public inbox for frysk@sourceware.org
 help / color / mirror / Atom feed
* common print-stack options
@ 2008-03-20 18:34 Andrew Cagney
  2008-03-26 12:06 ` Mark Wielaard
  0 siblings, 1 reply; 2+ messages in thread
From: Andrew Cagney @ 2008-03-20 18:34 UTC (permalink / raw)
  To: frysk

http://sourceware.org/bugzilla/show_bug.cgi?id=5978

FYI, all the utilities (fstack, ftrace, ferror, fcatch) now share a 
common set of options for specifying what a stack backtrace should look 
like, for instance:

$ ./frysk-core/frysk/bindir/ftrace -stack -full-path -sys=write -- /bin/ls
[...]
10774.10774 syscall write(1, "config.log                          
files...", 53)\
10774.10774  dumping stack trace:
Task #10774
#0 0x00110402 in __kernel_vsyscall () from 
/home/scratch/frysk/git/native/[vdso]
#1 0x00855135 in new_do_write () from /lib/libc-2.7.so
#2 0x0085541f in _IO_do_write@@GLIBC_2.1 () from /lib/libc-2.7.so
#3 0x00855d28 in _IO_file_overflow@@GLIBC_2.1 () from /lib/libc-2.7.so
#4 0x00858793 in __overflow () from /lib/libc-2.7.so
#5 0x0804ea57 in [unknown] from /bin/ls
#6 0x0804f026 in [unknown] from /bin/ls
#7 0x08050a9e in [unknown] from /bin/ls
#8 0x00808390 in __libc_start_main () from /lib/libc-2.7.so
#9 0x08049921 in [unknown] from /bin/ls
10774.10774 syscall leave write = -1 ERRNO=38

prints the stack backtrace withouth truncating paths.

I've marked down a brainstorming session on what the print-stack options 
should be for next wednesday's meeting.

Andrew


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: common print-stack options
  2008-03-20 18:34 common print-stack options Andrew Cagney
@ 2008-03-26 12:06 ` Mark Wielaard
  0 siblings, 0 replies; 2+ messages in thread
From: Mark Wielaard @ 2008-03-26 12:06 UTC (permalink / raw)
  To: Andrew Cagney; +Cc: frysk

Hi Andrew,

On Thu, 2008-03-20 at 14:33 -0400, Andrew Cagney wrote:
> http://sourceware.org/bugzilla/show_bug.cgi?id=5978
> 
> FYI, all the utilities (fstack, ftrace, ferror, fcatch) now share a 
> common set of options for specifying what a stack backtrace should look 
> like, for instance:
>
> I've marked down a brainstorming session on what the print-stack options 
> should be for next wednesday's meeting.

I think the suggested options in the bug report are nice:

Stack print options:
  -n, -number-of-frames NUMBER    number of frames to print. Use -n 0 or -n 
                                    all to print all frames.
  -f, -full-path                  print the full path to object files
  -a, -all                        Print all information that can be retrieved 
                                    about the stack
                                    this is equivalent to -print 
                                    functions,params,scopes,fullpath
  -v, -virtual                    Includes virtual frames in the stack trace.
                                    Virtual frames are artificial frames 
                                    corresponding to calls to inlined functions
  -c, -common                     print commonly used debug information:this 
                                    is equivalent to -print 
                                    functions,params,fullpath
      -print [item],...           itmes to print. Possible items:
                                    functions : print function names using 
                                    debug information
                                    scopes : print variables declared in each 
                                    scope within the function.
                                    params : print function parameters
                                    fullpath : print full executbale path

Having some options separate and available also in --print [item] is
slightly confusing though, I would propose to make them all "top-level"
options, or all --print [item] options.

Since this is a common option but just for stack trace printing, you
might want to change --print [item] to --stack [item]. Where a --stack
without items replaces the --common option. And maybe replace -a with
--stack all.

Cheers,

Mark

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2008-03-26 12:06 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2008-03-20 18:34 common print-stack options Andrew Cagney
2008-03-26 12:06 ` Mark Wielaard

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 read-only IMAP folder(s) and NNTP newsgroup(s).