From: P Z. <zol...@gm...> - 2009-10-16 17:10:21
|
Yes; it's caused by the debug information (for example I use the configure script with the --enable-debug=full). For instance if I remember correctly, I've once compiled pidgin wiht debug info... the result had something like 30MB main executable. After stripping, its size reduced to ~2MB. By the way, does the trunk compile for you (rev 532) ? For me, it complains about CurrecntSource constructor in ECCurrentSource constructor...? On Wed, 14 Oct 2009 06:55:25 +0200, Alan Grimes <ag...@sp...> wrote: > Anyone ever notice how absolutely ginormous ktechlab's binary size is? > > unstripped: > -rwxr-xr-x 1 atg tty 51,783,635 Oct 13 14:38 ktechlab > > (because linux is not half as good as a late version of DOS, I had to > manually add the comas for readability...) > > stripped: > > -rwxr-xr-x 1 atg tty 4,182,148 Oct 14 00:42 ktechlab-strip > > which is almost reasonable... > > This is one of the main reasons I've spent so much effort trying to > crush as much complexity out of the program as possible. =\ > > One of the biggest contributors is probably that we're carrying around a > hefty chunk of qt cuz we need a customized canvas implementation... > > Ideally, this would be a separately maintained library that would be > outside of the ktechlab binary... > > The other major factor is that we are maintaining, internally, a large > parts library which should be externalized (for the most part), as a > database. > > By way of comparison, kdevelop3's main binary is: > > atg@leenooks /usr/kde/3.5/bin $ ls -l kdevelop > -rwxr-xr-x 1 root root 425761 Sep 28 02:59 kdevelop > ***UNSTRIPPED!!!*** > > Seamonkey web browser? > > atg@leenooks /usr/lib/seamonkey $ ls -l seamonkey-bin > -rwxr-xr-x 1 root root 723289 Sep 25 23:57 seamonkey-bin > (again, unstripped) > > I'm sure you can find many other applications to weigh ktechlab > against... The moral of the story is that we have a problem. (which I > knew about years ago...) > |