On Mon, Feb 11, 2008 at 12:33:47PM +0100, Nikolai Pretzell wrote: > Frank Schönheit - Sun Microsystems Germany schrieb: >> What, except removing some "ugliness", which is alway highly subjective, >> is the gain of this change? > > Maintainability. > > The "uglyness" IMHO is not subjective, and a defect. Maintainability is > a major problem in any project as huge as OOo, so any improvement is > helpful. The external header guards have two kinds of impact: > - unnecessary code, > - a potential cause for difficult to find errors, when internal include > guards change. > Yep. External guards expose an implementation detail of the included header, that's a) likely to change (and _is_ changing already for those that unfortunately took the guards out of boost or stlport headers), and b) apparently fragile, as I found a whole lot of places where non-existing or misnamed (wrong case, etc.) headers were included - change the includes in one central header, and you risk build breakages across the whole office.
So, in my book, this change is justified, and only lets us suffer some pain now, instead of continuous pain in the future, at a possibly much more inconvient time. As I mentioned, I'll happily exclude dbaccess from this CWS, since you largely reworked includes there. Cheers, -- Thorsten --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
