• ketmar (4/4) Jan 23 2015 i see that 2.066.1 is landed in GDC (heh, that was sudden! had to fix=20
• Johannes Pfau (3/7) Jan 24 2015 I'd say testing stage. I'm still reducing one ARM codegen bug ;-)
• Iain Buclaw via D.gnu (6/13) Jan 24 2015 Agreed. I'd say it's only released when it hits branch. Now it's in
• Stefan Frijters (9/32) Jan 24 2015 I'm very much looking forward to this release: my code depends
• Iain Buclaw via D.gnu (11/37) Jan 24 2015 wrote:
• ketmar (5/20) Jan 24 2015 but it is in branch! ;-) i'm building my compiler from 'gdc-4.9' branch,...
• Iain Buclaw via D.gnu (3/23) Jan 24 2015 No one told me this. :o)
• Johannes Pfau (4/34) Jan 24 2015 Sorry, I already updated the branches as I needed a 4.9.2 backport for
• ketmar (5/7) Jan 24 2015 i think that it was the right move. this way people who are using gcc 4....
• Iain Buclaw via D.gnu (5/39) Jan 24 2015 Nah - I shouldn't have to have my finger dipped in everything. It's
• ketmar (6/11) Jan 24 2015 off: Iain (or Johannes), can you made gcc.builtin primitives `@nogc`,=20
• Johannes Pfau (6/19) Jan 24 2015 That was a simple fix:
• ketmar (10/32) Jan 24 2015 thank you. i made manu of small changes to druntime and phobos, includin...
• Johannes Pfau (7/44) Jan 25 2015 I thought you have ready to use patches. The nogc/nothrow changes are
• ketmar (9/14) Jan 25 2015 mainline core.atomic is already @nogc, but gdc is not. i'm using mainlin...

ketmar <ketmar ketmar.no-ip.org> writes:

i see that 2.066.1 is landed in GDC (heh, that was sudden! had to fix=20
almost all my private patches...). is it still in testing stage, or it's=20
already fully usable? and if it's usable, i believe that it worth=20
announcing in "D.announce" NG.=

Johannes Pfau <nospam example.com> writes:

Am Sat, 24 Jan 2015 06:32:57 +0000 (UTC)
schrieb ketmar <ketmar ketmar.no-ip.org>:

 i see that 2.066.1 is landed in GDC (heh, that was sudden! had to fix 
 almost all my private patches...). is it still in testing stage, or
 it's already fully usable? and if it's usable, i believe that it
 worth announcing in "D.announce" NG.

I'd say testing stage. I'm still reducing one ARM codegen bug ;-)

"Iain Buclaw via D.gnu" <d.gnu puremagic.com> writes:

On 24 Jan 2015 08:40, "Johannes Pfau via D.gnu" <d.gnu puremagic.com> wrote:
 Am Sat, 24 Jan 2015 06:32:57 +0000 (UTC)
 schrieb ketmar <ketmar ketmar.no-ip.org>:

 i see that 2.066.1 is landed in GDC (heh, that was sudden! had to fix
 almost all my private patches...). is it still in testing stage, or
 it's already fully usable? and if it's usable, i believe that it
 worth announcing in "D.announce" NG.

 I'd say testing stage. I'm still reducing one ARM codegen bug ;-)

Agreed.  I'd say it's only released when it hits branch.  Now it's in
master, it allows people to test out their projects with it.

I also need to write up a formal changelog of what's gone in, what hasn't,
and what has been backported from 2.067.

Iain.

"Stefan Frijters" <sfrijters gmail.com> writes:

On Saturday, 24 January 2015 at 08:47:33 UTC, Iain Buclaw via 
D.gnu wrote:
 On 24 Jan 2015 08:40, "Johannes Pfau via D.gnu" 
 <d.gnu puremagic.com> wrote:
 Am Sat, 24 Jan 2015 06:32:57 +0000 (UTC)
 schrieb ketmar <ketmar ketmar.no-ip.org>:

 i see that 2.066.1 is landed in GDC (heh, that was sudden! 
 had to fix
 almost all my private patches...). is it still in testing 
 stage, or
 it's already fully usable? and if it's usable, i believe 
 that it
 worth announcing in "D.announce" NG.

 I'd say testing stage. I'm still reducing one ARM codegen bug 
 ;-)

 Agreed.  I'd say it's only released when it hits branch.  Now 
 it's in
 master, it allows people to test out their projects with it.

 I also need to write up a formal changelog of what's gone in, 
 what hasn't,
 and what has been backported from 2.067.

 Iain.

I'm very much looking forward to this release: my code depends 
heavily on the 2.066 multidimensional array syntax and I'm eager 
to find out if GDC can give me some extra performance.
As for testing the current state of GDC: is there a way to 
cleanly add this testing stage version to my build matrix on 
Travis CI, e.g. is a binary available somewhere?

Stefan

"Iain Buclaw via D.gnu" <d.gnu puremagic.com> writes:

On 24 Jan 2015 10:20, "Stefan Frijters via D.gnu" <d.gnu puremagic.com>
wrote:
 On Saturday, 24 January 2015 at 08:47:33 UTC, Iain Buclaw via D.gnu wrote:
 On 24 Jan 2015 08:40, "Johannes Pfau via D.gnu" <d.gnu puremagic.com>


wrote:
 Am Sat, 24 Jan 2015 06:32:57 +0000 (UTC)
 schrieb ketmar <ketmar ketmar.no-ip.org>:

 i see that 2.066.1 is landed in GDC (heh, that was sudden! > had to




fix
 almost all my private patches...). is it still in testing > stage, or
 it's already fully usable? and if it's usable, i believe > that it
 worth announcing in "D.announce" NG.

 I'd say testing stage. I'm still reducing one ARM codegen bug ;-)


 Agreed.  I'd say it's only released when it hits branch.  Now it's in
 master, it allows people to test out their projects with it.

 I also need to write up a formal changelog of what's gone in, what


hasn't,
 and what has been backported from 2.067.

 Iain.


 I'm very much looking forward to this release: my code depends heavily on

the 2.066 multidimensional array syntax and I'm eager to find out if GDC
can give me some extra performance.
 As for testing the current state of GDC: is there a way to cleanly add

this testing stage version to my build matrix on Travis CI, e.g. is a
binary available somewhere?
 Stefan

I will be updating docker images this weekend.

https://github.com/ibuclaw/docker-gdc

ketmar <ketmar ketmar.no-ip.org> writes:

On Sat, 24 Jan 2015 08:47:25 +0000, Iain Buclaw via D.gnu wrote:

 On 24 Jan 2015 08:40, "Johannes Pfau via D.gnu" <d.gnu puremagic.com>
 wrote:
 Am Sat, 24 Jan 2015 06:32:57 +0000 (UTC)
 schrieb ketmar <ketmar ketmar.no-ip.org>:

 i see that 2.066.1 is landed in GDC (heh, that was sudden! had to fix
 almost all my private patches...). is it still in testing stage, or
 it's already fully usable? and if it's usable, i believe that it
 worth announcing in "D.announce" NG.

 I'd say testing stage. I'm still reducing one ARM codegen bug ;-)

=20
 Agreed.  I'd say it's only released when it hits branch.  Now it's in
 master, it allows people to test out their projects with it.

but it is in branch! ;-) i'm building my compiler from 'gdc-4.9' branch,=20
and it answers "2066L" on `pragma(msg, __VERSION__);`. that was the=20
reason for my question in the first place.

p.s. please, don't revert it! ;-)=

"Iain Buclaw via D.gnu" <d.gnu puremagic.com> writes:

On 24 January 2015 at 11:08, ketmar via D.gnu <d.gnu puremagic.com> wrote:
 On Sat, 24 Jan 2015 08:47:25 +0000, Iain Buclaw via D.gnu wrote:

 On 24 Jan 2015 08:40, "Johannes Pfau via D.gnu" <d.gnu puremagic.com>
 wrote:
 Am Sat, 24 Jan 2015 06:32:57 +0000 (UTC)
 schrieb ketmar <ketmar ketmar.no-ip.org>:

 i see that 2.066.1 is landed in GDC (heh, that was sudden! had to fix
 almost all my private patches...). is it still in testing stage, or
 it's already fully usable? and if it's usable, i believe that it
 worth announcing in "D.announce" NG.

 I'd say testing stage. I'm still reducing one ARM codegen bug ;-)

 Agreed.  I'd say it's only released when it hits branch.  Now it's in
 master, it allows people to test out their projects with it.

 but it is in branch! ;-) i'm building my compiler from 'gdc-4.9' branch,
 and it answers "2066L" on `pragma(msg, __VERSION__);`. that was the
 reason for my question in the first place.

 p.s. please, don't revert it! ;-)


No one told me this.  :o)

Iain.

Johannes Pfau <nospam example.com> writes:

Am Sat, 24 Jan 2015 11:56:31 +0000
schrieb "Iain Buclaw via D.gnu" <d.gnu puremagic.com>:

 On 24 January 2015 at 11:08, ketmar via D.gnu <d.gnu puremagic.com>
 wrote:
 On Sat, 24 Jan 2015 08:47:25 +0000, Iain Buclaw via D.gnu wrote:

 On 24 Jan 2015 08:40, "Johannes Pfau via D.gnu"
 <d.gnu puremagic.com> wrote:
 Am Sat, 24 Jan 2015 06:32:57 +0000 (UTC)
 schrieb ketmar <ketmar ketmar.no-ip.org>:

 i see that 2.066.1 is landed in GDC (heh, that was sudden! had
 to fix almost all my private patches...). is it still in
 testing stage, or it's already fully usable? and if it's
 usable, i believe that it worth announcing in "D.announce" NG.

 I'd say testing stage. I'm still reducing one ARM codegen bug ;-)

 Agreed.  I'd say it's only released when it hits branch.  Now it's
 in master, it allows people to test out their projects with it.

 but it is in branch! ;-) i'm building my compiler from 'gdc-4.9'
 branch, and it answers "2066L" on `pragma(msg, __VERSION__);`. that
 was the reason for my question in the first place.

 p.s. please, don't revert it! ;-)

 
 
 No one told me this.  :o)
 
 Iain.

Sorry, I already updated the branches as I needed a 4.9.2 backport for
testing. I should probably think twice about this next time :-)

ketmar <ketmar ketmar.no-ip.org> writes:

On Sat, 24 Jan 2015 13:33:50 +0100, Johannes Pfau wrote:

 Sorry, I already updated the branches as I needed a 4.9.2 backport for
 testing. I should probably think twice about this next time :-)

i think that it was the right move. this way people who are using gcc 4.9=20
(me ;-) can test it too. not that i'm doing compiler unittests or so, but=20
i'm working on my project with gdc, so there is a small chance to catch=20
some bugs anyway.=

"Iain Buclaw via D.gnu" <d.gnu puremagic.com> writes:

On 24 January 2015 at 12:33, Johannes Pfau via D.gnu
<d.gnu puremagic.com> wrote:
 Am Sat, 24 Jan 2015 11:56:31 +0000
 schrieb "Iain Buclaw via D.gnu" <d.gnu puremagic.com>:

 On 24 January 2015 at 11:08, ketmar via D.gnu <d.gnu puremagic.com>
 wrote:
 On Sat, 24 Jan 2015 08:47:25 +0000, Iain Buclaw via D.gnu wrote:

 On 24 Jan 2015 08:40, "Johannes Pfau via D.gnu"
 <d.gnu puremagic.com> wrote:
 Am Sat, 24 Jan 2015 06:32:57 +0000 (UTC)
 schrieb ketmar <ketmar ketmar.no-ip.org>:

 i see that 2.066.1 is landed in GDC (heh, that was sudden! had
 to fix almost all my private patches...). is it still in
 testing stage, or it's already fully usable? and if it's
 usable, i believe that it worth announcing in "D.announce" NG.

 I'd say testing stage. I'm still reducing one ARM codegen bug ;-)

 Agreed.  I'd say it's only released when it hits branch.  Now it's
 in master, it allows people to test out their projects with it.

 but it is in branch! ;-) i'm building my compiler from 'gdc-4.9'
 branch, and it answers "2066L" on `pragma(msg, __VERSION__);`. that
 was the reason for my question in the first place.

 p.s. please, don't revert it! ;-)


 No one told me this.  :o)

 Iain.

 Sorry, I already updated the branches as I needed a 4.9.2 backport for
 testing. I should probably think twice about this next time :-)

Nah - I shouldn't have to have my finger dipped in everything.  It's
just given me a list of more things to update today.

Iain

ketmar <ketmar ketmar.no-ip.org> writes:

On Sat, 24 Jan 2015 12:48:27 +0000, Iain Buclaw via D.gnu wrote:

 Sorry, I already updated the branches as I needed a 4.9.2 backport for
 testing. I should probably think twice about this next time :-)

=20
 Nah - I shouldn't have to have my finger dipped in everything.  It's
 just given me a list of more things to update today.

off: Iain (or Johannes), can you made gcc.builtin primitives ` nogc`,=20
please? i completely lost in the code that generates 'em, but it blocks=20
making "core.atomic" ` nogc`, and i can't build my project anymore. it's=20
ok for me to manually patch "core.atomic", but sadly, i can't do this=20
with gdc code.=

Johannes Pfau <nospam example.com> writes:

Am Sat, 24 Jan 2015 13:07:43 +0000 (UTC)
schrieb ketmar <ketmar ketmar.no-ip.org>:

 On Sat, 24 Jan 2015 12:48:27 +0000, Iain Buclaw via D.gnu wrote:
 
 Sorry, I already updated the branches as I needed a 4.9.2 backport
 for testing. I should probably think twice about this next time :-)

 
 Nah - I shouldn't have to have my finger dipped in everything.  It's
 just given me a list of more things to update today.

 
 off: Iain (or Johannes), can you made gcc.builtin primitives ` nogc`, 
 please? i completely lost in the code that generates 'em, but it
 blocks making "core.atomic" ` nogc`, and i can't build my project
 anymore. it's ok for me to manually patch "core.atomic", but sadly, i
 can't do this with gdc code.

That was a simple fix:
https://github.com/D-Programming-GDC/GDC/commit/aa83d2ac749baa552f250891378118b704d7c589

If you make changes to core.atomic for  nogc please provide a patch
or file a pull request.

ketmar <ketmar ketmar.no-ip.org> writes:

On Sat, 24 Jan 2015 23:35:55 +0100, Johannes Pfau wrote:

 Am Sat, 24 Jan 2015 13:07:43 +0000 (UTC)
 schrieb ketmar <ketmar ketmar.no-ip.org>:
=20
 On Sat, 24 Jan 2015 12:48:27 +0000, Iain Buclaw via D.gnu wrote:
=20
 Sorry, I already updated the branches as I needed a 4.9.2 backport
 for testing. I should probably think twice about this next time :-)

=20
 Nah - I shouldn't have to have my finger dipped in everything.  It's
 just given me a list of more things to update today.

=20
 off: Iain (or Johannes), can you made gcc.builtin primitives ` nogc`,
 please? i completely lost in the code that generates 'em, but it blocks
 making "core.atomic" ` nogc`, and i can't build my project anymore.
 it's ok for me to manually patch "core.atomic", but sadly, i can't do
 this with gdc code.

=20
 That was a simple fix:
 https://github.com/D-Programming-GDC/GDC/commit/

aa83d2ac749baa552f250891378118b704d7c589
=20
 If you make changes to core.atomic for  nogc please provide a patch or
 file a pull request.

thank you. i made manu of small changes to druntime and phobos, including=20
taking MonoTime from core.time, using std.utf from mainline git head,=20
'cause i need .byChar and so on. i also adding "nothow:  nogc:" here and=20
there in druntime headers. dunno if it all of any real value, though.

i can create some tickets in bugzilla with those patches, so you can pick=20
the ones you are interested in and just close all the others. or i can=20
simply dump 'em all in one ticket (but in different attaches). what will=20
be better?=

Johannes Pfau <nospam example.com> writes:

Am Sat, 24 Jan 2015 23:28:52 +0000 (UTC)
schrieb ketmar <ketmar ketmar.no-ip.org>:

 On Sat, 24 Jan 2015 23:35:55 +0100, Johannes Pfau wrote:
 
 Am Sat, 24 Jan 2015 13:07:43 +0000 (UTC)
 schrieb ketmar <ketmar ketmar.no-ip.org>:
 
 On Sat, 24 Jan 2015 12:48:27 +0000, Iain Buclaw via D.gnu wrote:
 
 Sorry, I already updated the branches as I needed a 4.9.2
 backport for testing. I should probably think twice about this
 next time :-)

 
 Nah - I shouldn't have to have my finger dipped in everything.
 It's just given me a list of more things to update today.

 
 off: Iain (or Johannes), can you made gcc.builtin primitives
 ` nogc`, please? i completely lost in the code that generates 'em,
 but it blocks making "core.atomic" ` nogc`, and i can't build my
 project anymore. it's ok for me to manually patch "core.atomic",
 but sadly, i can't do this with gdc code.

 
 That was a simple fix:
 https://github.com/D-Programming-GDC/GDC/commit/

 aa83d2ac749baa552f250891378118b704d7c589
 
 If you make changes to core.atomic for  nogc please provide a patch
 or file a pull request.

 
 thank you. i made manu of small changes to druntime and phobos,
 including taking MonoTime from core.time, using std.utf from mainline
 git head, 'cause i need .byChar and so on. i also adding "nothow:
  nogc:" here and there in druntime headers. dunno if it all of any
 real value, though.
 
 i can create some tickets in bugzilla with those patches, so you can
 pick the ones you are interested in and just close all the others. or
 i can simply dump 'em all in one ticket (but in different attaches).
 what will be better?

I thought you have ready to use patches. The nogc/nothrow changes are
useful if they're not in mainline druntime yet but I don't really have
the time to refactor such patches. So if you have ready to use patches
against mainline please attach them to bug reports. Otherwise I guess
somebody else will fix it in mainline druntime at some point.

ketmar <ketmar ketmar.no-ip.org> writes:

On Sun, 25 Jan 2015 11:00:04 +0100, Johannes Pfau wrote:

 I thought you have ready to use patches. The nogc/nothrow changes are
 useful if they're not in mainline druntime yet but I don't really have
 the time to refactor such patches. So if you have ready to use patches
 against mainline please attach them to bug reports. Otherwise I guess
 somebody else will fix it in mainline druntime at some point.

mainline core.atomic is already  nogc, but gdc is not. i'm using mainline=20
git head, with everyday updates, so i see alot of features that aren't in=20
"releases" and aren't in gdc. and i'm using some of that features, and=20
then i backporting some of them to gdc (sometimes with some dummy no-op=20
patches). so, for example, non- nogc atomic ops are gdc-specific and=20
absent in mainline. alot of core modules in mainline already has=20
"nothrow:  nogc:", but not gdc core modules. that are the patches i'm=20
applying to gdc -- mostly to up it to mainline.=