On Thu, Dec 11, 2014 at 04:01:04PM +0100, Michael Stahl wrote:
> forward every single method to a method of the pimpl class? i've never
> seen anybody do that
Oh, I did see that done a lot. Anyway, even without that, speed is the same
with -O3 and an abstract base class is still shorter. But as s
On 11.12.2014 13:49, Bjoern Michaelsen wrote:
> Pimpl
>
> (HXX):
>
> class AnotherThing;
> class RealThing;
>
> struct Thing {
> Thing(AnotherThing* pAnother);
> void DoSomething()
> { m_pImpl->DoSomething() }
> void DoSomethingElse()
> { m_pImpl->DoSomethingElse() }
On 12/11/2014 01:49 PM, Bjoern Michaelsen wrote:
To repeat: Im fine with adding Pimpls were sensible. But Id like to suggest to
maybe also consider ABCs were sensible.
...or KISS and have a plain struct pimpl
___
LibreOffice mailing list
LibreOffice@
Hi,
On Wed, Dec 10, 2014 at 09:09:04PM -0500, Kohei Yoshida wrote:
> 1) To reduce the size of our shared objects.
> 2) To improve our build time.
> [...]
> Please share your opinions.
Both are valuable goals and doers are deciders, so sounds good to me.
As a note though, I personally prefer an a
Kohei Yoshida wrote:
> 1) To reduce the size of our shared objects.
>
The above should not be underestimated for debug builds - there's
still lots of laptop hardware in student hands that has a physical
limit on RAM size (e.g. 4GB). I'd hate to have hacking LibO becoming
an absolutely unbearable ex
> Please share your opinions.
>
Well, I guess this email was caused by some complaints by me on IRC some
days ago... It was just that, complaints on IRC. Don't take it too
seriously. There will always be some workaround when debugging...
--tml
___
Libre
On 11/12/14 08:06, Miklos Vajna wrote:
> I very much appreciate the work you do here, if regular developers
> ignore the build time issues with buying faster and faster HW, then
> at some point we won't be able to attract new contributors.
As a gentoo-er, (and there are other similar distros :-)
ove both of the
> 2 aforementioned points, and I've only covered perhaps 10-20% of all
> possible candidates.
I don't have objection if _you_ are doing that :-) It's fine for me
when it is targeted, focused on the goals you've outlined.
But I did not know the reasons for pimpl-
ld be surprised if it had an impact on any non-debug parts of
link objects.) Another positive impact is improved build time with
today's toolchains (sans pch). (But remember that pimpl-ization will
always need to include into the include file, so don't go
overboard with erasing s
Hi Kohei,
On Wed, Dec 10, 2014 at 09:09:04PM -0500, Kohei Yoshida
wrote:
> 1) To reduce the size of our shared objects.
> 2) To improve our build time.
>
> 1) is apparently important for Michael Stahl, and I personally care
> about 2) since it directly affects my development efficiency.
>
> My
It has come to my attention that some folks are feeling a bit uneasy
about my recent pimplization of some of what I call "high impact" public
classes. I just wanted to make it clear that the reason I'm doing this
is in direct reponse to the discussion we recently had during our ESC
call wrt the bl
11 matches
Mail list logo
