Many thanks David,
After weeks of constant failure on regression tests (ret_longdouble_test2...
failure) I automatically run after each new commit, my beloved Raspberry Pi is
again happy with tcc.
Christian
From:
Hi all,
Because my automatic test has been broken for several days I don't know
exactly when but ret_longdouble_test2 fails on ARM (on RPi)
ret_longdouble_test2... failure
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
: 'Christian JULLIEN'
Subject: Re: [Tinycc-devel] New warning on ARM since few commits
Hi,
On Sat, 12 Apr 2014, Christian Jullien wrote:
No comments about this issue?
It shouldn't be a new warning. The code in question (libtcc1.c, ldouble_long
type, the __fixunsxfdi function, and the definition
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
Hi, semicolon is not superfluous because you're able to define an object
after structure definition. It allows you for example to define anonymous
structure like;
struct {
int x;
int y;
} myobj ;
Semicolon is there to say, definition is done AND I don't define variable of
this type
Hi all,
This is just a question I'm doing for myself.
Currently, when you do a ./configure; make you generate a tcc that is a gcc
compiled program implementing a tcc compiler. So far so good.
The challenge for a compiler is to compile itself.
Long ago, you were able to bootstrap gcc with any
[mailto:tinycc-devel-bounces+eligis=orange...@nongnu.org] On Behalf Of Thomas
Preudhomme
Sent: lundi 10 mars 2014 02:19
To: tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] tcc bootstrap
Le 2014-03-10 07:27, Patrick Lauer a écrit :
On Sunday 09 March 2014 17:12:11 Christian Jullien wrote:
I think
I second Daniel that a feature should not be removed before an obsolescent
period.
IMHO, you should add a warning now and you can remove OABI and FPA right
after the tcc new version is released.
If you prefer, you can fork arm file to xxx-oabi.c now and leave a comment
for those who want OABI
Currently, there is no (or very limited) number of warnings. No warning at
all on RPi, on Fedora 20 x86_64 on Windows x86/X86_64
IMHO, it's worth to chase them all and then add -Werror which does not
introduce new warnings to test. It only refuses new changes that add new
warnings.
I strongly
Since you pay attention of warnings (as -Wall implies, which is good), may I
suggest to also add -Werror ?
For all the projects I worked on, I always fought developers to improve code
quality. I never tolerated any warnings as, most of the time, a warning may
be a potential bug.
With nice
I confirm it works on RPi
I only have this warning:
In file included from ../libtcc.c:41:
./tccrun.c:245: warning: implicit declaration of function '__clear_cache'
- Message d'origine -
De : Thomas Preud'homme lt;robo...@celest.frgt;
Date ven. 07/02/2014 15:33 (GMT +01:00)
À :
] On Behalf Of
Christian JULLIEN
Sent: vendredi 7 février 2014 16:52
To: tinycc-devel@nongnu.org
Subject: [Tinycc-devel] RE :Re: Recent change breaks test3 on RPi
I confirm it works on RPi
I only have this warning:
In file included from ../libtcc.c:41:
./tccrun.c:245: warning: implicit declaration
gt; tcc should also honnor -fsigned-char or -funsigned-char option (gcc
gt; compatible) to change char default sign.
There is already such an option at least.
Ok, in that case --help should mention this
./tcc --help
tcc version 0.9.26 - Tiny C Compiler - Copyright (C) 2001-2006 Fabrice
Hi All,
This failure is new on RPi
test3
../tcc -B.. -I.. -I.. -I../include -DCONFIG_LDDIR=\lib\
-DCONFIG_MULTIARCHDIR=\arm-linux-gnueabihf\ -DTCC_TARGET_ARM
-DTCC_ARM_EABI -DTCC_ARM_HARDFLOAT -DTCC_ARM_VFP -DONE_SOURCE -run ../tcc.c
-B.. -I.. -I.. -I../include
Hi all,
Playing with clock_gettime and uint64_t/int64_t on RPi (ARM), I discovered
uint64_t = double conversion bug :
NOTE: tcc is compiled using --with-libgcc
#include stdint.h
#include stdio.h
static uint64_t x1 = 123456789;
static uint32_t x2 = 123456789;
int
main()
{
Hi,
This commit raises this error on RPi
aaa=3
@@ -556,7 +556,7 @@
__builtin_types_compatible_p(int *, int *) = 1
__builtin_types_compatible_p(int *, void *) = 0
__builtin_types_compatible_p(int *, const int *) = 0
-__builtin_types_compatible_p(char *, unsigned char *) = 0
...@gmx.de]
Sent: mardi 21 janvier 2014 15:04
To: Christian JULLIEN; tinycc-devel@nongnu.org
Cc: Kirill Smelkov
Subject: Re: [Tinycc-devel] RE :Re: mksh fails to build with bound check
I guess these errors aren't new, just make test didn't stop previously. I'll
leave it to others to decide
Compiling tcc on Windows x64 using gcc, I get:
../tccpe.c:1806:26: warning: cast from pointer to integer of different size
[-Wpointer-to-int-cast]
pe-start_addr = (DWORD)tcc_get_symbol_err(s1, start_symbol);
On Windows, DWORD is a 32bit value, you can't cast a 64bit pointer to a
If I recall correctly, MAP_ANONYMOUS is not true POSIX but an extension most
un*x system implement yet it is sometime call MAP_ANON.
To be portable, you must detect in ./configure if mmap supports either
MAP_ANONYMOUS or MAP_ANON and fallback to /tmp/xxx if none is supported.
Christian
-
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
Hi Daniel,
If I understand you well, you want to change RELICENSING file
From
Daniel Glöckner? arm-gen.c
To
Daniel GlöcknerNO arm-gen.c
On my side, I prefer a clear NO than a '?'
If I've really no objection about your choice (I
tcc and gcc are different. Gcc is an external tool that produces binary. There
is absolutely no problem for gcc to have GPL License.
The same applies for tcc for those who use tcc to produce binaries.
A very nice feature of tcc is that it can be used embedded in another program
to compile C code
So to answer your question you need to look at what files are used to create
libtcc on your system and the copyrights of each of those files
Currently, if you have a look at files having (or supposed to be) BSD Like
license (as x86_64-gen.c), the header is talking only about LGPL. It is
Hi all,
On my RPI tcc is 425082 bytes, after stripping it goes to 133148 which is a
big save, especially on small devices and low performance I/O on SD card.
Ok, I can strip myself but most GNU generated Makefile has both install and
install-strip entries.
WDYT?
Christian
Hi Thomas,
I this morning I decided to test your new runtime functions on my beloved
RPi.
As usual, I've made a fresh install:
And tried to compile using only ./configure; make
Then, I got:
test3
../tcc -B.. -I.. -I.. -I../include -DCONFIG_LDDIR=\lib\
gt; Normally an inline must be locally defined ?
Makes sense.
NO! inline keyword is only a hint to the compiler that a function is supposed
small enough to be inlined. It's perfectly legal for a compiler to ignore this
hint if it thinks it's too complicated to inline this code.
C standard
:
Le mercredi 11 décembre 2013, 09:28:07 Christian
JULLIEN a écrit :
I knew about the fact that it is a hint and I knew
even when inlined, the
function still needs to be output in case its
address is used. However I
Jared,
The file is in win32/include/math.h
It contains two times the exact same inline definition which is obviously an
error for any C compiler.
To me, all code from line 734 to 778 is duplicated from line 669 and should
be removed.
This file, as copyright notice says, is part of the w64
Hi x86/x64 Windows maintainers,
This morning, I chekouted the latest TCC version from mod to update tcc on
Windows (both x86/x64).
It used to work flawlessly (my latest successful build was made on October).
Using gcc, I now get an error because __fp_unordered_compare (long double x,
2013, Christian Jullien wrote:
Hi x86/x64 Windows maintainers,
This morning, I chekouted the latest TCC version from mod to update tcc on
Windows (both x86/x64).
It used to work flawlessly (my latest successful build was made on
October).
Using gcc, I now get an error because
Sorry if doubled-posted but previous replied (with attachment) has not been
distributed.
-Original Message-
From: Christian Jullien [mailto:eli...@orange.fr]
Sent: samedi 7 décembre 2013 12:26
To: 'tinycc-devel@nongnu.org'
Subject: RE: [Tinycc-devel] __fp_unordered_compare defined twice
a translator may use to implement any call to the function
in the same translation unit. It is unspecified whether a call to the
function uses the inline definition or the external definition. - C99,
6.7.4
Rob
On Sat, 7 Dec 2013, Christian Jullien wrote:
Sorry if doubled-posted but previous
Message-
From: Thomas Preud'homme [mailto:robo...@celest.fr]
Sent: dimanche 17 novembre 2013 11:32
To: tinycc-devel@nongnu.org
Cc: Christian JULLIEN
Subject: Re: [Tinycc-devel] RE : RE : Rewrite of gfunc_call on arm
Le dimanche 17 novembre 2013, 13:38:12 Thomas Preud'homme a écrit :
Le
-Original Message-
From: Thomas Preud'homme [mailto:robo...@celest.fr]
Sent: dimanche 17 novembre 2013 17:09
To: tinycc-devel@nongnu.org
Cc: Christian JULLIEN
Subject: Re: [Tinycc-devel] RE : RE : Rewrite of gfunc_call on arm
Le dimanche 17 novembre 2013, 13:38:12 Thomas Preud'homme a écrit
Except for
ret_2float_test... Segmentation fault
Already mentioned long ago, nothing seems to be broken.
NOTE: I see:
use libgccnbsp;nbsp;nbsp;nbsp;nbsp;nbsp; yes
Creating config.mak and config.h
make: Circular Makefile lt;- Makefile dependency dropped.
But I can't tell you if it new of not.
d'origine -
De : Christian JULLIEN lt;eli...@orange.frgt;
Date ven. 15/11/2013 12:47 (GMT +01:00)
À : tinycc-devel@nongnu.org lt;tinycc-devel@nongnu.orggt;
Objet : [Tinycc-devel] RE : Rewrite of gfunc_call on arm
Except for
ret_2float_test... Segmentation fault
Already mentioned long ago, nothing
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
Hi,
I think there no correct result.
On a system that requires 64bit alignment for uint64_t, a C compiler is free
to add extra bytes padding for either processor required alignment or for
faster access.
With gcc you can force packed struct using __attribute__((__packed__))
But compiler may add
Hi Daniel
ARM hardfloat: fix struct return with float/double args
Fixes the case where the structure is not returned in registers.
I thought it was related to ret_2float_test
At least on Rpi I still have:
ret_2float_test... Segmentation fault
C.
P.S. Compiled from a fresh git
-Original
: mercredi 1 mai 2013 17:23
To: tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] Recent changes segfault on Linux ARM
Hi Christian,
On Wed, May 01, 2013 at 04:44:09PM +0200, Christian Jullien wrote:
ARM hardfloat: fix struct return with float/double args
Fixes the case where the structure
Wouah!
Do you mean you're implementing eabi_ functions for ARM (such as div and
mod ?)
Total respect!!
I started to find a fast algo for div then tried to figure out how to implement
this in assembler and found it's a real pain.
I understand why your progress it is very (very) slow.
I you
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
;tinycc-devel@nongnu.orggt;
Cc : Christian Jullien lt;eli...@orange.frgt;
Objet : Re: [Tinycc-devel] Yesterday commit breaks tcc (at least on RPi)
Le mardi 19 mars 2013 06:39:40, Christian Jullien a écrit :
gt; Hi guys,
gt;
gt; I've no time to investigate but yesterday commit breaks tcc (at least
Hi guys,
I've no time to investigate but yesterday commit breaks tcc (at least on
RPi)
jullien@sims ~/tinycc $ git pull
From http://repo.or.cz/r/tinycc
d9dfd9c..6ed6a36 mob- origin/mob
Updating d9dfd9c..6ed6a36
Fast-forward
lib/libtcc1.c | 11 +++
tccrun.c |1 +
2
Hi,
The issue you have is not related to TCC, any C compiler will produce this
error.
If you define a variable in a header and you include this header from more
than one translation unit you get this variable define more than once and
linker complains.
-- decl.h
int i = 100;
-- source1.c
tcc_define_symbol(s, __STDC_VERSION__, 199901L);
This define pretends it is ISO/IEC 9899:1999 (E)
Because __STDC_VERSION__ is set to 199901L
This code should be legal:
#include stdio.h
#include wchar.h
int main()
{
#if defined(__STDC_VERSION__) (__STDC_VERSION__ = 199901L)
c:\tmpgit clone git://repo.or.cz/tinycc.git
Cloning into 'tinycc'...
remote: Counting objects: 5037, done.
remote: Compressing objects: 100% (1558/1558), done.
remote: Total 5037 (delta 3427), reused 5037 (delta 3427)
Receiving objects: 100% (5037/5037), 2.30 MiB | 404 KiB/s, done.
Resolving
Sent: samedi 16 février 2013 19:58
To: tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] What C version tcc is supposed to implement?
Hello Christian,
Am 16.02.2013 12:04, schrieb Christian Jullien:
tcc_define_symbol(s, __STDC_VERSION__, 199901L);
This define pretends it is ISO/IEC 9899
...@nongnu.org
[mailto:tinycc-devel-bounces+eligis=orange...@nongnu.org] On Behalf Of
Michael Matz
Sent: samedi 16 février 2013 21:18
To: tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] What C version tcc is supposed to implement?
Am 16.02.2013 20:16, schrieb Christian Jullien:
Thank you
Thanks Thomas, it works again now
-Original Message-
From: Thomas Preud'homme [mailto:robo...@celest.fr]
Sent: samedi 16 février 2013 23:39
To: Christian Jullien
Cc: tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] Can't git clone anymore
Le samedi 16 février 2013 18:28:52, Christian
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
BINGO! This line
ifeq (-$(findstring clang,$(CC))-,-gcc-)
returns false and GCC specific flags are not used, If I remove this if (and
associated else)
tcc compiles with no warnings as yesterday
Christian
- Message d'origine -
De : grischka lt;gris...@gmx.degt;
Date ven. 15/02/2013
Thomas, your last commit lets Makefile find again GCC and now correctly sets
CFLAGS with warnings we want to ignore.
Christian
- Message d'origine -
De : Thomas Preud'homme lt;robo...@celest.frgt;
Date ven. 15/02/2013 14:27 (GMT +01:00)
À : tinycc-devel@nongnu.org
Quick question, I thought that recent config work was supposed to solve, among
other things, the need to pass --with-libgcc (sorry if I'm wrong)
On RPi, while ./configure; make; make test works ROOTB
I still have to use ./configure --with-libgcs if I later do a make install
and want to use tcc
Even my last warning is gone.
My tests include;
tcc compilation
tcc tests
OpenLisp huge non regression tests
To me, tcc works like a charm, you can release 0.9.26 at any time.
I'll test of course all commits you'll do.
Great job all!
Christian
___
Sorry Thomas, made 5 mn ago
As usual, I used a fresh git clone to test:
makeinfo tcc-doc.texi
make -C tests test
make[1]: Entering directory `/home/jullien/tinycc/tests'
hello-exe
../tcc -B.. ../examples/ex1.c -o hello || (../tcc -vv; exit 1) amp;amp;
./hello
tcc:
@nongnu.orggt;, Christian
JULLIEN lt;eli...@orange.frgt;
Objet : Re: [Tinycc-devel] RE :Re: [+Last+] Call for testing
Le jeudi 14 février 2013 16:48:24, Christian JULLIEN a écrit :
gt; hello-exe
gt; ../tcc -B.. ../examples/ex1.c -o hello || (../tcc -vv; exit 1)
amp;amp
This commit still works on RPi but produces
test3
../tcc -B.. -DCONFIG_LDDIR=\lib64\ -DTCC_TARGET_X86_64 -DONE_SOURCE -run
../tcc.c -B.. -DCONFIG_LDDIR=\lib64\ -DTCC_TARGET_X86_64 -DONE_SOURCE -run
../tcc.c -B.. -DCONFIG_LDDIR=\lib64\ -DTCC_TARGET_X86_64 -DONE_SOURCE
After git pull on RPi I get:
$ make make test
gcc -o libtcc.o -c libtcc.c -DTCC_TARGET_ARM -DWITHOUT_LIBTCC -DTCC_ARM_EABI
-DTCC_ARM_HARDFLOAT -DTCC_ARM_VFP -I. -Wall -g -O2 -fno-strict-aliasing
-Wno-pointer-sign -Wno-sign-compare
libtcc.c: In function tcc_new:
libtcc.c:988:29: error:
*/
- Message d'origine -
De : grischka lt;gris...@gmx.degt;
Date mer. 13/02/2013 14:39 (GMT +01:00)
À : tinycc-devel@nongnu.org lt;tinycc-devel@nongnu.orggt;
Objet : Re: [Tinycc-devel] Last two warnings
Christian Jullien wrote:
gt; tccpp.c: In function ‘macro_subst’:
gt; tccpp.c:2803:12
: tinycc-devel-bounces+eligis=orange...@nongnu.org
[mailto:tinycc-devel-bounces+eligis=orange...@nongnu.org] On Behalf Of grischka
Sent: mercredi 6 février 2013 23:23
To: tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] Last two warnings
Christian Jullien wrote:
tccpp.c: In function ‘macro_subst
To: tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] New tcc display info?
Christian Jullien wrote:
(neither -vv nor -print-search-dirs is shown with -help)
True.
So I did my hopefully last commit for 0.9.26:
http://repo.or.cz/w/tinycc.git/commitdiff/82965594
--- grischka
As with gcc or cl, could tcc -v show processor/system info ?
For example, something similar to
tcc –v
tcc version 0.9.25 – WIN32 x86
tcc version 0.9.25 – WIN32 x86_64
tcc version 0.9.25 – Linux ARM hard float
Rationale, on Windows and Linux we can have both 32/64 bit version
CL shows
Microsoft
Still using my OpenLisp benchmarks, tcc 0.9.26 executes code around 10% faster
than previous version 0.9.25 on both 32 and 64bits. Congrat!
## Tiny C Compiler XP (32 bits interpreted)
5.554 s. (prev. version 6.323 s.)
## Tiny C Compiler XP (32 bits compiled VM LAP)
9.931 s. (prev.
Does it give the target architecture or the host architecture?
It should return target, we generally know on which host we are running on
(useless yet can be added as well if you like)
Why not just use readelf or equivalent to detect the platform?
You have no readelf or equivalent on Windows
scripting
capability in your application.
Miki.
On 8 February 2013 00:44, Christian Jullien lt;eli...@orange.frgt; wrote:
Hello Miki,
After ./configure; make
My config.mak contains TARGETOS=Darwin
A very simple hello world
tcc -o hello hello.c
Gives me:
/usr/lib/crt1.o:1: error
Hello, are you sure about order of your arguments?
Please try
$ CC=gcc-m-4.8 ../..configure
instead
Christian
- Message d'origine -
De : Akim Demaille lt;akim.demai...@gmail.comgt;
Date ven. 08/02/2013 09:36 (GMT +01:00)
À : tinycc-devel@nongnu.org lt;tinycc-devel@nongnu.orggt;
Objet :
Here are some benchmarks comparing tcc and gcc 4.7 on RPi.
Benchmarks are taken from my OpenLisp tests suite.
Source files for the bench consist of 15040 lignes of C code generated form
compiler
Summary: tcc compiles x41 faster but runs program x3.37 slower
== gcc
1) compile time
jullien@sims
Gcc is not using -O2 ... but -O3.
Here are the results with -O0:
Gcc loses 20% in program speed but compiles of course much faster
tcc still compiles around 10x faster but runs x3.5 times slower.
Christian
jullien@sims ~/openlisp/cbench $ time make
main.c
gabriel.c
boyer.c
browse.c
ctak.c
Hi,
You use -Wall to compile tcc, that's great it means you pay attention to
compiler warnings (I love that).
So may I insist to remove the very last two remaining warnings:
tccpp.c: In function 'macro_subst':
tccpp.c:2803:12: warning: '*((void *)cval+4)' is used uninitialized in this
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
on
RPi!
- Message d'origine -
De : Thomas Preud'homme lt;robo...@celest.frgt;
Date mar. 05/02/2013 10:52 (GMT +01:00)
À : Christian JULLIEN lt;eli...@orange.frgt;, tinycc-devel@nongnu.org
lt;tinycc-devel@nongnu.orggt;
Objet : Re: [Tinycc-devel] RE :Re: Another ARM (RPi) compiler bug
Le
I've no time tonight but tomorrow I'll try to play with
-z execstack and/or -fno-stack-protection
-Original Message-
From: grischka [mailto:gris...@gmx.de]
Sent: mardi 5 février 2013 11:09
To: tinycc-devel@nongnu.org
Cc: Christian JULLIEN
Subject: Re: [Tinycc-devel] current mod makes
to pass an option to ./configure to allow both, something
--with-run
M2c
C.
-Original Message-
From: tinycc-devel-bounces+eligis=orange...@nongnu.org
[mailto:tinycc-devel-bounces+eligis=orange...@nongnu.org] On Behalf Of
Christian Jullien
Sent: mardi 5 février 2013 20:17
To: 'grischka
I continue to test mod before 9.26
I only have access on Snow Leopard on which mod produces this error.
Any MacOS enthusiasts to investigate? Maybe because MacOS does not support
static compilation.
uname -a
Darwin macos.local 10.8.0 Darwin Kernel Version 10.8.0: Tue Jun 7 16:33:36 PDT
2011;
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
A clean reinstall on RPi gives me this error (as 2 warnings already known):
jullien@sims ~/tinycc $ make
gcc -o tccrun.o -c tccrun.c -DTCC_TARGET_ARM -DWITHOUT_LIBTCC -DTCC_ARM_EABI
-DTCC_ARM_HARDFLOAT -DCONFIG_MULTIARCHDIR=\arm-linux-gnueabihf\ -DTCC_ARM_VFP
-I.nbsp; -Wall -g -O2
I confirm it's fixed on RPi
- Message d'origine -
De : grischka lt;gris...@gmx.degt;
Date lun. 04/02/2013 17:28 (GMT +01:00)
À : Christian JULLIEN lt;eli...@orange.frgt;, tinycc-devel@nongnu.org
lt;tinycc-devel@nongnu.orggt;
Objet : Re: [Tinycc-devel] RE : Cleanups
Christian JULLIEN
Just checked this morning on F18 x86_64
[jullien@fedora18 tinycc]$ gcc -v
Using built-in specs.
COLLECT_GCC=/usr/bin/gcc
COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/4.7.2/lto-wrapper
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
Thanks, the reported bug is fixed yet my bignum lib still makes a segment
violation.
I'll investigate and return to you with a new test case.
C.
-Original Message-
From: tinycc-devel-bounces+eligis=orange...@nongnu.org
[mailto:tinycc-devel-bounces+eligis=orange...@nongnu.org] On Behalf
Wrong double - int conversion when double is 0
#include stdio.h
int
main()
{
// Ok if d 0
{
double d = 413.55;
long l = (long)d;
int i = (int)d;
printf(d=%f, i=%d (OK), l=%ld (OK)\n, d, i, l);
}
// BUG if d 0
{
-bounces+eligis=orange...@nongnu.org] On Behalf Of
Daniel Glöckner
Sent: dimanche 3 février 2013 23:52
To: tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] Another ARM (RPi) compiler bug
On Sun, Feb 03, 2013 at 10:14:16PM +0100, Christian Jullien wrote:
Wrong double - int conversion when double
-206.60 -206
-Original Message-
From: tinycc-devel-bounces+eligis=orange...@nongnu.org
[mailto:tinycc-devel-bounces+eligis=orange...@nongnu.org] On Behalf Of
Christian Jullien
Sent: lundi 4 février 2013 07:24
To: tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] Another ARM (RPi) compiler
Hi all,
Trying to port my (extremely portable Bignum lib:
http://sourceforge.net/projects/bigz/)
This code snippet, isolates the error I get:
#include stdio.h
void
main(void)
{
double dbl = 44.361420;
size_t len = 2;
printf(1) =%f %ld\n, dbl, len);
printf(2)
::
0 test(s) failed
-Original Message-
From: tinycc-devel-bounces+eligis=orange...@nongnu.org
[mailto:tinycc-devel-bounces+eligis=orange...@nongnu.org] On Behalf Of
Christian Jullien
Sent: samedi 2 février 2013 12:26
To: tinycc-devel@nongnu.org
Subject: [Tinycc-devel] ARM (RPi
From: Jerry Reed [mailto:irving2nd2...@gmail.com]
Sent: mardi 29 janvier 2013 22:52
To: Thomas Preud'homme
Cc: Christian Jullien; tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] Segmentation Faults and test failures on ARM
(Raspbian/Raspberry Pi)
Hi:
It will be tomorrow EST before I can
I confirm it also works when boostraped with gcc 4.7.2 on my fully up to
date RPi
$ uname -a
Linux *** 3.6.11+ #362 PREEMPT Tue Jan 22 14:52:21 GMT 2013 armv6l GNU/Linux
$ gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/arm-linux-gnueabihf/4.7/lto-wrapper
At this point, had to make a number of symlinks/copies into
/usr/include of header files from:
/usr/include/arm-linux-gnueabihf/ to /usr/include
These shouldn't be necessary. Can you try without the symlink and the mob
branch?
Hi, I just tested mod on my RPi and I still
I started to implement (some) of __aeabi missing functions, yet in C (see
attachments)
It is the right direction to go?
-Original Message-
From: tinycc-devel-bounces+eligis=orange...@nongnu.org
[mailto:tinycc-devel-bounces+eligis=orange...@nongnu.org] On Behalf Of
Christian Jullien
Trying to compile, tcc on my rpi box,
./configure complains because my proc (armv6l) is not supported.
$ uname -m
armv6l
arm|armv4l|armv5tel|armv5tejl|armv6j|armv7a|armv7l)
arm|armv4l|armv5tel|armv5tejl|armv6j|armv6l|armv7a|armv7l)
Fixed the issue.
Also:
$ more
Sorry if this is a known issue.
I'm Still running tcc on my Rpi box.
1) Download latest mob
2) Patch configure to add armv6l
3) ./configure; make - no error, 1 warning tccelf.c:1327:12:
warning: 'tcc_add_support' defined but not used
4) make install
So, expect for
, __aeabi_d2ulz)
are defined in ./libgcc/config/arm/* files from gcc source code.
-Original Message-
From: Thomas Preud'homme [mailto:robo...@celest.fr]
Sent: dimanche 11 novembre 2012 20:18
To: tinycc-devel@nongnu.org
Cc: Christian Jullien
Subject: Re: [Tinycc-devel] __aeabi_idivmod
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
Paulo,
I'm not tcc maintainer and I only speak for myself, reference falls in the
same category: incompatible C extensions.
This a debate that sooner than later comes for C implementations: should we
extend the language to make a better C?
There is already a better C which is called C++,
Personally, I can't say that tcc execution speed is close to other C
compilers.
Results from my OpenLisp benchmarks (see www.eligis.com)
All tests are run on the same machine.
You can see that TCC (Tiny C Compiler) is always between 3x to 10x slower!!!
than gcc and/or VC++.
The funny thing is
PM, Christian Jullien eli...@orange.fr wrote:
You can see that TCC (Tiny C Compiler) is always between 3x to 10x slower!!!
than gcc and/or VC++.
The original post being referred to was from me (i remember it well). Just
to clarify: my (continued) amazement was/is solely about the compilation
speed
As long as my opinion matters, I prefer to reserve C++ features to C++
compilers.
In order to keep tcc small and simple, I prefer to stick to C only features.
I think it's better to implement C11 features than to hack few C++ goodies.
M2c
Christian
-Original Message-
From:
Julia,
See also my reply for the same question
http://lists.nongnu.org/archive/html/tinycc-devel/2010-04/msg2.html
It contains an attached zip with source code that works on any branch.
Christian
From: tinycc-devel-bounces+eligis=orange...@nongnu.org
Or even better:
#!/usr/bin/env tcc -run
-Original Message-
From: tinycc-devel-bounces+eligis=orange...@nongnu.org
[mailto:tinycc-devel-bounces+eligis=orange...@nongnu.org] On Behalf Of
grischka
Sent: jeudi 7 juillet 2011 02:43
To: tinycc-devel@nongnu.org
Subject: Re: [Tinycc-devel] Allow
The same curious can compare performances (expressed in second) between gcc
and tcc (both 32/64 bits) on the same Windows 7 x64 machine.
While LAP compiled code (interpreted by a virtual machine) is generally
twice as fast as the interpreter, tcc is two times slower.
GNU GCC 4.5 (Mgw) XP
701 - 800 of 824 matches
Mail list logo
