Hi Lubos, On Wed, 30 Mar 2011 11:21:26 +0200 Lubos Lunak <l.lunak@suse.cz> wrote:
How does API have to do anything with marshalling?
Language bindings.
Do you mean UNO with that?
Yes.
And what do you estimate is the ratio of code that does marshalling to the rest of the code?
In the applications (and that is the major code block) practically everything, as everything has at least some UNO wrapper -- and UNO is also used a lot internally.
If it's used for marshalling, then it can't be changed. Or, if it can, then it doesn't matter if the data would be simply marshalled as int.
Some sal_* => "int" conversion that works on all current platforms does not have to on the next one coming around.
Are you sure you're not arguing for my way here? The only way to get the same type is if the codebase normally used int. Now it's full of different types (sal_* ones and standard ones, all mixed).
I think most code now uses sal_* as of now and less code uses the
standard ones. For example, searching for this stuff roughly a la
find clone -name "*.cxx"|xargs grep "unsigned int"|wc -l
in clone (excluding libs-extern):
sal_uInt32 unsigned int
clone without external 24005 1246
hxx 7560 256
cxx 16445 999
Thus my conclusion.
And I'm also not arguing for converting everything right now this very moment. But I don't see why we should have an easy task that moves the situation in the wrong direction.
Agreed. Changing stuff there will just create lots of useless merge conflicts. And if we see Oracle doing something stupid with sal_uLong, we should do our better solution after _merging_ that. The worst situtaton however is: they walk on to greener pastures because priorities changed this week and sal_uLong keeps hanging around. Best Regards, Bjoern -- https://launchpad.net/~bjoern-michaelsen
Attachment:
signature.asc
Description: PGP signature