On Sat, Jun 6, 2009 at 7:43 PM, Albin Tonnerre<[email protected]> wrote: > On Sat, Jun 06, 2009 at 07:20:39PM +0200, Cedric BAIL wrote : >> On Sat, Jun 6, 2009 at 5:48 PM, Albin Tonnerre<[email protected]> >> wrote: >> > On Sat, Jun 06, 2009 at 12:12:49PM -0300, Andre Dieb wrote : >> >> This patch contains my other two previous patches: >> >> >> >> - [PATCH] Fix eina mempool leak, init error handling, remove consumed >> >> TODO msg >> >> - [PATCH] Add checks to eina_module_new >> >> >> >> Also fixes the build error with --disable-magic-debug. >> > >> > I still think that when ecore_magic is disabled, it would be better to use >> > functions rather than preprocessor macros for things like >> > eina_magic_string_init(). Indeed, that would avoid exporting different >> > symbols >> > depending on whether eina has magic-debug enabled or not. If you use >> > macros, an >> > application using magic debug compiled against eina without magic-debug >> > will >> > have to be recompiled to benefit from it if it gets enabled in eina at a >> > later >> > point. If you use functions, this is no longer an issue. >> > >> > Thoughts ? >> >> I strongly disagree, when you disable magic debug in eina, you want to >> avoid it's cost at all. >>This include the call of the function call. >> And from my test, it does have a big impact to just enter an empty >> function. When we are speaking about performance, I think we can >> accept the little tradeoff of needing to recompile application using >> eina for maximum performance. > > If you're an application developer and use eina_magic, you're ready to pay > that > cost and more likely to rant about eina_magic being disabled than about any > performance loss, as you explicitely asked for it.
Well, in my opinion, when doing development I use magic debug for easing the stabilisation of my code, but in production I don't want to pay a price at all. I know that my hardware is certainly one of the slowest one where the EFL is running, and since a few years, it is constanly running better. This should continue. So please don't make the argument that we can accept a little drop in performance. We can measure, test and try way to do this differently, but we should not loose anything. > If you're talking about eina, nothing prevents us to do some trickery so that > functions calls are avoided internally, while still exporting the symbols. Well exporting the symbol could be done, if it's needed to prevent ABI break between two build of eina, one with magic debug and the other without. -- Cedric BAIL ------------------------------------------------------------------------------ OpenSolaris 2009.06 is a cutting edge operating system for enterprises looking to deploy the next generation of Solaris that includes the latest innovations from Sun and the OpenSource community. Download a copy and enjoy capabilities such as Networking, Storage and Virtualization. Go to: http://p.sf.net/sfu/opensolaris-get _______________________________________________ enlightenment-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
