www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - future of std.process?

reply Lutger Blijdestijn <lutger.blijdestijn gmail.com> writes:
Some time ago a new std.process branch was made, which included support for 
pipes. Is there still a plan to integrate this in phobos? Does it depend on 
a decision regarding the io design?
Dec 05 2010
parent reply "Lars T. Kyllingstad" <public kyllingen.NOSPAMnet> writes:
On Sun, 05 Dec 2010 15:51:18 +0100, Lutger Blijdestijn wrote:

 Some time ago a new std.process branch was made, which included support
 for pipes. Is there still a plan to integrate this in phobos? Does it
 depend on a decision regarding the io design?

That is still the plan, yes. The new std.process is pretty much done, and has been for a while, but its incorporation in Phobos is being blocked by bug 3979. (The bug was fixed a while ago, but the changes were almost immediately reverted by another bug fix...) -Lars
Dec 06 2010
next sibling parent reply Lutger Blijdestijn <lutger.blijdestijn gmail.com> writes:
Lars T. Kyllingstad wrote:

 On Sun, 05 Dec 2010 15:51:18 +0100, Lutger Blijdestijn wrote:
 
 Some time ago a new std.process branch was made, which included support
 for pipes. Is there still a plan to integrate this in phobos? Does it
 depend on a decision regarding the io design?

That is still the plan, yes. The new std.process is pretty much done, and has been for a while, but its incorporation in Phobos is being blocked by bug 3979. (The bug was fixed a while ago, but the changes were almost immediately reverted by another bug fix...) -Lars

Thanks. I've noticed your personal copy at github, is it useable in the meantime? It doesn't suffer from the same issue?
Dec 06 2010
parent Lutger Blijdestijn <lutger.blijdestijn gmail.com> writes:
Lars T. Kyllingstad wrote:

 On Tue, 07 Dec 2010 08:49:07 -0500, Steven Schveighoffer wrote:
 
 On Tue, 07 Dec 2010 03:47:53 -0500, Lars T. Kyllingstad
 <public kyllingen.nospamnet> wrote:
 
 On Mon, 06 Dec 2010 19:10:23 +0100, Lutger Blijdestijn wrote:

 Lars T. Kyllingstad wrote:

 On Sun, 05 Dec 2010 15:51:18 +0100, Lutger Blijdestijn wrote:

 Some time ago a new std.process branch was made, which included
 support for pipes. Is there still a plan to integrate this in
 phobos? Does it depend on a decision regarding the io design?

That is still the plan, yes. The new std.process is pretty much done, and has been for a while, but its incorporation in Phobos is being blocked by bug 3979. (The bug was fixed a while ago, but the changes were almost immediately reverted by another bug fix...) -Lars

Thanks. I've noticed your personal copy at github, is it useable in the meantime? It doesn't suffer from the same issue?

Yes, it works (and I just uploaded some minor changes that I had in my local repo). Bug 3979 only sets in once you try to name the module "std.process" and compile it together with the rest of Phobos. Note that the code in my github repo is for POSIX only. Steven Schveighoffer has done the Windows work, and I don't have his code.

That reminds me, I should make sure that doesn't get lost, it's not checked in anywhere... Maybe I should send you my code.

Sure, feel free to do so. :) I'm very curious to see how you solved the pipe stuff! Even though we can't include it in Phobos before 3979 is fixed, we can at least combine our code, publish it somewhere, and start the review process. Also, we should probably get the whole File buffering thing sorted out. That discussion kinda ebbed out without any good solution presenting itself... -Lars

Will you announce it if published? I'm interested in using it even though api is unstable, at least I have something until std.process is finished. I'll report back feedback / issues if you want.
Dec 07 2010
prev sibling next sibling parent "Lars T. Kyllingstad" <public kyllingen.NOSPAMnet> writes:
On Mon, 06 Dec 2010 19:10:23 +0100, Lutger Blijdestijn wrote:

 Lars T. Kyllingstad wrote:
 
 On Sun, 05 Dec 2010 15:51:18 +0100, Lutger Blijdestijn wrote:
 
 Some time ago a new std.process branch was made, which included
 support for pipes. Is there still a plan to integrate this in phobos?
 Does it depend on a decision regarding the io design?

That is still the plan, yes. The new std.process is pretty much done, and has been for a while, but its incorporation in Phobos is being blocked by bug 3979. (The bug was fixed a while ago, but the changes were almost immediately reverted by another bug fix...) -Lars

Thanks. I've noticed your personal copy at github, is it useable in the meantime? It doesn't suffer from the same issue?

Yes, it works (and I just uploaded some minor changes that I had in my local repo). Bug 3979 only sets in once you try to name the module "std.process" and compile it together with the rest of Phobos. Note that the code in my github repo is for POSIX only. Steven Schveighoffer has done the Windows work, and I don't have his code. Also be aware that there hasn't been any formal review of the code yet, so some design details may change. (The new UnbufferedFile type, for instance, didn't sit well with everyone.) So don't expect to be able to just s/ltk.process/std.process/ when it eventually gets into Phobos. ;) -Lars
Dec 07 2010
prev sibling next sibling parent "Steven Schveighoffer" <schveiguy yahoo.com> writes:
On Tue, 07 Dec 2010 03:47:53 -0500, Lars T. Kyllingstad  
<public kyllingen.nospamnet> wrote:

 On Mon, 06 Dec 2010 19:10:23 +0100, Lutger Blijdestijn wrote:

 Lars T. Kyllingstad wrote:

 On Sun, 05 Dec 2010 15:51:18 +0100, Lutger Blijdestijn wrote:

 Some time ago a new std.process branch was made, which included
 support for pipes. Is there still a plan to integrate this in phobos?
 Does it depend on a decision regarding the io design?

That is still the plan, yes. The new std.process is pretty much done, and has been for a while, but its incorporation in Phobos is being blocked by bug 3979. (The bug was fixed a while ago, but the changes were almost immediately reverted by another bug fix...) -Lars

Thanks. I've noticed your personal copy at github, is it useable in the meantime? It doesn't suffer from the same issue?

Yes, it works (and I just uploaded some minor changes that I had in my local repo). Bug 3979 only sets in once you try to name the module "std.process" and compile it together with the rest of Phobos. Note that the code in my github repo is for POSIX only. Steven Schveighoffer has done the Windows work, and I don't have his code.

That reminds me, I should make sure that doesn't get lost, it's not checked in anywhere... Maybe I should send you my code. -Steve
Dec 07 2010
prev sibling next sibling parent "Lars T. Kyllingstad" <public kyllingen.NOSPAMnet> writes:
On Tue, 07 Dec 2010 08:49:07 -0500, Steven Schveighoffer wrote:

 On Tue, 07 Dec 2010 03:47:53 -0500, Lars T. Kyllingstad
 <public kyllingen.nospamnet> wrote:
 
 On Mon, 06 Dec 2010 19:10:23 +0100, Lutger Blijdestijn wrote:

 Lars T. Kyllingstad wrote:

 On Sun, 05 Dec 2010 15:51:18 +0100, Lutger Blijdestijn wrote:

 Some time ago a new std.process branch was made, which included
 support for pipes. Is there still a plan to integrate this in
 phobos? Does it depend on a decision regarding the io design?

That is still the plan, yes. The new std.process is pretty much done, and has been for a while, but its incorporation in Phobos is being blocked by bug 3979. (The bug was fixed a while ago, but the changes were almost immediately reverted by another bug fix...) -Lars

Thanks. I've noticed your personal copy at github, is it useable in the meantime? It doesn't suffer from the same issue?

Yes, it works (and I just uploaded some minor changes that I had in my local repo). Bug 3979 only sets in once you try to name the module "std.process" and compile it together with the rest of Phobos. Note that the code in my github repo is for POSIX only. Steven Schveighoffer has done the Windows work, and I don't have his code.

That reminds me, I should make sure that doesn't get lost, it's not checked in anywhere... Maybe I should send you my code.

Sure, feel free to do so. :) I'm very curious to see how you solved the pipe stuff! Even though we can't include it in Phobos before 3979 is fixed, we can at least combine our code, publish it somewhere, and start the review process. Also, we should probably get the whole File buffering thing sorted out. That discussion kinda ebbed out without any good solution presenting itself... -Lars
Dec 07 2010
prev sibling parent "Lars T. Kyllingstad" <public kyllingen.NOSPAMnet> writes:
On Tue, 07 Dec 2010 15:51:09 +0100, Lutger Blijdestijn wrote:

 Lars T. Kyllingstad wrote:
 
 On Tue, 07 Dec 2010 08:49:07 -0500, Steven Schveighoffer wrote:
 
 On Tue, 07 Dec 2010 03:47:53 -0500, Lars T. Kyllingstad
 <public kyllingen.nospamnet> wrote:
 
 On Mon, 06 Dec 2010 19:10:23 +0100, Lutger Blijdestijn wrote:

 Lars T. Kyllingstad wrote:

 On Sun, 05 Dec 2010 15:51:18 +0100, Lutger Blijdestijn wrote:

 Some time ago a new std.process branch was made, which included
 support for pipes. Is there still a plan to integrate this in
 phobos? Does it depend on a decision regarding the io design?

That is still the plan, yes. The new std.process is pretty much done, and has been for a while, but its incorporation in Phobos is being blocked by bug 3979. (The bug was fixed a while ago, but the changes were almost immediately reverted by another bug fix...) -Lars

Thanks. I've noticed your personal copy at github, is it useable in the meantime? It doesn't suffer from the same issue?

Yes, it works (and I just uploaded some minor changes that I had in my local repo). Bug 3979 only sets in once you try to name the module "std.process" and compile it together with the rest of Phobos. Note that the code in my github repo is for POSIX only. Steven Schveighoffer has done the Windows work, and I don't have his code.

That reminds me, I should make sure that doesn't get lost, it's not checked in anywhere... Maybe I should send you my code.

Sure, feel free to do so. :) I'm very curious to see how you solved the pipe stuff! Even though we can't include it in Phobos before 3979 is fixed, we can at least combine our code, publish it somewhere, and start the review process. Also, we should probably get the whole File buffering thing sorted out. That discussion kinda ebbed out without any good solution presenting itself... -Lars

Will you announce it if published? I'm interested in using it even though api is unstable, at least I have something until std.process is finished. I'll report back feedback / issues if you want.

Sure, that would be great. :) -Lars
Dec 07 2010