Insight 5.0 bug?

Insight 5.0 bug?

[Stephen Smith]

The inight 5.0 debugger is having a hard time finding and loading the the source code for
a currently running program the is running remote.

The 20010306 snapshot works.

Was this a documented error in the 5.0 release ( the currrently shiping version with cygwin)

Spec

     host=i686-pc-cygwin
     target=powerpc-motorola-elf

sps

-------------------- Stack trace of Insight 5.0 ---------------------------------
couldn't stat "/htc/DEOSExamples/threads/threads.cpp": no such file or directory
    while executing
"file mtime $f"
    (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::_mtime_changed" body line 7)
    invoked from within
"_mtime_changed $filename"
    (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::FillSource" body line 10)
    invoked from within
"FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib"
    ("SOURCE" arm line 2)
    invoked from within
"switch $current(mode) {
    SOURCE {
      FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib
    }
    ASSEMBLY {
      FillAssembly..."
    (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::location" body line 26)
    invoked from within
"$twin location $tag $name $funcname $line $addr $pc_addr $lib"
    (object "::.srcwin0.srcwin" method "::SrcWin::location" body line 52)
    invoked from within
"location BROWSE_TAG [list $val "" $full 0 0 0 {}]"
    (object "::.srcwin0.srcwin" method "::SrcWin::_name" body line 22)
    invoked from within
"::.srcwin0.srcwin _name .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
    (in namespace inscope "::SrcWin" script line 1)
    invoked from within
"namespace inscope ::SrcWin {::.srcwin0.srcwin _name} .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
    ("after" script)errorCode is POSIX ENOENT {no such file or directory}

Re: Insight 5.0 bug?

[Fernando Nasser]

Yes,

This was a know bug in 5.0 that has been fixed in more recent sources.

Maybe Cygwin will have to go with a GDB snapshot instead of 5.0 (there
is also some reported difficulties with ARM/Angel debugging I believe).

Fernando


Stephen Smith wrote:
> 
> The inight 5.0 debugger is having a hard time finding and loading the the source code for
> a currently running program the is running remote.
> 
> The 20010306 snapshot works.
> 
> Was this a documented error in the 5.0 release ( the currrently shiping version with cygwin)
> 
> Spec
> 
>      host=i686-pc-cygwin
>      target=powerpc-motorola-elf
> 
> sps
> 
> -------------------- Stack trace of Insight 5.0 ---------------------------------
> couldn't stat "/htc/DEOSExamples/threads/threads.cpp": no such file or directory
>     while executing
> "file mtime $f"
>     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::_mtime_changed" body line 7)
>     invoked from within
> "_mtime_changed $filename"
>     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::FillSource" body line 10)
>     invoked from within
> "FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib"
>     ("SOURCE" arm line 2)
>     invoked from within
> "switch $current(mode) {
>     SOURCE {
>       FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib
>     }
>     ASSEMBLY {
>       FillAssembly..."
>     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::location" body line 26)
>     invoked from within
> "$twin location $tag $name $funcname $line $addr $pc_addr $lib"
>     (object "::.srcwin0.srcwin" method "::SrcWin::location" body line 52)
>     invoked from within
> "location BROWSE_TAG [list $val "" $full 0 0 0 {}]"
>     (object "::.srcwin0.srcwin" method "::SrcWin::_name" body line 22)
>     invoked from within
> "::.srcwin0.srcwin _name .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
>     (in namespace inscope "::SrcWin" script line 1)
>     invoked from within
> "namespace inscope ::SrcWin {::.srcwin0.srcwin _name} .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
>     ("after" script)errorCode is POSIX ENOENT {no such file or directory}

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

Re: Insight 5.0 bug?

[Stephen Smith]

Ok.  Or should a new version number be assigned?

Fernando Nasser wrote:

> Yes,
>
> This was a know bug in 5.0 that has been fixed in more recent sources.
>
> Maybe Cygwin will have to go with a GDB snapshot instead of 5.0 (there
> is also some reported difficulties with ARM/Angel debugging I believe).
>
> Fernando
>
> Stephen Smith wrote:
> >
> > The inight 5.0 debugger is having a hard time finding and loading the the source code for
> > a currently running program the is running remote.
> >
> > The 20010306 snapshot works.
> >
> > Was this a documented error in the 5.0 release ( the currrently shiping version with cygwin)
> >
> > Spec
> >
> >      host=i686-pc-cygwin
> >      target=powerpc-motorola-elf
> >
> > sps
> >
> > -------------------- Stack trace of Insight 5.0 ---------------------------------
> > couldn't stat "/htc/DEOSExamples/threads/threads.cpp": no such file or directory
> >     while executing
> > "file mtime $f"
> >     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::_mtime_changed" body line 7)
> >     invoked from within
> > "_mtime_changed $filename"
> >     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::FillSource" body line 10)
> >     invoked from within
> > "FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib"
> >     ("SOURCE" arm line 2)
> >     invoked from within
> > "switch $current(mode) {
> >     SOURCE {
> >       FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib
> >     }
> >     ASSEMBLY {
> >       FillAssembly..."
> >     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::location" body line 26)
> >     invoked from within
> > "$twin location $tag $name $funcname $line $addr $pc_addr $lib"
> >     (object "::.srcwin0.srcwin" method "::SrcWin::location" body line 52)
> >     invoked from within
> > "location BROWSE_TAG [list $val "" $full 0 0 0 {}]"
> >     (object "::.srcwin0.srcwin" method "::SrcWin::_name" body line 22)
> >     invoked from within
> > "::.srcwin0.srcwin _name .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
> >     (in namespace inscope "::SrcWin" script line 1)
> >     invoked from within
> > "namespace inscope ::SrcWin {::.srcwin0.srcwin _name} .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
> >     ("after" script)errorCode is POSIX ENOENT {no such file or directory}
>
> --
> Fernando Nasser
> Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
> 2323 Yonge Street, Suite #300
> Toronto, Ontario   M4P 2C9

Re: Insight 5.0 bug?

[Fernando Nasser]

Stephen Smith wrote:
> 
> Ok.  Or should a new version number be assigned?
> 

If the Cygwin team uses a snapshot they will properly identify the
version.

I believe GDB 5.1 is not far from being released (a couple of months?)
and it will have a better bug fix handling police, so these situations
will no longer happen.

Fernando

> Fernando Nasser wrote:
> 
> > Yes,
> >
> > This was a know bug in 5.0 that has been fixed in more recent sources.
> >
> > Maybe Cygwin will have to go with a GDB snapshot instead of 5.0 (there
> > is also some reported difficulties with ARM/Angel debugging I believe).
> >
> > Fernando
> >
> > Stephen Smith wrote:
> > >
> > > The inight 5.0 debugger is having a hard time finding and loading the the source code for
> > > a currently running program the is running remote.
> > >
> > > The 20010306 snapshot works.
> > >
> > > Was this a documented error in the 5.0 release ( the currrently shiping version with cygwin)
> > >
> > > Spec
> > >
> > >      host=i686-pc-cygwin
> > >      target=powerpc-motorola-elf
> > >
> > > sps
> > >
> > > -------------------- Stack trace of Insight 5.0 ---------------------------------
> > > couldn't stat "/htc/DEOSExamples/threads/threads.cpp": no such file or directory
> > >     while executing
> > > "file mtime $f"
> > >     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::_mtime_changed" body line 7)
> > >     invoked from within
> > > "_mtime_changed $filename"
> > >     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::FillSource" body line 10)
> > >     invoked from within
> > > "FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib"
> > >     ("SOURCE" arm line 2)
> > >     invoked from within
> > > "switch $current(mode) {
> > >     SOURCE {
> > >       FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib
> > >     }
> > >     ASSEMBLY {
> > >       FillAssembly..."
> > >     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::location" body line 26)
> > >     invoked from within
> > > "$twin location $tag $name $funcname $line $addr $pc_addr $lib"
> > >     (object "::.srcwin0.srcwin" method "::SrcWin::location" body line 52)
> > >     invoked from within
> > > "location BROWSE_TAG [list $val "" $full 0 0 0 {}]"
> > >     (object "::.srcwin0.srcwin" method "::SrcWin::_name" body line 22)
> > >     invoked from within
> > > "::.srcwin0.srcwin _name .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
> > >     (in namespace inscope "::SrcWin" script line 1)
> > >     invoked from within
> > > "namespace inscope ::SrcWin {::.srcwin0.srcwin _name} .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
> > >     ("after" script)errorCode is POSIX ENOENT {no such file or directory}
> >
> > --
> > Fernando Nasser
> > Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
> > 2323 Yonge Street, Suite #300
> > Toronto, Ontario   M4P 2C9

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

Re: Insight 5.0 bug?

[Fernando Nasser]

Christopher Faylor wrote:
> 
> On Mon, Mar 12, 2001 at 10:45:01AM -0700, Stephen Smith wrote:
> >Ok.  Or should a new version number be assigned?
> 
> I do append the snapshot date to the version number.  I really can't
> invent a version number.  I have to use whatever is in CVS.
> 

AFAIK this is the right way of doing it.

Having the date allows whoever is looking into a problem to determine if
the code already has a fix or not, for one thing.

There was some talking into removing the release version string from the
CVS file in the GDB list recently.  This will further avoid any
misleading references.

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

Re: Insight 5.0 bug?

[Stephen Smith]

After some more testing,  it is only a problem for the files (*.h *.cpp *.c) in the /htc/DEOSExamples/threads directory.
This is the directory that contains function main().

Stephen Smith wrote:

> The inight 5.0 debugger is having a hard time finding and loading the the source code for
> a currently running program the is running remote.
>
> The 20010306 snapshot works.
>
> Was this a documented error in the 5.0 release ( the currrently shiping version with cygwin)
>
> Spec
>
>      host=i686-pc-cygwin
>      target=powerpc-motorola-elf
>
> sps
>
> -------------------- Stack trace of Insight 5.0 ---------------------------------
> couldn't stat "/htc/DEOSExamples/threads/threads.cpp": no such file or directory
>     while executing
> "file mtime $f"
>     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::_mtime_changed" body line 7)
>     invoked from within
> "_mtime_changed $filename"
>     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::FillSource" body line 10)
>     invoked from within
> "FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib"
>     ("SOURCE" arm line 2)
>     invoked from within
> "switch $current(mode) {
>     SOURCE {
>       FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib
>     }
>     ASSEMBLY {
>       FillAssembly..."
>     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::location" body line 26)
>     invoked from within
> "$twin location $tag $name $funcname $line $addr $pc_addr $lib"
>     (object "::.srcwin0.srcwin" method "::SrcWin::location" body line 52)
>     invoked from within
> "location BROWSE_TAG [list $val "" $full 0 0 0 {}]"
>     (object "::.srcwin0.srcwin" method "::SrcWin::_name" body line 22)
>     invoked from within
> "::.srcwin0.srcwin _name .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
>     (in namespace inscope "::SrcWin" script line 1)
>     invoked from within
> "namespace inscope ::SrcWin {::.srcwin0.srcwin _name} .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
>     ("after" script)errorCode is POSIX ENOENT {no such file or directory}
>
> --
> Want to unsubscribe from this list?
> Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: Insight 5.0 bug?

[Fernando Nasser]

Stephen Smith wrote:
> 
> After some more testing,  it is only a problem for the files (*.h *.cpp *.c) in the /htc/DEOSExamples/threads directory.
> This is the directory that contains function main().
> 

Yes, that was one of the places where file names were not being properly
converted before being passed on to Tcl.  I believe David Taylor was the
one who first detected it.  There were at least two more other
situations that did not make the conversion.  They are all fixed now.

Fernando


> Stephen Smith wrote:
> 
> > The inight 5.0 debugger is having a hard time finding and loading the the source code for
> > a currently running program the is running remote.
> >
> > The 20010306 snapshot works.
> >
> > Was this a documented error in the 5.0 release ( the currrently shiping version with cygwin)
> >
> > Spec
> >
> >      host=i686-pc-cygwin
> >      target=powerpc-motorola-elf
> >
> > sps
> >
> > -------------------- Stack trace of Insight 5.0 ---------------------------------
> > couldn't stat "/htc/DEOSExamples/threads/threads.cpp": no such file or directory
> >     while executing
> > "file mtime $f"
> >     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::_mtime_changed" body line 7)
> >     invoked from within
> > "_mtime_changed $filename"
> >     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::FillSource" body line 10)
> >     invoked from within
> > "FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib"
> >     ("SOURCE" arm line 2)
> >     invoked from within
> > "switch $current(mode) {
> >     SOURCE {
> >       FillSource t $tagname $filename $funcname $line $addr $pc_addr $lib
> >     }
> >     ASSEMBLY {
> >       FillAssembly..."
> >     (object "::.srcwin0.srcwin.container.pane1.childsite.con" method "::SrcTextWin::location" body line 26)
> >     invoked from within
> > "$twin location $tag $name $funcname $line $addr $pc_addr $lib"
> >     (object "::.srcwin0.srcwin" method "::SrcWin::location" body line 52)
> >     invoked from within
> > "location BROWSE_TAG [list $val "" $full 0 0 0 {}]"
> >     (object "::.srcwin0.srcwin" method "::SrcWin::_name" body line 22)
> >     invoked from within
> > "::.srcwin0.srcwin _name .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
> >     (in namespace inscope "::SrcWin" script line 1)
> >     invoked from within
> > "namespace inscope ::SrcWin {::.srcwin0.srcwin _name} .srcwin0.srcwin.container.pane3.childsite.con.name threads.cpp"
> >     ("after" script)errorCode is POSIX ENOENT {no such file or directory}
> >
> > --
> > Want to unsubscribe from this list?
> > Check out: http://cygwin.com/ml/#unsubscribe-simple

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

Re: Insight 5.0 bug?

[Christopher Faylor]

On Mon, Mar 12, 2001 at 08:50:48AM -0700, Stephen Smith wrote:
>The inight 5.0 debugger is having a hard time finding and loading the the source code for
>a currently running program the is running remote.
>
>The 20010306 snapshot works.
>
>Was this a documented error in the 5.0 release ( the currrently shiping version with cygwin)

Cygwin does not "ship with" version 5.0.  It uses a CVS snapshot.

Some people have reported problems, some people have no problems.  I'm one
of the people with no problems.

I will eventually make a new release of gdb available for cygwin.  Until that
time use of snapshots or CVS is recommended.

cgf