Kai Backman wrote:

>   However, finding these problems is significantly faster if it is
> easy for everyone to -test- them, which means the CWS should be
> integrated sooner with issues rather than later when it's totally
> perfect. It is not turned on by default. You can build using both
> ways. You RE guys hate gigantic CWS's, maybe we should save some fixes
> for pchfix05 and later?

So it is a good thing that after 2.2 we will have 6 months until the
next release where we can let the code settle. We should try to find as
much problems beforehand though.

>  The bigger question is why are such constructs used in the first
> place? Modifying a hxx file using a define is just plain bad
> programming in most cases. 
Agreed. But it's pointless to ask these questions more than 10 years
after the code was written by people that are no longer with us. I
assume they had their reason - even if this reason was that they just
didn't know a better alternative. I assume that they just wanted to
avoid the explicit argument and thought that this is the easiest
approach. Anyway, we all agree that we should throw this out.

>> another point that worries me is the different list of includes
>> depending on using pch or not.
> 
>   Yes! Headers that are order dependent are usually broken in more
> than one ways anyway. Finding them by turning on pch would be a Good
> Thing.

Wouldn't be making sure that each and every header compiles as it is a
good idea for a first step? E.g. the item header files don't fulfil this
request but I'm sure there are more headers that don't compile.

>> at the moment, i'm not sure that this "all you can include" approach
>> will come to an happy end. maybe cautiously selecting a set of includes
>> for the precompile hxx would be more appropriate regarding the two
>> issues i currently see...
> 
>   At this point we have 8645 #include statements in the the various
> pch files. It took martink and me about a month to come up with this
> list using our simple strategy. Who will do this cautious selection?
> Would it perhaps be a better idea to include those fixes in pchfix05?

Correct me if I'm wrong, but aren't the headers used with pch different
to the headers used for compiling a particular source file anyway, even
if you have created the minimum set of needed headers? So if you use pch
you automatically agree that the definitions seen while compiling a
particular source file may differ in case some header files are evil[TM].

Ciao,
Mathias

-- 
Mathias Bauer (mba) - Project Lead OpenOffice.org Writer
OpenOffice.org Engineering at Sun: http://blogs.sun.com/GullFOSS
Please don't reply to "[EMAIL PROTECTED]".
I use it for the OOo lists and only rarely read other mails sent to it.

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to