[Bug translator/11553] New: need better message for return probe on inline function

[Bug translator/11553] New: need better message for return probe on inline function

[jkenisto at us dot ibm dot com]

If I request a .return probe on a function that the compiler decides to inline,
the error message I get is something like this:
semantic error: no match while resolving probe point
process("./a.out").function("inc").return
... which is technically true, but not as helpful as it could be.  It'd be nice
if you could add something like:
(This function is inlined, so no return probe is possible.)

It is indeed documented that you can't have return probes on inline functions;
but have pity on the novice user, who may not be aware of this limitation and/or
gcc's penchant for inlining functions -- even those you've explicitly asked it
not to inline.

-- 
           Summary: need better message for return probe on inline function
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: translator
        AssignedTo: systemtap at sources dot redhat dot com
        ReportedBy: jkenisto at us dot ibm dot com


http://sourceware.org/bugzilla/show_bug.cgi?id=11553

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug translator/11553] need better message for return probe on inline function

[jistone at redhat dot com]

------- Additional Comments From jistone at redhat dot com  2010-04-28 17:54 -------
This is very similar to bug #5076.  It gets even harder when there's a wildcard
in the function name -- if every match happens to be an inline, how should we
warn about this?

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11553

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug translator/11553] need better message for return probe on inline function

[jkenisto at us dot ibm dot com]

------- Additional Comments From jkenisto at us dot ibm dot com  2010-04-28 18:34 -------
(In reply to comment #1)
> This is very similar to bug #5076.

As I should have mentioned, I looked at both #4413 and #5076 before I filed
this.  This is related, but a different issue.

> It gets even harder when there's a wildcard
> in the function name -- if every match happens to be an inline, how should we
> warn about this?

The message you currently give is the same no matter whether it's an
un-wildcarded name with no match (i.e., no return probe possible) or a
wildcarded name with 0 matches, right?  It'd be nice if you could notice that
multiple functions were found but couldn't be probed, and change the message to:
(These functions are inlined, so no return probes are possible.)

The question you didn't ask -- What, if anything, do we say if some matched
functions are inline and some aren't? -- is an interesting one, but the scope of
that issue is larger than the one this PR is concerned with.  stap -l is
presumably enlightening in those situations.

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11553

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug translator/11553] need better message for return probe on inline function

[fche at redhat dot com]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|systemtap at sources dot    |fche at redhat dot com
                   |redhat dot com              |
             Status|NEW                         |ASSIGNED


http://sourceware.org/bugzilla/show_bug.cgi?id=11553

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.