www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - LLVM progress

reply "Flamaros" <flamaros.xavier gmail.com> writes:
I read an interesting news about LLVM :
http://www.phoronix.com/scan.php?page=news_item&px=MTQ1NjI

It's certainly a great things for ldc too.

Is there a hope to see ldc as the main D compiler if it's full 
featured under Windows?
Sep 07 2013
next sibling parent reply Benjamin Thaut <code benjamin-thaut.de> writes:
Am 07.09.2013 15:20, schrieb Flamaros:
 I read an interesting news about LLVM :
 http://www.phoronix.com/scan.php?page=news_item&px=MTQ1NjI

 It's certainly a great things for ldc too.

 Is there a hope to see ldc as the main D compiler if it's full featured
 under Windows?

I doubt that LDC has compile times that are 3-5 times longer then those of DMD (even in debug without optimization). D advertises with short compile times, so unless LDC improves I don't see this happening. And thats only one of the points. Kind Regards Benjamin Thaut
Sep 07 2013
parent Walter Bright <newshound2 digitalmars.com> writes:
On 9/7/2013 9:42 AM, Benjamin Thaut wrote:
 I doubt that LDC has compile times that are 3-5 times longer then those of DMD
 (even in debug without optimization). D advertises with short compile times, so
 unless LDC improves I don't see this happening. And thats only one of the
points.

Having 3 different D compilers with different strengths and weaknesses spurs improvements in all of them. When I was at GoingNative2013, it was pretty obvious to me that the playful and friendly competition between gcc, clang, and vc has improved all three greatly.
Sep 07 2013
prev sibling next sibling parent Russel Winder <russel winder.org.uk> writes:
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Sat, 2013-09-07 at 10:08 -0700, Walter Bright wrote:
[=E2=80=A6]
 Having 3 different D compilers with different strengths and weaknesses sp=

 improvements in all of them. When I was at GoingNative2013, it was pretty=

 obvious to me that the playful and friendly competition between gcc, clan=

 vc has improved all three greatly.

As has been proved in many areas of life, having multiple players in a game validates the game. Having multiple compilers, books, IDEs etc. for D programming is a mark that D is player in the programming languages game. Sadly D is still not competing against C++ in the way it deserves. Of course C++ is now a niche language. The primary "war" just now is native vs. VM, and VM remains in the ascendency. Go and Rust are the "poster children" for native due to their backers. The questions is whether D should position itself in this "war". I say yes. There needs to be more books on D, and use of D in various areas. QtD, GtkD, (wxD?), D drivers for SQL, Mongo, Couch, Redis, Neo, Riak, etc. all need to be high quality and pushed via reports and talks at non-D conferences. Vibe.d is a huge possibility now that Node.js is losing it's "lustre" and Vert.x and Go are getting serious traction. (At least in the small start-ups arena.) D in GCC and D on LLVM are, for me, far more important than DMD, since they provide penetration into the market via the market leaders. D on Linux via GCC and LLVM, D on OX S via LLVM, (and on Windows, I suppose, via any route :-). The issue for me is to stop worrying about internal contemplative reflection on 10 years of D evolution, and get knowledge of the real-world use of D out there and in people's faces. Stop looking inward and start looking outward. This is the trick Go and Rust have picked up, albeit not as well as they could. D is a major player in the GCC and LLVM worlds, let's take this as read and exploit it for the future of high-quality, effective and pleasurable native code development. --=20 Russel. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n= et 41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Sep 07 2013
prev sibling next sibling parent "Flamaros" <flamaros.xavier gmail.com> writes:
On Saturday, 7 September 2013 at 19:35:47 UTC, Russel Winder 
wrote:
 On Sat, 2013-09-07 at 10:08 -0700, Walter Bright wrote:
 […]
 Having 3 different D compilers with different strengths and 
 weaknesses spurs improvements in all of them. When I was at 
 GoingNative2013, it was pretty obvious to me that the playful 
 and friendly competition between gcc, clang, and vc has 
 improved all three greatly.

As has been proved in many areas of life, having multiple players in a game validates the game. Having multiple compilers, books, IDEs etc. for D programming is a mark that D is player in the programming languages game. Sadly D is still not competing against C++ in the way it deserves. Of course C++ is now a niche language. The primary "war" just now is native vs. VM, and VM remains in the ascendency. Go and Rust are the "poster children" for native due to their backers. The questions is whether D should position itself in this "war". I say yes. There needs to be more books on D, and use of D in various areas. QtD, GtkD, (wxD?), D drivers for SQL, Mongo, Couch, Redis, Neo, Riak, etc. all need to be high quality and pushed via reports and talks at non-D conferences. Vibe.d is a huge possibility now that Node.js is losing it's "lustre" and Vert.x and Go are getting serious traction. (At least in the small start-ups arena.) D in GCC and D on LLVM are, for me, far more important than DMD, since they provide penetration into the market via the market leaders. D on Linux via GCC and LLVM, D on OX S via LLVM, (and on Windows, I suppose, via any route :-).

That also my concern, LLVM tends to replace gcc as C/C++ compiler. LLVM promise to simplify languages compatibility, Apple show us how much is important to improve developers productivity. Google think in the same way with the Go. I think the LLVM message is : developers would be more productive if compiler generate better reports, can aggregate more pieces of software and have better tools (IDE, static analyzer, debugger). In this way D and LLVM philosophies seems compatible.
 The issue for me is to stop worrying about internal 
 contemplative
 reflection on 10 years of D evolution, and get knowledge of the
 real-world use of D out there and in people's faces. Stop 
 looking inward
 and start looking outward. This is the trick Go and Rust have 
 picked up,
 albeit not as well as they could.  D is a major player in the 
 GCC and
 LLVM worlds, let's take this as read and exploit it for the 
 future of
 high-quality, effective and pleasurable native code development.

Sep 07 2013
prev sibling next sibling parent "Chris" <wendlec tcd.ie> writes:
Not to mention the importance of ARM support.
Sep 09 2013
prev sibling next sibling parent Manu <turkeyman gmail.com> writes:
--001a11333dc8bdf5d704e5f2d387
Content-Type: text/plain; charset=UTF-8

On 7 September 2013 23:20, Flamaros <flamaros.xavier gmail.com> wrote:

 I read an interesting news about LLVM :
 http://www.phoronix.com/scan.**php?page=news_item&px=MTQ1NjI<http://www.phoronix.com/scan.php?page=news_item&px=MTQ1NjI>

 It's certainly a great things for ldc too.

 Is there a hope to see ldc as the main D compiler if it's full featured
 under Windows?

This sounds like good news. I'm very interested to hear more news from LLDB. It will be a marvelous day when LDC 'just works' on Windows. No worrying about object/library formats, exception handling, runtime libraries, or debuginfo compatibility. --001a11333dc8bdf5d704e5f2d387 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr">On 7 September 2013 23:20, Flamaros <span dir=3D"ltr">&lt;= <a href=3D"mailto:flamaros.xavier gmail.com" target=3D"_blank">flamaros.xav= ier gmail.com</a>&gt;</span> wrote:<br><div class=3D"gmail_extra"><div clas= s=3D"gmail_quote"> <blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p= x #ccc solid;padding-left:1ex">I read an interesting news about LLVM :<br> <a href=3D"http://www.phoronix.com/scan.php?page=3Dnews_item&;px=3DMTQ1N= jI" target=3D"_blank">http://www.phoronix.com/scan.<u></u>php?page=3Dnews_i= tem&amp;px=3DMTQ1NjI</a><br> <br> It&#39;s certainly a great things for ldc too.<br> <br> Is there a hope to see ldc as the main D compiler if it&#39;s full featured= under Windows?<br> </blockquote></div><br></div><div class=3D"gmail_extra">This sounds like go= od news. I&#39;m very interested to hear more news from LLDB.</div><div cla= ss=3D"gmail_extra">It will be a marvelous day when LDC &#39;just works&#39;= on Windows. No worrying about object/library formats, exception handling, = runtime libraries, or debuginfo compatibility.</div> </div> --001a11333dc8bdf5d704e5f2d387--
Sep 09 2013
prev sibling next sibling parent "Flamaros" <flamaros.xavier gmail.com> writes:
On Monday, 9 September 2013 at 10:29:56 UTC, Chris wrote:
 Not to mention the importance of ARM support.

It's an evidence that is a major target today. :-) I work only on ARM compatible application for my day job, that why DQuick is tough to be easy to port those devices.
Sep 09 2013
prev sibling next sibling parent "Chris" <wendlec tcd.ie> writes:
On Monday, 9 September 2013 at 15:56:17 UTC, Flamaros wrote:
 On Monday, 9 September 2013 at 10:29:56 UTC, Chris wrote:
 Not to mention the importance of ARM support.

It's an evidence that is a major target today. :-) I work only on ARM compatible application for my day job, that why DQuick is tough to be easy to port those devices.

We'll need ARM support too for our applications, sooner or later.
Sep 10 2013
prev sibling parent "Flamaros" <flamaros.xavier gmail.com> writes:
On Tuesday, 10 September 2013 at 08:36:21 UTC, Chris wrote:
 On Monday, 9 September 2013 at 15:56:17 UTC, Flamaros wrote:
 On Monday, 9 September 2013 at 10:29:56 UTC, Chris wrote:
 Not to mention the importance of ARM support.

It's an evidence that is a major target today. :-) I work only on ARM compatible application for my day job, that why DQuick is tough to be easy to port those devices.

We'll need ARM support too for our applications, sooner or later.

Are you telling it's a need for D community or for your company? If it's for D community I hope be able to help a little when compilers will be ready.
Sep 10 2013