From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 7399 invoked by alias); 2 Nov 2009 17:03:54 -0000 Received: (qmail 4800 invoked by uid 48); 2 Nov 2009 17:03:42 -0000 Date: Mon, 02 Nov 2009 17:03:00 -0000 Message-ID: <20091102170342.4799.qmail@sourceware.org> From: "chwang at redhat dot com" To: systemtap@sources.redhat.com In-Reply-To: <20091031154558.10877.fche@redhat.com> References: <20091031154558.10877.fche@redhat.com> Reply-To: sourceware-bugzilla@sourceware.org Subject: [Bug translator/10877] improve probe point error reporting X-Bugzilla-Reason: AssignedTo Mailing-List: contact systemtap-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Post: List-Help: , Sender: systemtap-owner@sourceware.org X-SW-Source: 2009-q4/txt/msg00384.txt.bz2 ------- Additional Comments From chwang at redhat dot com 2009-11-02 17:03 ------- (In reply to comment #0) > Currently we store just one tok* per probe_point, instead of one > per probe point element (the stuff around the dots). This leads > to hard-to-interpret error messages > > % stap -p2 -e 'probe kernel.junkction("sdf") {}' > semantic error: probe point mismatch at position 1 (alternatives: > function(number) function(string) mark(string) statement(number) > statement(string) trace(string)): identifier 'kernel' at :1:7 while > resolving probe point kernel.junkction("sdf") > source: probe kernel.junkction("sdf") {} > ^ > Pass 2: analysis failed. Try again with another '--vp 01' option. > > Instead, we should store a tok* per probe point element, so that an > error can point the caret at just the right spot, and we can drop > the "at position N" from the error description body. To confirm, you are suggesting the use of: > source: probe kernel.junkction("sdf") {} > ^ Yes? -- http://sourceware.org/bugzilla/show_bug.cgi?id=10877 ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.