On 04/22/2013 01:17 PM, Bjoern Michaelsen wrote:
On Mon, Apr 22, 2013 at 01:04:30PM +0200, Stephan Bergmann wrote:
On 04/17/2013 04:45 PM, Bjoern Michaelsen wrote:
On Wed, Apr 17, 2013 at 04:13:15PM +0200, Stephan Bergmann wrote:
What good is the intermediate move? That is, was there at least one
situation in which it would have helped your debugging if an
existing RTL_LOGFILE_* call had been behaving like SAL_LOG?
Yes, there had been: I was trying to wrap my haed around the
observer-pattern-gone-bad of vcl-callbacks going into
framework/uielement/menubarmanager.cxx to see if there is something going wrong
with the order of execution. RTL_LOGFILE provides some helpful ad-hoc hints
there without manually setting bazillion breakpoints or adding SAL_INFOs there.
I think Tor's comment is to the point. Ultimately, we'll want
SAL_WARN/SAL_INFO enabled in more builds, so we shouldn't get too
carried away with adding too many SAL_INFOs for tracing purposes.
I can see your point for SAL_WARN, but not really for SAL_INFO. I cant think of
a scenario where a buildwide enabling of SAL_INFO makes much sense, however I
can see much use in enabling it for one lib/module/whatever, and for that it
doesnt have to be too restrained.
I'm not sure we're talking about the same thing here. With "enabled" I
meant having SAL_WARN/SAL_INFO expand to actual code (instead of having
the compiler optimize it away). Whether that code then also does some
logging depends on further constraints (SAL_LOG env var), but my point
was that the code itself does incur a certain (small) runtime cost.
Stephan
_______________________________________________
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice
