Re: [Libreoffice] Assertions and Logging

 First of all: I see this has been already committed, after giving people only 
slightly more than the weekend and the day after it to react, which I 
consider too short for something with as large scope as this. Some of us 
don't work on the weekends and have a backlog when they come back on Monday.

On Friday 18 of November 2011, Stephan Bergmann wrote:
...
> - The implementation is somewhat careful to produce as little code as
> possible at the call-site of the new macros, and to keep the code path
> for suppressed logs as short as possible.  However, the C++-stream-style
> macros will potentially always be more expensive than the
> C-printf-format-style ones.

 Does it really matter, when it is used only in debug builds anyway? Debug 
builds are slower already anyway, and if the cost of generating the output is 
considered expensive, what about the cost of wherever the output will end up?

> One open question is which set of macros (the C-printf-format-style ones
> or the C++-stream-style ones) to give the "natural" names without
> additional suffix.

 IMO that is a rather simpler decision: The natural names should be used by 
the only set that exists, for a number of reasons:
- the C variants don't work with C++ strings (or other non-trivial data), 
which makes them next to useless
- this is debug code, so the slight performance cost is negligible
- where do we have non-C++ code that has actually needs to use this 
functionality and cannot just have the filename extension changed?
- people will mix the two sets randomly, with all the consequences (mess, 
wondering when to use which one)
- the _S suffix is ugly and it is the variant that'll eventually be used more
- if there are any (real) problems with the C++ variant, perhaps they should 
be simply rather fixed/improved
- WTH should we intentionally have two sets of one functionality again?

-- 
 Lubos Lunak
 l.lunak@suse.cz