Fernando Nasser
Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
2323 Yonge Street, Suite #300
Toronto, Ontario   M4P 2C9

Re: [RFA/Testsuite] help.exp

[Elena Zannoni]

Fernando Nasser writes:
 > Elena Zannoni wrote:
 > > 
 > > The recent changes to help.exp eliminating the wildcards introduced a few
 > > failures on the powerpc.
 > > 
 > > How about this to deal with the difference in output for the powerpc
 > > vs. other targets?
 > > 
 > > Elena
 > > 
 > Hi Elena,
 > 
 > I would rather have a '.*' in there instead of the test for a specific
 > target.
 > 

That is the way it was originally. That works fine with the powerpc.

 > You don't need to resubmit: just add the '.*' and post the final
 > version.
 > 

OK.

 > Regards,
 > Fernando
 > 
 > 
 > > 2002-04-23  Elena Zannoni  <ezannoni@redhat.com>
 > > 
 > >         * gdb.base/help.exp: Change 'help status' to allow for PowerPC
 > >         differences.
 > > 
 > > Index: help.exp
 > > ===================================================================
 > > RCS file: /cvs/uberbaum/gdb/testsuite/gdb.base/help.exp,v
 > > retrieving revision 1.10
 > > diff -u -p -r1.10 help.exp
 > > --- help.exp    29 Mar 2002 01:19:11 -0000      1.10
 > > +++ help.exp    23 Apr 2002 23:04:23 -0000
 > > @@ -535,7 +535,13 @@ gdb_test "help source" "Read commands fr
 > >  # test help stack
 > >  gdb_test "help stack" "Examining the stack\..*\[\r\n\]+When the program being debugged stops, gdb selects the innermost frame\.\[\r\n\]+The commands below can be used to select other frames by number or address\.\[\r\n\]+List of commands:\[\r\n\]+backtrace -- Print backtrace of all stack frames\[\r\n\]+bt -- Print backtrace of all stack frames\[\r\n\]+down -- Select and print stack frame called by this one\[\r\n\]+frame -- Select and print a stack frame\[\r\n\]+return -- Make selected stack frame return to its caller\[\r\n\]+select-frame -- Select a stack frame without printing anything\[\r\n\]+up -- Select and print stack frame that called this one\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help stack"
 > >  # test help status
 > > -gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
 > > +
 > > +if [istarget "powerpc-*-*"] then {
 > > +    gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+info powerpc -- Various POWERPC info specific commands\[\r\n\]+info powerpc altivec -- Display the contents of the AltiVec registers\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
 > > +} else {
 > > +    gdb_test "help status" "Status inquiries\.\[\r\n\]+List of commands:\[\r\n\]+info -- Generic command for showing things about the program being debugged\[\r\n\]+show -- Generic command for showing things about the debugger\[\r\n\]+Type \"help\" followed by command name for full documentation\.\[\r\n\]+Command name abbreviations are allowed if unambiguous\." "help status"
 > > +}
 > > +
 > >  # test help support
 > >  # FIXME -- Ultrix hangs randomly on this very long output from gdb and
 > >  # continues with its output only if something is sent to gdb.
 > 
 > -- 
 > Fernando Nasser
 > Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
 > 2323 Yonge Street, Suite #300
 > Toronto, Ontario   M4P 2C9