fSDK© - Frankie's SDK for PellesC

[frankie]

New version:
- 12/Feb/2017 -- V.0.1.2.1 (12.Feb.2017)
  - Fixed header xmllite.h (discovered by TimoVjl).
  - Added XML demo from TimoVjl

To upgrade compiler with fSDK kit download installer from github and run it
fSDK can be browsed on GitHub

[Grincheux]

How is it possible to help you in this project?

[frankie]

Hello Grincheux
Thank-you very much.
fSDK is an open project!
You can help testing the SDK in all your projects and, when you found a problem, creating a small piece of code (the smaller the better) that reproduce the problem. Then send it to me so I can debug it.
If you have already found the problem on yourself please post the fix here, or send it to me.
You can also create a branch on GitHub to publish the fix, add more features, modify something, ... 

[Grincheux]

I told to Timo that there were 2 undefined

Code: [Select]

#define THREAD_MODE_BACKGROUND_BEGIN 0x00010000
#define THREAD_MODE_BACKGROUND_END 0x00020000

It is ok with the new version downloaded yestirday.

Ok for the tests.

[frankie]

Thanks Grincheux 

[TheRaven]

O.O
Awesome (even now)!

[frankie]

New version:
- 11/Feb/2018 -- V.0.1.2.2 (11.Feb.2018)
  - Fixed header MsHtmlC.h
  - Added WMI Samples

To upgrade compiler with fSDK kit download installer from github and run it
fSDK can be browsed on GitHub

Last pre-release

[MrBcx]

To upgrade compiler with fSDK kit download installer from github and run it
fSDK can be browsed on GitHub
-------------------------------------------------------------------
This is a challenging project targeting the port of the last complete MS SDK to pellesC.

^^^ That is an understatement, if ever there was one!

Fantastic accomplishment Frankie and one that is very much appreciated.   Thank You!

[frankie]

To upgrade compiler with fSDK kit download installer from github and run it
fSDK can be browsed on GitHub
-------------------------------------------------------------------
This is a challenging project targeting the port of the last complete MS SDK to pellesC.

^^^ That is an understatement, if ever there was one!

Fantastic accomplishment Frankie and one that is very much appreciated.   Thank You!

Thank you 
I haven't received any complain up to now, so I was thinking that nobody was interested as usually happens here, apart of course the few that are still active on the site. But the starring of the project on Git-Hub is increasing recently, so I'm thinking that maybe someone will report bugs or requests. For this reason I'll wait still some time then I'll release it.
Thanks again to all that liked my work  .

[frankie]

As soon as the poinst will be fixed I'll upgrade the fSDK to meet new PellesC Release 9.00.
The new installer will not overwrite distribution SDK, but copy files in a back-up directory to allow working with both SDK.
The new PellesC 9.00 release allows the upgrade of IDE and tools without reinstalling the SDK, this will allow tools upgrading without reinstalling the SDK.
If you have any comment/feature request could be the right time to talk.

[TimoVJL]

What symbols are missing ?
What headers and libs are used ?

[cnoob]

What symbols are missing ?
What headers and libs are used ?

It was very careless of me. I can't reproduce the problem. fSDK is fine. The error I mentioned happen when I tried SGL (a project on this forums to shorten code of Win32 App, seem to be dead). It failed to build because missing symbol when I correctly linked sgl64.lib. I thought it's because I didn't have the correct header installed (most of the time I fix this type of error is to add missing header) so I try fSDK instead full WinSDK (too big) but receive the same error. The fault is of SGL, not fSDK. I'm sorry 

[Abraham]

It should be

Code: [Select]

            typedef struct foo
            {
                int bar;
            } foo;
not

Code: [Select]

            struct foo
            {
                int bar;
            } foo;
The missing typedef is there for so long.

The verbose way would be

Code: [Select]

            typedef struct foo foo; // global namespace

            struct foo // tagname namespace
            {
                int bar;
            };
tagname namespace is "shared" between struct, union, and enum.

Code: [Select]

            struct some
            {
                int a;
            };
            union some // collision of tagname
            {
                int a;
                unsigned int b;
            };
If an object is simple enough to be typecast, do not use unions.

Note that

Code: [Select]

            struct foo
            {
                int bar;
            } foo;
Is a "struct foo" global object of name "foo".

[frankie]

It should be

Code: [Select]

            typedef struct foo
            {
                int bar;
            } foo;
not

Code: [Select]

            struct foo
            {
                int bar;
            } foo;
The missing typedef is there for so long.

thank you for the report, I fixed the post.

[Abraham]

xnamath.h

You are missing some intrinsic "type punning" with those unions.

Obviously Line 1701

Code: [Select]

return reinterpret_cast<const __m128 *>(&vTemp)[0];could compile in C?

Ah, my casting.

Cannot cast from __m128i to __m128 without first creating a union then type pun it.

The fundamental flaw is that __m128 (floats), __m128i (integers), and __m128d (doubles) are introduced at different times, so they are defined as separate unions. We need to combine them into a single universal union for easier type punning instead of reinterpret_cast from C++.