Agree, first it should be deprecated.
сб, 18 июн. 2022 г. в 15:12, Philipp Klaus Krause :
>
> Am 17.06.22 um 20:55 schrieb Sergey Belyashov:
> > Imho, it is better to break the use of __at in favor of __at().
> >
>
> While that might be a clenaer solution, and was my fi
Imho, it is better to break the use of __at in favor of __at().
пт, 17 июн. 2022 г. в 20:13, Philipp Klaus Krause :
>
> While working on C2X support in the parser (and trying to fix some
> existing bugs in the parser already affecting earlier standards), it
> looks to me like SDCC painted itself
I think, built-ins may cover all basic algorithms which pure C
implementation is very very complex (count of bits, odd/even...) but in
assembler is very simple. Moreover, in most cases these functions should be
inline, because call prepare+prolog+epilog of function some time took same
number of res
Hi,
Try use CXXFLAGS and LDFLAGS instead:
CXXFLAGS=... LDFLAGS=... ./configure ...
Or run configure with --help key, possible it has special option for the
boost path.
___
Sdcc-user mailing list
Sdcc-user@lists.sourceforge.net
https://lists.sourceforge.ne
If you update from svn (not clean build) working copy, then just run
config.status in build directory and than make all.
Best regards,
Sergey Belyashov
пт, 11 сент. 2020 г., 21:25 William Henning :
> I am trying to build the latest version. The compiler and pic16 libraries
> build fine, b
t solution.
Best regards,
Sergey Belyashov
чт, 9 июля 2020 г., 20:30 Philipp Klaus Krause :
> Am 09.07.20 um 19:06 schrieb Sergey Belyashov:
> > Imho, permit concatenation with same prefixes (if first of strings is
> > prefixed but some other not, permit too), reject if they d
Imho, permit concatenation with same prefixes (if first of strings is
prefixed but some other not, permit too), reject if they differs.
чт, 9 июля 2020 г., 19:57 Philipp Klaus Krause :
> In the C standard it is implementation-defined what happens if prefixed
> string literals can be concatenated
Hi,
Philipp changed author of patch #326 in commit:
https://sourceforge.net/p/sdcc/code/11694/ to me, but I'm not the author of
this patch. Please revert back.
Best regards,
Sergey Belyashov
___
Sdcc-user mailing list
Sdcc-user@lists.sourceforg
Please use standard names: sf, zf, hf, pf, nf, cf.
Philipp, I suggest to put some small help to the header of all peephole
definition files.
чт, 4 июн. 2020 г. в 16:57, Sebastian Riedel :
> On 04.06.20 14:20, Sergey Belyashov wrote:
> > Z80 has the same flags as GBZ80, but also i
Z80 has the same flags as GBZ80, but also it has p/v (parity or signed
overflow) and S (sign) flags. And order is different:
GBZ80: Z N H C 0 0 0 0
Z80: S Z 5 H 3 P/V N C
чт, 4 июн. 2020 г. в 15:12, Sebastian Riedel :
> On 04.06.20 12:44, Philipp Klaus Krause wrote:
> > The barrier works the oth
Hi,
I think it is historical. Rewrite rules, recompile, run regression tests
and publish the patch.
ср, 3 июня 2020 г., 22:26 Sebastian Riedel :
> Is there a reason why one wouldn't use: (z80/gbz80)
>
> replace restart {
> ld %2, %4
> ld %1, %3
> } by {
> ; broug
Hi,
Imho, it is better to add two pair rules which converts (hl+/-) to ldi/ldd.
And others use ldd/ldi only.
ср, 27 мая 2020 г., 10:24 Sebastian Riedel :
> Hi, another question regarding the peep hole optimizer.
>
> How does it handle aliases?
>
> Need there be doubled rules for ldd a,(hl)/ld a,(
Hi,
Do $ apt-cache search texinfo
Select proper package and do $ apt-get install
вс, 17 мая 2020 г., 15:46 Saurabh Bansode :
> Hello,
> I am using Ubuntu 16.04, Installed boost via apt-get install
> libboost-system-dev, and then ran ./configure --disable-pic14-port
> --disable-pic16-port as I
Hi,
notUsed isn't implemented for gdbz80 port.
сб, 16 мая 2020 г., 14:58 Sebastian Riedel :
> Hi,
>
> I’m trying to write peep hole rules for the case that A can be trashed.
> I thought `notUsed('a')` is for that use case, at least it looked like
> that in the peeph-gbz80.def.
> But when I compil
Hi,
C have not tool to control output section order. Strictly speaking, high
level languages does not known anything about sections/segments. It is a
linker job.
So, you should prepare special data for each linker type.
Best regards,
Sergey Belyashov
вс, 26 апр. 2020 г. в 22:02, Philipp Klaus
Hi,
Z80 port has --constseg support, but it looks incomplete.
int f1() {
static const char text[] = "some text"; //text will be placed in _CODE
segment
puts(text);
}
int f2() {
puts("some text"); //text will be placed in constseg segment
}
Sergey Belyashov
пн, 13
Hi,
SDCC uses very simple scheme for storing initialization data. It generates
two copies of variables, one copy is initialization data which is stored
after code section, and copied to variables area on startup. I think, it is
caused by very simple linkers supported. More complex linkers can do th
hl
ld -2(ix),l
ld -1(ix),h
ld hl,100$
ex (sp),hl
push bc
ret
100$:
ld a,-4(ix)
out (BANK_PORT),a
ld c,-2(ix)
ld b,-1(ix)
push bc
ret
Best regards,
Sergey Belyashov
сб, 7 марта 2020 г., 22:15 Sergey Belyashov :
> Hi,
> My project is used about 200 ROM pages and 33 R
are expected. Most of callers will have their own local
variables. These four bytes are required to make banked call cheaper: do
not use tranpoline in caller or move frame in banked_call implementation.
Best regards,
Sergey Belyashov
вс, 8 марта 2020 г., 6:26 Alan Cox :
> > 5. Add optimizat
(BANK_PORT),a
ld c,-2(ix)
ld b,-1(ix)
push bc
ret
Best regards,
Sergey Belyashov
___
Sdcc-user mailing list
Sdcc-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sdcc-user
20 matches
Mail list logo