Recent Posts

1

Projects developed with Pelles C / Re: vCardz_i an address book/contact list program

« Last post by John Z on May 14, 2022, 03:32:00 pm »

In case any forum member are using vCardz_i - A new version is posted on SourceForge - https://sourceforge.net/projects/vcardz/  which was updated using Pelles C version 11.  Version 2.5.3.2 has many improvements over 2.1.5 mentioned above. 

Feedback always welcome.

John Z

2

Bug reports / Re: Pelles 11 'Find dialog' possible non-critical bug

« Last post by frankie on May 11, 2022, 01:34:07 pm »

Yes, I confirm the bug.

3

Bug reports / Pelles 11 'Find dialog' possible non-critical bug

« Last post by John Z on May 11, 2022, 01:19:28 pm »

Has anyone else experienced this?  I don't think it is my system or installation.

Running Windows 11,  plus Pelles C version 11.

When I use the Find dialog box and pick a previously searched find string from the drop down list the 'Find Next' button does not become active.  It seems it is only active after actually typing in the 'Find What' field. Kind of defeats the purpose of the drop down box.  Example screen shots attached.

John Z

4

Bug reports / Re: expected 'signed char' but found '__m128i'

« Last post by frankie on May 08, 2022, 09:16:12 pm »

Thanks for explanation Pelle

5

Bug reports / Re: expected 'signed char' but found '__m128i'

« Last post by Pelle on May 08, 2022, 06:52:37 pm »

Unfortunately you are allowed to initialize individual elements of a SIMD type with MSVC (it would have been much better if SIMD types always behaved like BLOBs, like in other compilers). This complicates parsing of initializers for SIMD types (with or without /Ze): it can be a SIMD, or a union of arrays with various scalar types and lengths.

The initializer parsing needs to be rewritten anyway, for other/better reasons than this, but this is going to take some considerable time...

6

Bug reports / Re: expected 'signed char' but found '__m128i'

« Last post by frankie on May 08, 2022, 02:52:25 pm »

You're welcome Robert.
I made some more tests, specifically using the zero initialization for structures:

Code: [Select]

  struct processed_utf_bytes previous = {0};
This, as should be expected, works.
But compiling your header file triggers an internal compiler error:

Code: [Select]

fatal error: Internal error: cast_tree(#1).
Maybe Pelle can have a look on it.
On the other hand, using the same initialization in a simple program as below compiles and runs without errors:

Code: [Select]

#include <stdio.h>
#include <immintrin.h>

struct test__m128i
{
  __m128i rawbytes;
  __m128i high_nibbles;
  __m128i carried_continuations;
};

int main(int argc, char *argv[])
{
struct test__m128i t = {0};
printf("Hello World! %lld\n", *(long long int *)&t);
}
At this point is hard to say where the problem is, probably inside the compiler itself.
Let wait for Pelle... 

7

Bug reports / Re: expected 'signed char' but found '__m128i'

« Last post by Robert on May 06, 2022, 08:37:58 pm »

Thanks frankie, that gives me a go ahead for now.

8

Bug reports / Re: expected 'signed char' but found '__m128i'

« Last post by frankie on May 06, 2022, 02:33:49 pm »

Pelles C 11.00.2, This code

Code: [Select]

struct processed_utf_bytes previous = {.rawbytes = _mm_setzero_si128(),
                                       .high_nibbles = _mm_setzero_si128(),
                                       .carried_continuations =
                                           _mm_setzero_si128()};

is producing these errors

Code: [Select]

simdutf8check.h(162): error #2082: Invalid initialization type; expected 'signed char' but found '__m128i'.
simdutf8check.h(163): error #2082: Invalid initialization type; expected 'signed char' but found '__m128i'.
simdutf8check.h(165): error #2082: Invalid initialization type; expected 'signed char' but found '__m128i'.

Apparently you can't call intrinsic function inside an initialization. Changing initialization to:

Code: [Select]

struct processed_utf_bytes previous;
previous.rawbytes = _mm_setzero_si128();
previous.high_nibbles = _mm_setzero_si128();
previous.carried_continuations = _mm_setzero_si128();
Seems to work.
Anyway there are still many warnings:

Code: [Select]

warning #2071: Overflow or truncation in constant expression.


9

Bug reports / expected 'signed char' but found '__m128i'

« Last post by Robert on May 06, 2022, 06:38:33 am »

Pelles C 11.00.2, This code

Code: [Select]

struct processed_utf_bytes previous = {.rawbytes = _mm_setzero_si128(),
                                       .high_nibbles = _mm_setzero_si128(),
                                       .carried_continuations =
                                           _mm_setzero_si128()};

is producing these errors

Code: [Select]

simdutf8check.h(162): error #2082: Invalid initialization type; expected 'signed char' but found '__m128i'.
simdutf8check.h(163): error #2082: Invalid initialization type; expected 'signed char' but found '__m128i'.
simdutf8check.h(165): error #2082: Invalid initialization type; expected 'signed char' but found '__m128i'.
where the struct members have been declared as

Code: [Select]

struct processed_utf_bytes {
  __m128i rawbytes;
  __m128i high_nibbles;
  __m128i carried_continuations;
};

in the attached file.

10

Bug reports / Re: Linker Error *SOLVED*

« Last post by Mikael on April 30, 2022, 10:56:33 pm »

Yeah, I can understand it's a thankless task to keep adding support for an endless stream of new stuff no-one cares about. Sorry if I sounded negative, I was just tearing my hair out for a while...

I couldn't care less about the vast majority of stuff Microsoft adds either, but once in a while they do come up with something useful.

VirtualAlloc2 gives us the power to do some nifty things, like growing an array without moving any data around, just tell the OS to remap virtual memory. Calling the OS is pretty slow to begin with, but it can still be a win for big arrays.

Thanks again for the help!