digitalmars.D - Exceptional News

David Nadlinger (16/16) Feb 11 2013 Looks like we are finally getting somewhere. :)

Iain Buclaw (5/21) Feb 11 2013 Rather amusing. :-)
Andrei Alexandrescu (3/18) Feb 11 2013 Awesome!!!

Iain Buclaw (7/28) Feb 11 2013 This feature has been around for some time (somewhere in lexer.c iirc) -

Jesse Phillips (4/7) Feb 14 2013 LLVM hasn't had support for Exceptions in Windows, Hence the

=?UTF-8?B?U8O2bmtlIEx1ZHdpZw==?= (3/20) Feb 11 2013 Great news!
Dmitry Olshansky (5/20) Feb 11 2013 Nice one!

David Nadlinger (5/7) Feb 14 2013 It would be 4; MinGW for Win64 uses SEH, and LLVM still doesn't

Nick Sabalausky (8/26) Feb 11 2013 I took my mind a minute to process the point of that ("Is

David Nadlinger (12/15) Feb 14 2013 Don't get too excited. ;) There is still a lot to do, several

bearophile (5/9) Feb 14 2013 Why don't you create one zip or rar, that a user unpacks, adds a

David Nadlinger (6/13) Feb 14 2013 Because it's not ready for general consumption yet. For binary

bearophile (6/7) Feb 14 2013 If trying the compiler is so complex/hard/time consuming as it

David Nadlinger (6/11) Feb 14 2013 If you don't want to contribute to LDC development, you shouldn't

Andrej Mitrovic (7/11) Feb 14 2013 I suppose I'll give this a try tonight. However these steps should
jerro (5/7) Feb 14 2013 It seems there is a bug in the MinGW TLS support patch linked

David Nadlinger (4/11) Feb 14 2013 Yes, always those last-minute refactorings…

"David Nadlinger" <see klickverbot.at> writes:

Looks like we are finally getting somewhere. :)

---
version (Windows) {} else static assert(false);
void main() {
     try {
     	throw new Exception("This exception...");
     } catch (Exception e) {
     	import std.stdio;
     	writeln(e.msg, " ... has been caught for you by ", 
__VENDOR__, ".");
     }
}
---
This exception... ... has been caught for you by LDC.
---

David

Feb 11 2013

Iain Buclaw <ibuclaw ubuntu.com> writes:

On 11 February 2013 13:00, David Nadlinger <see klickverbot.at> wrote:

 Looks like we are finally getting somewhere. :)

 ---
 version (Windows) {} else static assert(false);
 void main() {
     try {
         throw new Exception("This exception...");
     } catch (Exception e) {
         import std.stdio;
         writeln(e.msg, " ... has been caught for you by ", __VENDOR__,
 ".");
     }
 }
 ---
 This exception... ... has been caught for you by LDC.
 ---

 David


Rather amusing. :-)

-- 
Iain Buclaw

*(p < e ? p++ : p) = (c & 0x0f) + '0';

Feb 11 2013

Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:

On 2/11/13 8:00 AM, David Nadlinger wrote:
 Looks like we are finally getting somewhere. :)

 ---
 version (Windows) {} else static assert(false);
 void main() {
 try {
 throw new Exception("This exception...");
 } catch (Exception e) {
 import std.stdio;
 writeln(e.msg, " ... has been caught for you by ", __VENDOR__, ".");
 }
 }
 ---
 This exception... ... has been caught for you by LDC.
 ---

 David

Awesome!!!

Andrei

Feb 11 2013

Iain Buclaw <ibuclaw ubuntu.com> writes:

On 11 February 2013 13:29, Andrei Alexandrescu <
SeeWebsiteForEmail erdani.org> wrote:

 On 2/11/13 8:00 AM, David Nadlinger wrote:

 Looks like we are finally getting somewhere. :)

 ---
 version (Windows) {} else static assert(false);
 void main() {
 try {
 throw new Exception("This exception...");
 } catch (Exception e) {
 import std.stdio;
 writeln(e.msg, " ... has been caught for you by ", __VENDOR__, ".");
 }
 }
 ---
 This exception... ... has been caught for you by LDC.
 ---

 David

 Awesome!!!

 Andrei

This feature has been around for some time (somewhere in lexer.c iirc) -
have you really only just added LDC into it David?

-- 
Iain Buclaw

*(p < e ? p++ : p) = (c & 0x0f) + '0';

Feb 11 2013

"Jesse Phillips" <Jessekphillips+D gmail.com> writes:

On Monday, 11 February 2013 at 13:33:28 UTC, Iain Buclaw wrote:
 This feature has been around for some time (somewhere in 
 lexer.c iirc) -
 have you really only just added LDC into it David?

LLVM hasn't had support for Exceptions in Windows, Hence the 
begining

version (Windows) {} else static assert(false);

Feb 14 2013

=?UTF-8?B?U8O2bmtlIEx1ZHdpZw==?= <sludwig outerproduct.org> writes:

Am 11.02.2013 14:00, schrieb David Nadlinger:
 Looks like we are finally getting somewhere. :)
 
 ---
 version (Windows) {} else static assert(false);
 void main() {
     try {
         throw new Exception("This exception...");
     } catch (Exception e) {
         import std.stdio;
         writeln(e.msg, " ... has been caught for you by ", __VENDOR__, ".");
     }
 }
 ---
 This exception... ... has been caught for you by LDC.
 ---
 
 David

Great news!

How is it implemented? SEH? Did LLVM get any recent additions in this area?

Feb 11 2013

Dmitry Olshansky <dmitry.olsh gmail.com> writes:

11-Feb-2013 17:00, David Nadlinger пишет:
 Looks like we are finally getting somewhere. :)

 ---
 version (Windows) {} else static assert(false);
 void main() {
      try {
          throw new Exception("This exception...");
      } catch (Exception e) {
          import std.stdio;
          writeln(e.msg, " ... has been caught for you by ", __VENDOR__,
 ".");
      }
 }
 ---
 This exception... ... has been caught for you by LDC.
 ---

Nice one!

There got to be size_t.sizeof printed somewhere ... :)


-- 
Dmitry Olshansky

Feb 11 2013

"David Nadlinger" <see klickverbot.at> writes:

On Monday, 11 February 2013 at 17:19:16 UTC, Dmitry Olshansky 
wrote:
 Nice one!

 There got to be size_t.sizeof printed somewhere ... :)

It would be 4; MinGW for Win64 uses SEH, and LLVM still doesn't 
support that (although there is hope).

David

Feb 14 2013

Nick Sabalausky <SeeWebsiteToContactMe semitwist.com> writes:

On Mon, 11 Feb 2013 14:00:17 +0100
"David Nadlinger" <see klickverbot.at> wrote:

 Looks like we are finally getting somewhere. :)
 
 ---
 version (Windows) {} else static assert(false);
 void main() {
      try {
      	throw new Exception("This exception...");
      } catch (Exception e) {
      	import std.stdio;
      	writeln(e.msg, " ... has been caught for you by ", 
 __VENDOR__, ".");
      }
 }
 ---
 This exception... ... has been caught for you by LDC.
 ---
 
 David

I took my mind a minute to process the point of that ("Is
__VENDOR__ a new thing? Is it something about nested exceptions? What's
with the version(Windows)...?")

...and then my thoughts jumped straight to: Holy shit!!

:)

Fantastic news!

Feb 11 2013

"David Nadlinger" <see klickverbot.at> writes:

On Monday, 11 February 2013 at 21:50:35 UTC, Nick Sabalausky 
wrote:
 ...and then my thoughts jumped straight to: Holy shit!!

 :)

 Fantastic news!

Don't get too excited. ;) There is still a lot to do, several 
test cases/unit tests still fail (mainly FP precision and ABI 
related).

But, yes, now there is a real chance of Windows being a 
"first-class target" for LDC in the near future.

I would very much appreciate any help on crushing the remaining 
bugs (it's not rocket science), see 
http://wiki.dlang.org/Building_LDC_on_MinGW_x86 for build 
instructions.

David

Feb 14 2013

"bearophile" <bearophileHUGS lycos.com> writes:

David Nadlinger:

 I would very much appreciate any help on crushing the remaining 
 bugs (it's not rocket science), see 
 http://wiki.dlang.org/Building_LDC_on_MinGW_x86 for build 
 instructions.

Why don't you create one zip or rar, that a user unpacks, adds a 
path, and tries the compiler?

Bye,
bearophile

Feb 14 2013

"David Nadlinger" <see klickverbot.at> writes:

On Thursday, 14 February 2013 at 23:19:38 UTC, bearophile wrote:
 David Nadlinger:

 I would very much appreciate any help on crushing the 
 remaining bugs (it's not rocket science), see 
 http://wiki.dlang.org/Building_LDC_on_MinGW_x86 for build 
 instructions.

 Why don't you create one zip or rar, that a user unpacks, adds 
 a path, and tries the compiler?

Because it's not ready for general consumption yet. For binary 
packages for the other platforms, see the last release 
announcement: 
http://forum.dlang.org/thread/yxklohfbaltbcdnkpnqw forum.dlang.org.

David

Feb 14 2013

"bearophile" <bearophileHUGS lycos.com> writes:

David Nadlinger:

 Because it's not ready for general consumption yet.

If trying the compiler is so complex/hard/time consuming as it 
looks, I will probably not try it. And maybe other people will do 
the same. So it will take even more time to find its bugs.

Bye,
bearophile

Feb 14 2013

"David Nadlinger" <see klickverbot.at> writes:

On Friday, 15 February 2013 at 00:02:00 UTC, bearophile wrote:
 David Nadlinger:

 Because it's not ready for general consumption yet.

 If trying the compiler is so complex/hard/time consuming as it 
 looks, I will probably not try it. And maybe other people will 
 do the same. So it will take even more time to find its bugs.

If you don't want to contribute to LDC development, you shouldn't 
try it on MinGW yet. Simple as that.

When the test suite has turned green or near-green, we will of 
course release binary packages.

David

Feb 14 2013

Andrej Mitrovic <andrej.mitrovich gmail.com> writes:

On 2/14/13, David Nadlinger <see klickverbot.at> wrote:
 I would very much appreciate any help on crushing the remaining
 bugs (it's not rocket science), see
 http://wiki.dlang.org/Building_LDC_on_MinGW_x86 for build
 instructions.

I suppose I'll give this a try tonight. However these steps should
really be part of a nice and simple D script that does this
automatically. Only then can we be sure the instructions are up to
date, + it will save time for potential contributors.

If I get a working build done I'll try writing such a script. I'm
eager to escape the wrath of Optlink on win32. :p

Feb 14 2013

"jerro" <a a.com> writes:

 http://wiki.dlang.org/Building_LDC_on_MinGW_x86 for build 
 instructions.

It seems there is a bug in the MinGW TLS support patch linked 
from that page. This line:

+        DAG.getExternalSymbol("_tls_array"), getPointerTy());

should be

+        DAG.getExternalSymbol("_tls_array", getPointerTy()));

Feb 14 2013

"David Nadlinger" <see klickverbot.at> writes:

On Friday, 15 February 2013 at 02:41:47 UTC, jerro wrote:
 http://wiki.dlang.org/Building_LDC_on_MinGW_x86 for build 
 instructions.

 It seems there is a bug in the MinGW TLS support patch linked 
 from that page. This line:

 +        DAG.getExternalSymbol("_tls_array"), getPointerTy());

 should be

 +        DAG.getExternalSymbol("_tls_array", getPointerTy()));

Yes, always those last-minute refactorings…

I fixed the Gist an hour or so ago.

David

Feb 14 2013

D Programming

C/C++ Programming

Other

digitalmars.D - Exceptional News