This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
[Bug translator/16615] don't require access to dwarf_query in has_single_line_record()
- From: "jlebon at redhat dot com" <sourceware-bugzilla at sourceware dot org>
- To: systemtap at sourceware dot org
- Date: Thu, 17 Apr 2014 21:56:36 +0000
- Subject: [Bug translator/16615] don't require access to dwarf_query in has_single_line_record()
- Auto-submitted: auto-generated
- References: <bug-16615-6586 at http dot sourceware dot org/bugzilla/>
https://sourceware.org/bugzilla/show_bug.cgi?id=16615
Jonathan Lebon <jlebon at redhat dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution|--- |FIXED
--- Comment #4 from Jonathan Lebon <jlebon at redhat dot com> ---
Merged into master.
(In reply to Jonathan Lebon from comment #3)
> We no longer rely on dwarf_getsrc_file(), which is slow, imprecise, and has
> funny behaviours.
Hmmm, that sounded a bit harsh in retrospect. Just to clarify,
dwarf_getsrc_file() does the job, but isn't what we need here. It scours all
CUs for any Dwarf_Lines in the target srcfile, when all we're interested in is
the current CU. This is one of the main reasons it is slower. If no line
records are found for the wanted lineno, it picks the next lineno with line
records. This complicates things for us, since we want to be precise.
These are the reasons why it's not a good fit for what we're trying to achieve
here (but may work well in other situations).
--
You are receiving this mail because:
You are the assignee for the bug.