On Thu, 2012-02-09 at 13:14 +0200, Tor Lillqvist wrote:
On the other hand, desktop/Pagein_common.mk mentions all except fwm.
:-) prolly that needs removing I guess.
(But then fwm is clearly the very smallest of them, so whether it is
included in a merged super-fwk or not shouldn't matter much IMHO.)
What does this mean, that also fwl gets "paged in" anyway?
Personally I don't much care; the sizes are reasonably small anyway
compared to the systematic bloat that we suffer everywhere. It is
amazing to me that libfwl is 350kB stripped when you see what is (not)
in it :-)
$ grep ';' framework/source/classes/fwlresid.cxx framework/source/dispatch/mailtodispatcher.cxx
framework/source/dispatch/oxt_handler.cxx framework/source/dispatch/popupmenudispatcher.cxx
framework/source/dispatch/servicehandler.cxx framework/source/recording/dispatchrecorder.cxx
framework/source/recording/dispatchrecordersupplier.cxx framework/source/register/registertemp.cxx
framework/source/services/dispatchhelper.cxx framework/source/services/license.cxx
framework/source/services/mediatypedetectionhelper.cxx
framework/source/services/uriabbreviation.cxx framework/source/uielement/fontmenucontroller.cxx
framework/source/uielement/fontsizemenucontroller.cxx
framework/source/uielement/footermenucontroller.cxx
framework/source/uielement/headermenucontroller.cxx
framework/source/uielement/langselectionmenucontroller.cxx
framework/source/uielement/logoimagestatusbarcontroller.cxx
framework/source/uielement/logotextstatusbarcontroller.cxx
framework/source/uielement/macrosmenucontroller.cxx
framework/source/uielement/newmenucontroller.cxx framework/source/uielement/popupmenucontroller.cxx
framework/source/uielement/simpletextstatusbarcontroller.cxx
framework/source/uielement/toolbarsmenucontroller.cxx | wc -l
1681
We have 1700 lines of code, producing:
$ ls -lh /tmp/fwk/
-rwxr-xr-x 1 michael users 355K Feb 9 12:31 libfwllo.so
When stripped; 200+ bytes per line; the binary is 20K larger than the
source that went in. Hence my hope that we can do better size-wise with
some heavy-duty optimisation. But thoughts appreciated, it's nice to
save 500k on start if we can do it easily.
I guess it depends on the coupling; if we could create a two library
solution - the "libmerged" for startup, and a "libbloat" for when
corner-case features are used, we could perhaps partition our problem
into two bits; but hard to know where the best cut is there.
So, should I continue merging into one fwk_new = fwe+fwi+fwk+fwl+fwm
(mostly done, seems to work on OS X, still compiling on Windows), or
redo as two libs fwk_new = fwe+fwi+fwk and fwl_new = fwl+fwm ?
I'd do whatever is easiest :-)
Thanks,
Michael.
--
michael.meeks@suse.com <><, Pseudo Engineer, itinerant idiot
Context
Privacy Policy |
Impressum (Legal Info) |
Copyright information: Unless otherwise specified, all text and images
on this website are licensed under the
Creative Commons Attribution-Share Alike 3.0 License.
This does not include the source code of LibreOffice, which is
licensed under the Mozilla Public License (
MPLv2).
"LibreOffice" and "The Document Foundation" are
registered trademarks of their corresponding registered owners or are
in actual use as trademarks in one or more countries. Their respective
logos and icons are also subject to international copyright laws. Use
thereof is explained in our
trademark policy.