RFA: parse output from `info sources' one filename at a time

RFA: parse output from `info sources' one filename at a time

[Jim Blandy]

This eliminates an `ERROR' when the test is run under Linux.  The
output from `info sources' includes all the shared library source
files, making it so long that it overflows Expect's buffer.

2002-02-13  Jim Blandy  <jimb@redhat.com>

	* gdb.asm/asm-source.exp: Parse the output from `info sources' one
	filename at a time, and watch for the ones we want to see.

Index: gdb/testsuite/gdb.asm/asm-source.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.asm/asm-source.exp,v
retrieving revision 1.17
diff -c -r1.17 asm-source.exp
*** gdb/testsuite/gdb.asm/asm-source.exp	2002/01/18 00:13:00	1.17
--- gdb/testsuite/gdb.asm/asm-source.exp	2002/02/14 00:05:30
***************
*** 202,211 ****
  	"Current source file is .*asmsrc2.s.*Source language is asm.*" \
  	"info source asmsrc2.s"
  
! # Try 'info sources'
! gdb_test "info sources" \
! 	"Source files .*asmsrc\[12\].s.*asmsrc\[12\].s.*" \
! 	"info sources"
  
  # Try 'info line'
  gdb_test "info line" \
--- 202,238 ----
  	"Current source file is .*asmsrc2.s.*Source language is asm.*" \
  	"info source asmsrc2.s"
  
! # Try 'info sources'.  This can produce a lot of output on systems
! # with dynamic linking, where the system's shared libc was compiled
! # with debugging info; for example, on Linux, this produces 47kb of
! # output.  So we consume it as we go.
! send_gdb "info sources\n"
! set seen_asmsrc_1 0
! set seen_asmsrc_2 0
! gdb_expect {
!     -re "^\[^,\]*asmsrc1.s(, |\[\r\n\]+)" {
!         set seen_asmsrc_1 1
!         exp_continue
!     }
!     -re "^\[^,\]*asmsrc2.s(, |\[\r\n\]+)" {
!         set seen_asmsrc_2 1
!         exp_continue
!     }
!     -re ", " { 
!         exp_continue
!     }
!     -re "$gdb_prompt $" {
!         if {$seen_asmsrc_1 && $seen_asmsrc_2} {
!             pass "info sources"
!         } else {
!             fail "info sources"
!         }
!     }
!     timeout {
!             fail "info sources (timeout)"
!     }
! }
!         
  
  # Try 'info line'
  gdb_test "info line" \

Re: RFA: parse output from `info sources' one filename at a time

[Michael Snyder]

Jim Blandy wrote:
> 
> This eliminates an `ERROR' when the test is run under Linux.  The
> output from `info sources' includes all the shared library source
> files, making it so long that it overflows Expect's buffer.

Not the maintainer, but I recommend acceptance.


> 
> 2002-02-13  Jim Blandy  <jimb@redhat.com>
> 
>         * gdb.asm/asm-source.exp: Parse the output from `info sources' one
>         filename at a time, and watch for the ones we want to see.
> 
> Index: gdb/testsuite/gdb.asm/asm-source.exp
> ===================================================================
> RCS file: /cvs/src/src/gdb/testsuite/gdb.asm/asm-source.exp,v
> retrieving revision 1.17
> diff -c -r1.17 asm-source.exp
> *** gdb/testsuite/gdb.asm/asm-source.exp        2002/01/18 00:13:00     1.17
> --- gdb/testsuite/gdb.asm/asm-source.exp        2002/02/14 00:05:30
> ***************
> *** 202,211 ****
>         "Current source file is .*asmsrc2.s.*Source language is asm.*" \
>         "info source asmsrc2.s"
> 
> ! # Try 'info sources'
> ! gdb_test "info sources" \
> !       "Source files .*asmsrc\[12\].s.*asmsrc\[12\].s.*" \
> !       "info sources"
> 
>   # Try 'info line'
>   gdb_test "info line" \
> --- 202,238 ----
>         "Current source file is .*asmsrc2.s.*Source language is asm.*" \
>         "info source asmsrc2.s"
> 
> ! # Try 'info sources'.  This can produce a lot of output on systems
> ! # with dynamic linking, where the system's shared libc was compiled
> ! # with debugging info; for example, on Linux, this produces 47kb of
> ! # output.  So we consume it as we go.
> ! send_gdb "info sources\n"
> ! set seen_asmsrc_1 0
> ! set seen_asmsrc_2 0
> ! gdb_expect {
> !     -re "^\[^,\]*asmsrc1.s(, |\[\r\n\]+)" {
> !         set seen_asmsrc_1 1
> !         exp_continue
> !     }
> !     -re "^\[^,\]*asmsrc2.s(, |\[\r\n\]+)" {
> !         set seen_asmsrc_2 1
> !         exp_continue
> !     }
> !     -re ", " {
> !         exp_continue
> !     }
> !     -re "$gdb_prompt $" {
> !         if {$seen_asmsrc_1 && $seen_asmsrc_2} {
> !             pass "info sources"
> !         } else {
> !             fail "info sources"
> !         }
> !     }
> !     timeout {
> !             fail "info sources (timeout)"
> !     }
> ! }
> !
> 
>   # Try 'info line'
>   gdb_test "info line" \

Re: RFA: parse output from `info sources' one filename at a time

[Fernando Nasser]

Michael Snyder wrote:
> 
> Jim Blandy wrote:
> >
> > This eliminates an `ERROR' when the test is run under Linux.  The
> > output from `info sources' includes all the shared library source
> > files, making it so long that it overflows Expect's buffer.
> 
> Not the maintainer, but I recommend acceptance.
> 

We don't have a co-maintainer for testsuite/gdb.asm...

Who was it that was fixing it all ove and adding lots
of things lately?  Would be a good candidate for a "volunteer" ;-)


Thanks for looking into it Michael.

Please check it in Jim, and thanks for the patch.

Regards to all,
Fernando




> >
> > 2002-02-13  Jim Blandy  <jimb@redhat.com>
> >
> >         * gdb.asm/asm-source.exp: Parse the output from `info sources' one
> >         filename at a time, and watch for the ones we want to see.
> >
> > Index: gdb/testsuite/gdb.asm/asm-source.exp
> > ===================================================================
> > RCS file: /cvs/src/src/gdb/testsuite/gdb.asm/asm-source.exp,v
> > retrieving revision 1.17
> > diff -c -r1.17 asm-source.exp
> > *** gdb/testsuite/gdb.asm/asm-source.exp        2002/01/18 00:13:00     1.17
> > --- gdb/testsuite/gdb.asm/asm-source.exp        2002/02/14 00:05:30
> > ***************
> > *** 202,211 ****
> >         "Current source file is .*asmsrc2.s.*Source language is asm.*" \
> >         "info source asmsrc2.s"
> >
> > ! # Try 'info sources'
> > ! gdb_test "info sources" \
> > !       "Source files .*asmsrc\[12\].s.*asmsrc\[12\].s.*" \
> > !       "info sources"
> >
> >   # Try 'info line'
> >   gdb_test "info line" \
> > --- 202,238 ----
> >         "Current source file is .*asmsrc2.s.*Source language is asm.*" \
> >         "info source asmsrc2.s"
> >
> > ! # Try 'info sources'.  This can produce a lot of output on systems
> > ! # with dynamic linking, where the system's shared libc was compiled
> > ! # with debugging info; for example, on Linux, this produces 47kb of
> > ! # output.  So we consume it as we go.
> > ! send_gdb "info sources\n"
> > ! set seen_asmsrc_1 0
> > ! set seen_asmsrc_2 0
> > ! gdb_expect {
> > !     -re "^\[^,\]*asmsrc1.s(, |\[\r\n\]+)" {
> > !         set seen_asmsrc_1 1
> > !         exp_continue
> > !     }
> > !     -re "^\[^,\]*asmsrc2.s(, |\[\r\n\]+)" {
> > !         set seen_asmsrc_2 1
> > !         exp_continue
> > !     }
> > !     -re ", " {
> > !         exp_continue
> > !     }
> > !     -re "$gdb_prompt $" {
> > !         if {$seen_asmsrc_1 && $seen_asmsrc_2} {
> > !             pass "info sources"
> > !         } else {
> > !             fail "info sources"
> > !         }
> > !     }
> > !     timeout {
> > !             fail "info sources (timeout)"
> > !     }
> > ! }
> > !
> >
> >   # Try 'info line'
> >   gdb_test "info line" \

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

Re: RFA: parse output from `info sources' one filename at a time

[Michael Snyder]

Fernando Nasser wrote:
> 
> Michael Snyder wrote:
> >
> > Jim Blandy wrote:
> > >
> > > This eliminates an `ERROR' when the test is run under Linux.  The
> > > output from `info sources' includes all the shared library source
> > > files, making it so long that it overflows Expect's buffer.
> >
> > Not the maintainer, but I recommend acceptance.
> >
> 
> We don't have a co-maintainer for testsuite/gdb.asm...
> 
> Who was it that was fixing it all ove and adding lots
> of things lately? 

That would be me.

> Would be a good candidate for a "volunteer" ;-)

Yep.  That's why I looked at this patch.
We can call me a co-maintainer for gdb.asm if we like.

> Thanks for looking into it Michael.
> 
> Please check it in Jim, and thanks for the patch.
> 
> Regards to all,
> Fernando
> 
> > >
> > > 2002-02-13  Jim Blandy  <jimb@redhat.com>
> > >
> > >         * gdb.asm/asm-source.exp: Parse the output from `info sources' one
> > >         filename at a time, and watch for the ones we want to see.
> > >
> > > Index: gdb/testsuite/gdb.asm/asm-source.exp
> > > ===================================================================
> > > RCS file: /cvs/src/src/gdb/testsuite/gdb.asm/asm-source.exp,v
> > > retrieving revision 1.17
> > > diff -c -r1.17 asm-source.exp
> > > *** gdb/testsuite/gdb.asm/asm-source.exp        2002/01/18 00:13:00     1.17
> > > --- gdb/testsuite/gdb.asm/asm-source.exp        2002/02/14 00:05:30
> > > ***************
> > > *** 202,211 ****
> > >         "Current source file is .*asmsrc2.s.*Source language is asm.*" \
> > >         "info source asmsrc2.s"
> > >
> > > ! # Try 'info sources'
> > > ! gdb_test "info sources" \
> > > !       "Source files .*asmsrc\[12\].s.*asmsrc\[12\].s.*" \
> > > !       "info sources"
> > >
> > >   # Try 'info line'
> > >   gdb_test "info line" \
> > > --- 202,238 ----
> > >         "Current source file is .*asmsrc2.s.*Source language is asm.*" \
> > >         "info source asmsrc2.s"
> > >
> > > ! # Try 'info sources'.  This can produce a lot of output on systems
> > > ! # with dynamic linking, where the system's shared libc was compiled
> > > ! # with debugging info; for example, on Linux, this produces 47kb of
> > > ! # output.  So we consume it as we go.
> > > ! send_gdb "info sources\n"
> > > ! set seen_asmsrc_1 0
> > > ! set seen_asmsrc_2 0
> > > ! gdb_expect {
> > > !     -re "^\[^,\]*asmsrc1.s(, |\[\r\n\]+)" {
> > > !         set seen_asmsrc_1 1
> > > !         exp_continue
> > > !     }
> > > !     -re "^\[^,\]*asmsrc2.s(, |\[\r\n\]+)" {
> > > !         set seen_asmsrc_2 1
> > > !         exp_continue
> > > !     }
> > > !     -re ", " {
> > > !         exp_continue
> > > !     }
> > > !     -re "$gdb_prompt $" {
> > > !         if {$seen_asmsrc_1 && $seen_asmsrc_2} {
> > > !             pass "info sources"
> > > !         } else {
> > > !             fail "info sources"
> > > !         }
> > > !     }
> > > !     timeout {
> > > !             fail "info sources (timeout)"
> > > !     }
> > > ! }
> > > !
> > >
> > >   # Try 'info line'
> > >   gdb_test "info line" \
> 
> --
> Fernando Nasser
> Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
> 2323 Yonge Street, Suite #300
> Toronto, Ontario   M4P 2C9

Re: RFA: parse output from `info sources' one filename at a time

[Fernando Nasser]

Michael Snyder wrote:
> 
> Yep.  That's why I looked at this patch.
> We can call me a co-maintainer for gdb.asm if we like.
> 

Great!  Thanks!

Please add a line for gdb.asm with your name on it as an obvious
fix to MAINTAINERS.

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

Old business: gdb.asm maintainership

[Michael Snyder]

[-- Attachment #1: Type: text/plain, Size: 418 bytes --]

Michael Snyder wrote:
> 
> Fernando Nasser wrote:
> >
> > We don't have a co-maintainer for testsuite/gdb.asm...
> >
> > Who was it that was fixing it all ove and adding lots
> > of things lately?
> 
> That would be me.
> 
> > Would be a good candidate for a "volunteer" ;-)
> 
> Yep.  That's why I looked at this patch.
> We can call me a co-maintainer for gdb.asm if we like.

I guess it's time I made this official:

[-- Attachment #2: asm --]
[-- Type: text/plain, Size: 116 bytes --]

2002-04-17  Michael Snyder  <msnyder@redhat.com>

	* MAINTAINERS: Add myself as co-maintainer of testsuite/gdb.asm.

Re: Old business: gdb.asm maintainership

[Fernando Nasser]

Thanks Michael.

Regards,
Fernando

Michael Snyder wrote:
> 
> Michael Snyder wrote:
> >
> > Fernando Nasser wrote:
> > >
> > > We don't have a co-maintainer for testsuite/gdb.asm...
> > >
> > > Who was it that was fixing it all ove and adding lots
> > > of things lately?
> >
> > That would be me.
> >
> > > Would be a good candidate for a "volunteer" ;-)
> >
> > Yep.  That's why I looked at this patch.
> > We can call me a co-maintainer for gdb.asm if we like.
> 
> I guess it's time I made this official:
> 
>   ------------------------------------------------------------------------
> 2002-04-17  Michael Snyder  <msnyder@redhat.com>
> 
>         * MAINTAINERS: Add myself as co-maintainer of testsuite/gdb.asm.

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

Re: RFA: parse output from `info sources' one filename at a time

[Jim Blandy]

I've committed this.

Jim Blandy <jimb@cygnus.com> writes:

> This eliminates an `ERROR' when the test is run under Linux.  The
> output from `info sources' includes all the shared library source
> files, making it so long that it overflows Expect's buffer.
> 
> 2002-02-13  Jim Blandy  <jimb@redhat.com>
> 
> 	* gdb.asm/asm-source.exp: Parse the output from `info sources' one
> 	filename at a time, and watch for the ones we want to see.

Re: RFA: parse output from `info sources' one filename at a time

[Michael Elizabeth Chastain]

Jim Blandy writes:
> This eliminates an `ERROR' when the test is run under Linux.  The
> output from `info sources' includes all the shared library source
> files, making it so long that it overflows Expect's buffer.

Interestingly, I don't get any ERROR from gdb.asm/asm-source.exp
on native i686-pc-linux-gnu.

My test script may be different from yours.  More likely, my
toolchain may be different from yours.  I am using stock tcl 8.3.4,
expect 5.33.0, dejagnu 1.4.2 compiled from source (not whatever
version comes with my Red Hat Linux).

Don't get me wrong, I'm all in favor of the patch.  It's good for the
testsuite to be "liberal in what it expects".  I just want to note that
there are several different versions of expect and dejagnu in the
world and their differences are materializing in test results.

I have a side project to package up my "standard gnu toolchain"
scripts.  They start by downloading 100 megabytes of pristine tarballs
and then build from scratch.

Michael C

Re: RFA: parse output from `info sources' one filename at a time

[Daniel Jacobowitz]

On Thu, Feb 21, 2002 at 03:36:55PM -0600, Michael Elizabeth Chastain wrote:
> Jim Blandy writes:
> > This eliminates an `ERROR' when the test is run under Linux.  The
> > output from `info sources' includes all the shared library source
> > files, making it so long that it overflows Expect's buffer.
> 
> Interestingly, I don't get any ERROR from gdb.asm/asm-source.exp
> on native i686-pc-linux-gnu.
> 
> My test script may be different from yours.  More likely, my
> toolchain may be different from yours.  I am using stock tcl 8.3.4,
> expect 5.33.0, dejagnu 1.4.2 compiled from source (not whatever
> version comes with my Red Hat Linux).
> 
> Don't get me wrong, I'm all in favor of the patch.  It's good for the
> testsuite to be "liberal in what it expects".  I just want to note that
> there are several different versions of expect and dejagnu in the
> world and their differences are materializing in test results.

I'm not arguing that expect/dejagnu version skew hasn't caused problems
in the past, but that's unrelated to this.  You don't get the error
because your C library does not contain debug info; no standard
desktop/server distribution leaves it there, generally.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: RFA: parse output from `info sources' one filename at a time

[Michael Elizabeth Chastain]

Daniel Jacobowitz write:
> I'm not arguing that expect/dejagnu version skew hasn't caused problems
> in the past, but that's unrelated to this.  You don't get the error
> because your C library does not contain debug info; no standard
> desktop/server distribution leaves it there, generally.

Oh, foo!  You are right, the C library is another free variable in
the testing system.

I wonder if I'll ever get my arms around the entire set of variables
for a reproducible test run.

Sad but enlightened,

Michael C