[prev in list] [next in list] [prev in thread] [next in thread]
List: wine-devel
Subject: Re: [v6 1/7] d3dx9: Introduce preshaders in effect.
From: Paul Gofman <gofmanp () gmail ! com>
Date: 2016-03-27 23:23:04
Message-ID: CAFJEe9K0LW7LoVtES65MjycshMqD8OgZ8+RokvqbZWsUat-J7A () mail ! gmail ! com
[Download RAW message or body]
[Attachment #2 (multipart/alternative)]
I will resend tomorrow morning with these changes and the old+these changes
list.
I can surely fix object creation (presumably just free memory on copy over
duplicate object at the first place). But maybe we should also skip object
data creation for object_id 0 at all? In any case it would be probably
easier to deal with it after finishing off the current patchset, I already
have a few other things for d3dx effects on top of it and it is becoming
difficult to manage :)
On Mon, 28 Mar 2016 at 02:07, Matteo Bruni <matteo.mystral@gmail.com> wrote:
> 2016-03-28 0:54 GMT+02:00 Paul Gofman <gofmanp@gmail.com>:
> > Yes, sorry, this is the case. It looks like I already changed it later in
> > preliminary shared parameters (pool) implementation, but did not get
> back to
> > the right patch in the sequence where it is also a problem.
> > Should I resend the patches with just these changes or wait for your
> > comments on the remaining part?
>
> You can resend, I haven't managed to seriously review the other
> patches in this iteration anyway :/ I'll try to take some time to
> review them tomorrow either way.
> Only noticed a nitpick, in patch 2/7 there is an if with no brackets
> followed by its else with them. You should add brackets to the if too.
>
> > PS I think there might be also a similar problem for object creation in
> > effect parsing when there are multiple object_id 0 in effect (which is
> > accompanied by a warning that object is already created which is seen
> quite
> > often), data copying code does not free existing pointer.
>
> Right. I'll write a patch for that, if you don't get there first.
>
[Attachment #5 (text/html)]
I will resend tomorrow morning with these changes and the old+these changes \
list.<br>I can surely fix object creation (presumably just free memory on copy over \
duplicate object at the first place). But maybe we should also skip object data \
creation for object_id 0 at all? In any case it would be probably easier to deal with \
it after finishing off the current patchset, I already have a few other things for \
d3dx effects on top of it and it is becoming difficult to manage :)<br><div \
class="gmail_quote"><div dir="ltr">On Mon, 28 Mar 2016 at 02:07, Matteo Bruni <<a \
href="mailto:matteo.mystral@gmail.com">matteo.mystral@gmail.com</a>> \
wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 \
.8ex;border-left:1px #ccc solid;padding-left:1ex">2016-03-28 0:54 GMT+02:00 Paul \
Gofman <<a href="mailto:gofmanp@gmail.com" \
target="_blank">gofmanp@gmail.com</a>>:<br> > Yes, sorry, this is the case. It \
looks like I already changed it later in<br> > preliminary shared parameters \
(pool) implementation, but did not get back to<br> > the right patch in the \
sequence where it is also a problem.<br> > Should I resend the patches with just \
these changes or wait for your<br> > comments on the remaining part?<br>
<br>
You can resend, I haven't managed to seriously review the other<br>
patches in this iteration anyway :/ I'll try to take some time to<br>
review them tomorrow either way.<br>
Only noticed a nitpick, in patch 2/7 there is an if with no brackets<br>
followed by its else with them. You should add brackets to the if too.<br>
<br>
> PS I think there might be also a similar problem for object creation in<br>
> effect parsing when there are multiple object_id 0 in effect (which is<br>
> accompanied by a warning that object is already created which is seen quite<br>
> often), data copying code does not free existing pointer.<br>
<br>
Right. I'll write a patch for that, if you don't get there first.<br>
</blockquote></div>
[Attachment #6 (text/plain)]
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic