POV-Ray : Newsgroups : povray.unofficial.patches : Hgpovray38, current state Server Time
6 Dec 2021 06:17:56 EST (-0500)
  Hgpovray38, current state (Message 38 to 47 of 47)  
<<< Previous 10 Messages Goto Initial 10 Messages
From: William F Pokorny
Subject: Re: Hgpovray38, current state
Date: 25 Oct 2020 08:04:34
Message: <5f9569d2$1@news.povray.org>
On 10/25/20 5:57 AM, Le_Forgeron wrote:
> Le 25/10/2020 à 10:33, Le_Forgeron a écrit :
>> Le 25/10/2020 à 02:35, Bald Eagle a écrit :
...
>>
>> Commenting the media in the last union, get the same bad result as
>> hgpovray38 with povray master 3.8;
>>

I'd not have looked if not for the claim of the same bad results in 
povray master 3.8, but I branched off the last commit to master and I do 
not see that line in tracepixel.cpp commented. (I've not run the test case)

If any help, I went on to look at the state of tracepixel.cpp for every 
commit since the move the github. I see twelve up to v3.8 master head 
and none have the precompute... line commented.

> 
> Ok, I can fix it... but I do not understand yet why it happened and how
> it works.
> 
> diff --git a/source/core/render/tracepixel.cpp
> b/source/core/render/tracepixel.cpp
> index 6d661b07..50d420a1 100644
> --- a/source/core/render/tracepixel.cpp
> +++ b/source/core/render/tracepixel.cpp
> @@ -235,7 +235,7 @@ void TracePixelCameraData::SetupCamera(const Camera&
> cam)
>       bool normalise = false;
>       camera = cam;
>       useFocalBlur = false;
> -    //precomputeContainingInteriors = true;
> +    precomputeContainingInteriors = true;
>       cameraDirection = camera.Direction;
>       cameraRight = camera.Right;
>       cameraUp =  camera.Up;
>


Post a reply to this message

From: Le Forgeron
Subject: Re: Hgpovray38, current state
Date: 25 Oct 2020 08:45:43
Message: <5f957377$1@news.povray.org>
Latest package:

https://github.com/LeForgeron/povray/releases/tag/v3.8.0.alpha%2BHg.229.Betterave

> https://github.com/LeForgeron/povray/releases/tag/v3.8.0.alpha%2BHg.229.Betterave


Post a reply to this message

From: Mr
Subject: Re: Hgpovray38, current state
Date: 26 Oct 2020 02:40:00
Message: <web.5f966e17b933d8376adeaecb0@news.povray.org>
Le_Forgeron <jgr### [at] freefr> wrote:
> Latest package:
>
> https://github.com/LeForgeron/povray/releases/tag/v3.8.0.alpha%2BHg.229.Betterave
>
> > https://github.com/LeForgeron/povray/releases/tag/v3.8.0.alpha%2BHg.229.Betterave

Thanks ! It does render with blue background now.
Next issue is when I try to render the grenadine scene, I get a "Parse Error:
Cannot open include file lemon.inc." (disregarding any +L option).
That however could also happen with standard povray 3.8.


Post a reply to this message

From: Mr
Subject: Re: Hgpovray38, current state
Date: 26 Oct 2020 03:10:01
Message: <web.5f967589b933d8376adeaecb0@news.povray.org>
"Mr" <nomail@nomail> wrote:
> Le_Forgeron <jgr### [at] freefr> wrote:
> > Latest package:
> >
> > https://github.com/LeForgeron/povray/releases/tag/v3.8.0.alpha%2BHg.229.Betterave
> >
> > >
https://github.com/LeForgeron/povray/releases/tag/v3.8.0.alpha%2BHg.229.Betterave
>
> Thanks ! It does render with blue background now.
> Next issue is when I try to render the grenadine scene, I get a "Parse Error:
> Cannot open include file lemon.inc." (disregarding any +L option).
> That however could also happen with standard povray 3.8.

Sorry, ignore the grenadine (non)issue, I was making a typo.


Post a reply to this message

From: Mr
Subject: Re: Hgpovray38, current state
Date: 6 Nov 2020 03:15:01
Message: <web.5fa5058db933d8376adeaecb0@news.povray.org>
Le_Forgeron <jgr### [at] freefr> wrote:
> Le 24/10/2020 à 17:59, Le_Forgeron a écrit :
> > Le 23/10/2020 à 16:10, Mr a écrit :
> >> Le_Forgeron <jgr### [at] freefr> wrote:
> >>> For the people who do not want to bother with git, Hgpovray is now
> >>> available as sources
> >>>
> >>>
> >>>
https://github.com/LeForgeron/povray/releases/tag/v3.8.0.alpha%2BHg.226-Thermidor-30
> >>>
> >>>>
https://github.com/LeForgeron/povray/releases/tag/v3.8.0.alpha%2BHg.226-Thermidor-30
> >>>
> >>> More works on the agenda :
> >>>
> >>> http://wiki.povray.org/content/User:Le_Forgeron/HgPovray38
> >>
> >> Hello, ...With the set of default scenes, the advanced/abyss.pov shows a black
> >> background behind submarine instead of blue with current 3.8 plain pov master.
> >>
> >>
> >
> > Can you identify the version of hgpovray38 ?
> >
> Oups, my bad. 226 Thermidor is 2 years old.
>
> I will check with the latest head.

Hi, I tried building on windows again with "betterave", and after reimplementing
all the changes previously mentioned in this thread I manage to push again many
of the errors away, but fail to see what to do about the wrong type line 485 of
windows/pvengine.cpp :
https://github.com/LeForgeron/povray/blob/Hgpovray/windows/pvengine.cpp#L485

At the end of a struct array typed WORD, having -1 values maybe to specify end
of array that looks like list of all keys that allow to scroll.

It seems that some difference relied upon *outside* this file allows vanilla pov
to build with that same technique.


Is there maybe some library that I could change to more recent version? (I
already pushed many errors by replacing boost folder... But maybe I want too far
with that one in a too recent version?


Post a reply to this message

From: Mr
Subject: Re: Hgpovray38, current state
Date: 6 Nov 2020 06:45:00
Message: <web.5fa5365eb933d8376adeaecb0@news.povray.org>
"Mr" <nomail@nomail> wrote:
> Le_Forgeron <jgr### [at] freefr> wrote:
> > Le 24/10/2020 à 17:59, Le_Forgeron a écrit :
> > > Le 23/10/2020 à 16:10, Mr a écrit :
> > >> Le_Forgeron <jgr### [at] freefr> wrote:
> > >>> For the people who do not want to bother with git, Hgpovray is now
> > >>> available as sources
> > >>>
> > >>>
> > >>>
https://github.com/LeForgeron/povray/releases/tag/v3.8.0.alpha%2BHg.226-Thermidor-30
> > >>>
> > >>>>
https://github.com/LeForgeron/povray/releases/tag/v3.8.0.alpha%2BHg.226-Thermidor-30
> > >>>
> > >>> More works on the agenda :
> > >>>
> > >>> http://wiki.povray.org/content/User:Le_Forgeron/HgPovray38
> > >>
> > >> Hello, ...With the set of default scenes, the advanced/abyss.pov shows a black
> > >> background behind submarine instead of blue with current 3.8 plain pov master.
> > >>
> > >>
> > >
> > > Can you identify the version of hgpovray38 ?
> > >
> > Oups, my bad. 226 Thermidor is 2 years old.
> >
> > I will check with the latest head.
>
> Hi, I tried building on windows again with "betterave", and after reimplementing
> all the changes previously mentioned in this thread I manage to push again many
> of the errors away, but fail to see what to do about the wrong type line 485 of
> windows/pvengine.cpp :
> https://github.com/LeForgeron/povray/blob/Hgpovray/windows/pvengine.cpp#L485
>
> At the end of a struct array typed WORD, having -1 values maybe to specify end
> of array that looks like list of all keys that allow to scroll.
>
> It seems that some difference relied upon *outside* this file allows vanilla pov
> to build with that same technique.
>
>
> Is there maybe some library that I could change to more recent version? (I
> already pushed many errors by replacing boost folder... But maybe I want too far
> with that one in a too recent version?

Adding :
-Wno-c++11-narrowing

into additional compiler options seems to have solved that... next issues coming
soon :-)


Post a reply to this message

From: Mr
Subject: Re: Hgpovray38, current state
Date: 13 Nov 2020 11:15:01
Message: <web.5faeb074b933d8376adeaecb0@news.povray.org>
"Mr" <nomail@nomail> wrote:
> >
> > Is there maybe some library that I could change to more recent version? (I
> > already pushed many errors by replacing boost folder... But maybe I want too far
> > with that one in a too recent version?
>
> Adding :
> -Wno-c++11-narrowing
>
> into additional compiler options seems to have solved that... next issues coming
> soon :-)


It now tells me as error for
libraries\boost\boost\type_index\detail\compile_time_type_info.hpp ,  line 279
(" boost::typeindex::detail::failed_to_get_function_name();") :
no matching function for call to 'failed_to_get_function_name'



....and as info for \libraries\boost\boost\type_index\ctti_type_index.hpp , line
117 :
in instantiation of member function 'boost::detail::ctti<void>::n' requested
here

also info for libraries\boost\boost\type_index\detail\compile_time_type_info.hpp
, line 87 :
candidate template ignored: couldn't infer template argument 'T'

If I click to show definition of 'T', below are lines :
"TypeIndex library could not detect your compiler.Please make the
BOOST_TYPE_INDEX_FUNCTION_SIGNATURE macro use correct compiler macro for getting
the whole function name. Define BOOST_TYPE_INDEX_CTTI_USER_DEFINED_PARSING to
correct value after that."

I am using Visual Studio 2019 building Release x64 target, with LLVM as a
compiler, ISO standard Language version C++14, --disable-io-restrictions
-Wno-c++11-narrowing, and with all other changes mentioned before in this
discussion. any clue what else to change to keep/(start?) moving forward?


Post a reply to this message

From: Mr
Subject: Re: Hgpovray38, current state
Date: 28 Feb 2021 17:20:00
Message: <web.603c1617b933d8376adeaecb0@news.povray.org>
Hi, I made yet another attempt at building on Windows with other parameters and
received different warnings mentioning :

unknown type name '__m64'

Is there any chance to replace / enhance definition of this in the source code
with something more cross platform  ?


Post a reply to this message

From: Le Forgeron
Subject: Re: Hgpovray38, current state
Date: 1 Mar 2021 06:56:51
Message: <603cd683$1@news.povray.org>
Le 28/02/2021 à 23:15, Mr a écrit :
> Hi, I made yet another attempt at building on Windows with other parameters and
> received different warnings mentioning :
> 
> unknown type name '__m64'
> 
> Is there any chance to replace / enhance definition of this in the source code
> with something more cross platform  ?
> 
> 
> 

Can you elaborate a bit more (location of the error ?) ?

What is puzzling is that something like __m64 starts with double
underscore and that is something which should be reserved to compiler
implementation.

Any capture of the full message ?


Post a reply to this message

From: Mr
Subject: Re: Hgpovray38, current state
Date: 1 Mar 2021 11:30:00
Message: <web.603d1603b933d8376adeaecb0@news.povray.org>
Le_Forgeron <jgr### [at] freefr> wrote:
> Le 28/02/2021 à 23:15, Mr a écrit :
> > Hi, I made yet another attempt at building on Windows with other parameters and
> > received different warnings mentioning :
> >
> > unknown type name '__m64'
> >
> > Is there any chance to replace / enhance definition of this in the source code
> > with something more cross platform  ?
> >
> >
> >
>
> Can you elaborate a bit more (location of the error ?) ?
>
> What is puzzling is that something like __m64 starts with double
> underscore and that is something which should be reserved to compiler
> implementation.
>
> Any capture of the full message

 Indeed includes belonged to LLVM /CLANG 10 for VC... The thing is I know too
little, especially about expected options and get lost in trial and errors.

A few clues or reminders of what "should" normally  be the best options could
help me to get there quicker. (Using a Windows 10 with an I7 CPU) current
settings: when loading a file I let it update toolset from v140 to version142

Configuration: Debug, Platform: x64
Windows SDK version 10 ... Better with the default of 8.1?
Floating point mode Precise... Better with strict or fast?

MFC: Use Standard Windows Libraries... Better with use MFC from a static
library? or from a shared dll?
Desactivate language extensions : Yes (/Za)... ok?
iso norm for C version?  trying with MSVC inherited... Better with C 11 or C 17?
iso norm for C++ version?  trying with default std C++14 (nothing older)...
Better with 17 or "latest"?

Character set currently undefined  in project file, better with unicode or MBCS?

No Common Language Runtime

Compile as C++(/TP) ... Better with compile as C Code (/TC)?
With that, I get all the errors mentionned in this discussion, so I stopped
assuming that the code should not be touched and am about to start again to the
suggested changes. I was only searching for a working build setup witout it
first : so I now switched the toolset from Visual Studio 2019 (v142) to LLVM
(clang-cl).. better with LLVM? any clues?


Post a reply to this message

<<< Previous 10 Messages Goto Initial 10 Messages

Copyright 2003-2021 Persistence of Vision Raytracer Pty. Ltd.