digitalmars.D - [D-Programming-Deimos] Review process?
- Johannes Pfau (19/19) Nov 05 2011 Now that deimos is an organization and we can't use pull requests, how
- Dejan Lekic (3/6) Nov 05 2011 Johannes, I did not know this - good news IMHO! :)
- Dejan Lekic (3/3) Nov 05 2011 OK, I believe I understand how it goes there - you basically need to hav...
- Gor Gyolchanyan (2/5) Nov 05 2011
- mta`chrono (3/3) Nov 08 2011 Does anyone know the last state of Deimos? What are we going to use now.
- Gor Gyolchanyan (8/11) Nov 08 2011 The first one is handy because you can have everything you want in a
- Andrea Fontana (8/24) Nov 08 2011 Could the same idea (of a single repository) be extended to support swig
- Gor Gyolchanyan (11/32) Nov 08 2011 Using SWIG for porting C libraries to D is like using a chainsaw for
- Steve Teale (12/16) Nov 08 2011 I still have more fundamental questions. Maybe they've been answered, bu...
- Gor Gyolchanyan (11/29) Nov 08 2011 s. Is that still the boundary?
- Jude Young (12/12) Nov 12 2011 Ok, guys..
- Andrei Alexandrescu (6/18) Nov 12 2011 Not sure when I robbed you of the first one :o).
- Jude Young (9/32) Nov 12 2011 It was meant as a joke, ncurses wasn't quite ready the first time I
- Walter Bright (3/5) Nov 12 2011 I've been working on the openssl stuff, but it ain't ready yet. I'm wond...
- Andrei Alexandrescu (4/9) Nov 12 2011 I think we should do all we can to avoid deimos becoming a
- Jude Young (9/9) Nov 12 2011 No offense taken Andrei.
- Jude Young (2/2) Nov 12 2011 Oops misinterpreted.
- Andrei Alexandrescu (10/18) Nov 12 2011 I think I remember now, thanks. It was something almost complete, almost...
- Walter Bright (2/3) Nov 12 2011 https://github.com/D-Programming-Deimos/ncurses
- Walter Bright (2/4) Nov 12 2011 What are they? Homepage urls?
- Walter Bright (6/15) Nov 12 2011 Great!
- Walter Bright (3/4) Nov 12 2011 https://github.com/D-Programming-Deimos/libmysql
- Walter Bright (3/6) Nov 12 2011 Yes
- JimB (7/15) Nov 12 2011 I can't help but notice your attention given to (i.e., where your mind i...
- Walter Bright (4/12) Nov 12 2011 https://github.com/D-Programming-Deimos/liblzma
- Johannes Pfau (9/22) Nov 13 2011 Thanks. My fork of that repository is here:
- Walter Bright (2/19) Nov 13 2011 Make a subdirectory called C and store the C .h files there.
- David Nadlinger (5/6) Nov 13 2011 What is the point of keeping the C headers in a separate subdirectory?
- Walter Bright (3/9) Nov 13 2011 I haven't looked at it yet.
- Jude Young (17/17) Nov 14 2011 Ok, I think that the ncurses bindings are about ready.
- Walter Bright (7/24) Nov 14 2011 The latter (.d files at the top level).
- Jude Young (13/22) Nov 14 2011 Over 2000 lines of code in the curses.h file.
- Jacob Carlborg (7/21) Nov 14 2011 That doesn't sound like a good idea. I think there should be one top
- Jude Young (4/27) Nov 15 2011 conflicts how?
- Jacob Carlborg (6/14) Nov 13 2011 I have bindings for Ruby and Clang ready. I'm going to wait with the
- maarten van damme (1/1) Nov 13 2011 will the bindings project from dsource be added?
Now that deimos is an organization and we can't use pull requests, how do we get new bindings into deimos? I'd still like to get my liblzma bindings in there: https://github.com/jpf91/deimos/tree/liblzma And I still wonder whether we are allowed to bundle simple examples with the bindings: https://github.com/jpf91/deimos/tree/examples https://github.com/jpf91/deimos/blob/examples/examples/liblzma/xz_pipe_comp.d Also, "The Deimos Manifesto" should be moved to the deimos organization. I guess organizations can't have wikis, so we probably have to add a 'fake' project for that? (Or we could keep a wiki at https://github.com/D-Programming-Language/deimos and link to the projects in https://github.com/D-Programming-Deimos ?) And the deimos project (https://github.com/D-Programming-Language/deimos) should probably display a note that it has moved to https://github.com/D-Programming-Deimos -- Johannes Pfau
Nov 05 2011
On Sat, 05 Nov 2011 10:28:25 +0100, Johannes Pfau wrote:Now that deimos is an organization and we can't use pull requests, how do we get new bindings into deimos?Johannes, I did not know this - good news IMHO! :) I wonder the same question as I have some stuff to contribute as well.
Nov 05 2011
OK, I believe I understand how it goes there - you basically need to have a project on GitHub, and ask Deimos (organisation) admin to add your project to the list of Deimos projects. Am I correct?
Nov 05 2011
I'd gladly start making a libjpeg port :-) On Sat, Nov 5, 2011 at 4:00 PM, Dejan Lekic <dejan.lekic gmail.com> wrote:OK, I believe I understand how it goes there - you basically need to have a project on GitHub, and ask Deimos (organisation) admin to add your project to the list of Deimos projects. Am I correct?
Nov 05 2011
Does anyone know the last state of Deimos? What are we going to use now. This: https://github.com/D-Programming-Language/deimos Or that: https://github.com/D-Programming-Deimos/
Nov 08 2011
The first one is handy because you can have everything you want in a single place and don't need to clone every lib separately. The second one is handy because you don't have to carry around tons of unneeded code. For that reason, I propose us to keep the code in separate repositories and refer to them from a single repository in the form of submodules. On Tue, Nov 8, 2011 at 2:00 PM, mta`chrono <chrono mta-international.net> wrote:Does anyone know the last state of Deimos? What are we going to use now. This: https://github.com/D-Programming-Language/deimos Or that: https://github.com/D-Programming-Deimos/
Nov 08 2011
Could the same idea (of a single repository) be extended to support swig based portings?=20 I've tried to port one of my project (quite complex data access layer) to D using swig, and it works fine (and it's quite easy to do). It would be useful for c++ class-based libraries... Il giorno mar, 08/11/2011 alle 14.08 +0400, Gor Gyolchanyan ha scritto:The first one is handy because you can have everything you want in a single place and don't need to clone every lib separately. The second one is handy because you don't have to carry around tons of unneeded code. =20 For that reason, I propose us to keep the code in separate repositories and refer to them from a single repository in the form of submodules. =20 On Tue, Nov 8, 2011 at 2:00 PM, mta`chrono <chrono mta-international.net>=wrote:.Does anyone know the last state of Deimos? What are we going to use now=This: https://github.com/D-Programming-Language/deimos Or that: https://github.com/D-Programming-Deimos/
Nov 08 2011
Using SWIG for porting C libraries to D is like using a chainsaw for vascular incisions: It's extremely heavy and unproductive. This is all because SWIG generates intermediate binary, which forwards calls to C functions, which is completely unnecessary. Furthermore, It translates macro constants into lists of const variables. This is very unproductive as well, because macro definitions don't have addresses and const variables do. A human translator would've put those in a single enum (providing a whole bunch of extra functionality, like obtaining the string of the enum symbols). On Tue, Nov 8, 2011 at 2:47 PM, Andrea Fontana <advmail katamail.com> wrote:Could the same idea (of a single repository) be extended to support swig based portings? I've tried to port one of my project (quite complex data access layer) to D using swig, and it works fine (and it's quite easy to do). It would be useful for c++ class-based libraries... Il giorno mar, 08/11/2011 alle 14.08 +0400, Gor Gyolchanyan ha scritto: The first one is handy because you can have everything you want in a single place and don't need to clone every lib separately. The second one is handy because you don't have to carry around tons of unneeded code. For that reason, I propose us to keep the code in separate repositories and refer to them from a single repository in the form of submodules. On Tue, Nov 8, 2011 at 2:00 PM, mta`chrono <chrono mta-international.netwrote: Does anyone know the last state of Deimos? What are we going to use now. This: https://github.com/D-Programming-Language/deimos Or that: https://github.com/D-Programming-Deimos/
Nov 08 2011
On Tue, 08 Nov 2011 11:00:30 +0100, mta`chrono wrote:Does anyone know the last state of Deimos? What are we going to use now. This: https://github.com/D-Programming-Language/deimos Or that: https://github.com/D-Programming-Deimos/I still have more fundamental questions. Maybe they've been answered, but if so I missed it. Walter's original CAPI Manifesto wanted straight translations of C headers. Is that still the boundary? I have done mysql.d and that would be close to fitting in there, but I also did mysqld.d, and that is way off. It's not Phobos material because of license considerations, but it's considerably beyond being a translation of a C header file. Where would that go? There are other pretty basic questions - who is the Deimos admin? How do you go about getting things in there? Steve
Nov 08 2011
Walter's original CAPI Manifesto wanted straight translations of C header=s. Is that still the boundary? Yes. The entire idea of Deimos is to provide an interface to original C libraries with absolutely minimum amount of modifications.I have done mysql.d and that would be close to fitting in there, but I al=so did mysqld.d, and that is way off. It's not Phobos material because of l= icense considerations, but it's considerably beyond being a translation of = a C header file. Where would that go? Those kind of wrappers are very useful in general. I think it would be a good idea to also have a place to keep adapted and safe versions of the Deimos. On Tue, Nov 8, 2011 at 5:47 PM, Steve Teale <steve.teale britseyeview.com> wrote:On Tue, 08 Nov 2011 11:00:30 +0100, mta`chrono wrote:Does anyone know the last state of Deimos? What are we going to use now. This: https://github.com/D-Programming-Language/deimos Or that: https://github.com/D-Programming-Deimos/I still have more fundamental questions. Maybe they've been answered, but if so I missed it. Walter's original CAPI Manifesto wanted straight translations of C headers. Is that still the boundary? I have done mysql.d and that would be close to fitting in there, but I also did mysqld.d, and that is way off. It's not Phobos material because of license considerations, but it's considerably beyond being a translation of a C header file. Where would that go? There are other pretty basic questions - who is the Deimos admin? How do you go about getting things in there? Steve
Nov 08 2011
Ok, guys.. does anyone know the answers to these questions? I have ncurses, 0MQ(ZeroMQ? ZMQ?), and CZMQ(high-level C wrapper for 0MQ). (They aren't 100% mine, but I helped! =D) They all should work, I don't have any problems with them. (haven't tested CZMQ yet...) I haven't decided what my next project is, but I LIKE working on these. And I think it's a shame that they are just sittin around, collecting dust... Will someone please tell me what needs to be done to get these into Deimos? Come on Andrei, give me a second chance. ; )
Nov 12 2011
On 11/12/11 7:34 PM, Jude Young wrote:Ok, guys.. does anyone know the answers to these questions? I have ncurses, 0MQ(ZeroMQ? ZMQ?), and CZMQ(high-level C wrapper for 0MQ). (They aren't 100% mine, but I helped! =D) They all should work, I don't have any problems with them. (haven't tested CZMQ yet...) I haven't decided what my next project is, but I LIKE working on these. And I think it's a shame that they are just sittin around, collecting dust... Will someone please tell me what needs to be done to get these into Deimos? Come on Andrei, give me a second chance. ; )Not sure when I robbed you of the first one :o). Deimos is a very young initiative. We don't have a set process for it, and it needs critical mass to succeed. At this point I personally think we should have a low barrier of acceptance. Andrei
Nov 12 2011
On Sat 12 Nov 2011 08:07:55 PM CST, Andrei Alexandrescu wrote:On 11/12/11 7:34 PM, Jude Young wrote:It was meant as a joke, ncurses wasn't quite ready the first time I tried to get it in. Apparently the low barrier wasn't low enough the first time around. haha. Well, there may not be a set process, but I do see a completely empty repo labeled openssl. I will guarantee that mine are further along than that! =D Just give me a heads up when y'all are ready to start adding stuff.Ok, guys.. does anyone know the answers to these questions? I have ncurses, 0MQ(ZeroMQ? ZMQ?), and CZMQ(high-level C wrapper for 0MQ). (They aren't 100% mine, but I helped! =D) They all should work, I don't have any problems with them. (haven't tested CZMQ yet...) I haven't decided what my next project is, but I LIKE working on these. And I think it's a shame that they are just sittin around, collecting dust... Will someone please tell me what needs to be done to get these into Deimos? Come on Andrei, give me a second chance. ; )Not sure when I robbed you of the first one :o). Deimos is a very young initiative. We don't have a set process for it, and it needs critical mass to succeed. At this point I personally think we should have a low barrier of acceptance. Andrei
Nov 12 2011
On 11/12/2011 6:28 PM, Jude Young wrote:Well, there may not be a set process, but I do see a completely empty repo labeled openssl.I've been working on the openssl stuff, but it ain't ready yet. I'm wondering if a work-in-progress would be worthwhile to put in it.
Nov 12 2011
On 11/12/11 8:46 PM, Walter Bright wrote:On 11/12/2011 6:28 PM, Jude Young wrote:I think we should do all we can to avoid deimos becoming a stuff-that-almost-works graveyard. AndreiWell, there may not be a set process, but I do see a completely empty repo labeled openssl.I've been working on the openssl stuff, but it ain't ready yet. I'm wondering if a work-in-progress would be worthwhile to put in it.
Nov 12 2011
No offense taken Andrei. I completely understand where you were coming from. It was presumptuous of me to attempt to get it in that quickly. Walter, zeromq is here: https://github.com/1100110/zeromq It should be entirely up-to-date and correct. I haven't encountered any bugs yet, and I'm currently trying to port the C examples in the ZMQ tutorial to D. CZMQ compiles cleanly, but I don't trust it enough yet...
Nov 12 2011
On 11/12/11 8:28 PM, Jude Young wrote:It was meant as a joke, ncurses wasn't quite ready the first time I tried to get it in. Apparently the low barrier wasn't low enough the first time around. haha. Well, there may not be a set process, but I do see a completely empty repo labeled openssl. I will guarantee that mine are further along than that! =D Just give me a heads up when y'all are ready to start adding stuff.I think I remember now, thanks. It was something almost complete, almost done, and almost working. It was also coming from a very recent contributor. No offense, but we've had terrible experience with people who came out of nowhere, made "almost" contributions, and disappeared. I think we need a minimal assurance that the stuff is either 100% finished and deployable, or some clear track record of the submitter indicating that an incomplete first commit will be followed up. Thanks, Andrei
Nov 12 2011
On 11/12/2011 5:34 PM, Jude Young wrote:I have ncurses,https://github.com/D-Programming-Deimos/ncurses
Nov 12 2011
On 11/12/2011 5:34 PM, Jude Young wrote:0MQ(ZeroMQ? ZMQ?), and CZMQ(high-level C wrapper for 0MQ).What are they? Homepage urls?
Nov 12 2011
On 11/8/2011 5:47 AM, Steve Teale wrote:Walter's original CAPI Manifesto wanted straight translations of C headers. Is that still the boundary?Yes.I have done mysql.d and that would be close to fitting in there,Great!but I also did mysqld.d, and that is way off. It's not Phobos material because of license considerations, but it's considerably beyond being a translation of a C header file. Where would that go?At the moment, there isn't a central place for it yet. For the time being, I'd recommend setting up your own github account for it.There are other pretty basic questions - who is the Deimos admin? How do you go about getting things in there?The admins who can insert a new project are myself, Andrei, and Brad Roberts.
Nov 12 2011
On 11/8/2011 5:47 AM, Steve Teale wrote:I have done mysql.d and that would be close to fitting in there,https://github.com/D-Programming-Deimos/libmysql Full speed ahead!
Nov 12 2011
On 11/8/2011 2:00 AM, mta`chrono wrote:Does anyone know the last state of Deimos? What are we going to use now. This: https://github.com/D-Programming-Language/deimosNoOr that: https://github.com/D-Programming-Deimos/Yes
Nov 12 2011
"Walter Bright" <newshound2 digitalmars.com> wrote in message news:j9n97j$t4a$1 digitalmars.com...On 11/8/2011 2:00 AM, mta`chrono wrote:I can't help but notice your attention given to (i.e., where your mind is at these days?) to websites and such "technology". Is that what old compiler-writers do in retirement and where they go... and fade away...? I have a parallel thought on that (i.e., this scenario, be it "true" or not, reminds me of another scenario).Does anyone know the last state of Deimos? What are we going to use now. This: https://github.com/D-Programming-Language/deimosNoOr that: https://github.com/D-Programming-Deimos/Yes
Nov 12 2011
On 11/5/2011 2:28 AM, Johannes Pfau wrote:Now that deimos is an organization and we can't use pull requests, how do we get new bindings into deimos? I'd still like to get my liblzma bindings in there: https://github.com/jpf91/deimos/tree/liblzmahttps://github.com/D-Programming-Deimos/liblzma Next step: create a pull request for it.And I still wonder whether we are allowed to bundle simple examples with the bindings: https://github.com/jpf91/deimos/tree/examples https://github.com/jpf91/deimos/blob/examples/examples/liblzma/xz_pipe_comp.dI suppose if you make a subdirectory for them, that'd be fine.
Nov 12 2011
Walter Bright wrote:On 11/5/2011 2:28 AM, Johannes Pfau wrote:Thanks. My fork of that repository is here: https://github.com/jpf91/liblzma/ But I decided to keep the C headers in a special branch and github only allows to add one branch per pull request? Should I file 6 pull requests (2 branches + 4 tags) or is there a simpler way to merge this?Now that deimos is an organization and we can't use pull requests, how do we get new bindings into deimos? I'd still like to get my liblzma bindings in there: https://github.com/jpf91/deimos/tree/liblzmahttps://github.com/D-Programming-Deimos/liblzma Next step: create a pull request for it.-- Johannes PfauAnd I still wonder whether we are allowed to bundle simple examples with the bindings: https://github.com/jpf91/deimos/tree/examples https://github.com/jpf91/deimos/blob/examples/examples/liblzma/xz_pipe_comp.dI suppose if you make a subdirectory for them, that'd be fine.
Nov 13 2011
On 11/13/2011 1:56 AM, Johannes Pfau wrote:Walter Bright wrote:Make a subdirectory called C and store the C .h files there.On 11/5/2011 2:28 AM, Johannes Pfau wrote:Thanks. My fork of that repository is here: https://github.com/jpf91/liblzma/ But I decided to keep the C headers in a special branch and github only allows to add one branch per pull request? Should I file 6 pull requests (2 branches + 4 tags) or is there a simpler way to merge this?Now that deimos is an organization and we can't use pull requests, how do we get new bindings into deimos? I'd still like to get my liblzma bindings in there: https://github.com/jpf91/deimos/tree/liblzmahttps://github.com/D-Programming-Deimos/liblzma Next step: create a pull request for it.
Nov 13 2011
On 11/13/11 11:13 AM, Walter Bright wrote:Make a subdirectory called C and store the C .h files there.What is the point of keeping the C headers in a separate subdirectory? Johannes' automatic merging suggestion on d.D.announce seemed much more elegant to me, what's your opinion on that? David
Nov 13 2011
On 11/13/2011 1:30 PM, David Nadlinger wrote:On 11/13/11 11:13 AM, Walter Bright wrote:So the user can compare the D ones with the correct C ones.Make a subdirectory called C and store the C .h files there.What is the point of keeping the C headers in a separate subdirectory?Johannes' automatic merging suggestion on d.D.announce seemed much more elegant to me, what's your opinion on that?I haven't looked at it yet.
Nov 13 2011
Ok, I think that the ncurses bindings are about ready. Just something to remember: ncurses is HUGE. There are probably hidden bugs somewhere. So there are a few questions that I'd like answered before I send a pull request. 1. layout of the files. I see that lzma is currently using C folder to house the C.h and deimos to house the .d files. Is this the standard that all of the repos should follow? I have simply put the .d files in the top level. Which would be the preferred method? 2. What will be the criteria for deciding what gets into Deimos? I understand not wanting to allow cruft and unsupportable bindings into Deimos, but at the same time, some of the projects would benefit from bug reports. 3. Can we get a set standard way to request that a new repo be added? I have ZeroMQ ready, but I'm a little unsure who to contact about adding it. I don't really want to bother anyone...
Nov 14 2011
On 11/14/2011 5:54 AM, Jude Young wrote:Ok, I think that the ncurses bindings are about ready. Just something to remember: ncurses is HUGE.The header files, too?There are probably hidden bugs somewhere. So there are a few questions that I'd like answered before I send a pull request. 1. layout of the files. I see that lzma is currently using C folder to house the C.h and deimos to house the .d files. Is this the standard that all of the repos should follow? I have simply put the .d files in the top level. Which would be the preferred method?The latter (.d files at the top level).2. What will be the criteria for deciding what gets into Deimos? I understand not wanting to allow cruft and unsupportable bindings into Deimos, but at the same time, some of the projects would benefit from bug reports.Deimos is for interfaces to publicly available C libraries. For inclusion to Deimos, those libraries ought to be notable, high quality, and reasonably well known, i.e. you can do an "apt get" for them on Ubuntu.3. Can we get a set standard way to request that a new repo be added? I have ZeroMQ ready, but I'm a little unsure who to contact about adding it. I don't really want to bother anyone...Just letting me, Andrei, or Brad know will do for now.
Nov 14 2011
On Mon 14 Nov 2011 11:06:20 PM CST, Walter Bright wrote:On 11/14/2011 5:54 AM, Jude Young wrote:Over 2000 lines of code in the curses.h file. Which wouldn't be a problem except quite a bit of it is preprocessor directives. When a large percentage of the functions are defined macros it can get very tedious very quickly. I'm not about to count the number of function declarations, but it looks to be over 300... just going by line numbers. I'd count that as huge, but maybe I simply don't have enough experience yet. hahah.Ok, I think that the ncurses bindings are about ready. Just something to remember: ncurses is HUGE.The header files, too?The latter (.d files at the top level).Great, cause that's what I did. =PDeimos is for interfaces to publicly available C libraries. For inclusion to Deimos, those libraries ought to be notable, high quality, and reasonably well known, i.e. you can do an "apt get" for them on Ubuntu.I'll keep that in mind when looking for my next project. Ok, thank you for the time.
Nov 14 2011
On 2011-11-15 06:06, Walter Bright wrote:On 11/14/2011 5:54 AM, Jude Young wrote:That doesn't sound like a good idea. I think there should be one top level package that has the same name as the library and but every D module in that directory, keeping any directory structure that the C headers have. Otherwise there will be a lot of conflicts. -- /Jacob CarlborgOk, I think that the ncurses bindings are about ready. Just something to remember: ncurses is HUGE.The header files, too?There are probably hidden bugs somewhere. So there are a few questions that I'd like answered before I send a pull request. 1. layout of the files. I see that lzma is currently using C folder to house the C.h and deimos to house the .d files. Is this the standard that all of the repos should follow? I have simply put the .d files in the top level. Which would be the preferred method?The latter (.d files at the top level).
Nov 14 2011
On Tue 15 Nov 2011 01:58:21 AM CST, Jacob Carlborg wrote:On 2011-11-15 06:06, Walter Bright wrote:conflicts how? I'd like to hear a simple example if you don't mind. I don't fully understand all of the implications..On 11/14/2011 5:54 AM, Jude Young wrote:That doesn't sound like a good idea. I think there should be one top level package that has the same name as the library and but every D module in that directory, keeping any directory structure that the C headers have. Otherwise there will be a lot of conflicts.Ok, I think that the ncurses bindings are about ready. Just something to remember: ncurses is HUGE.The header files, too?There are probably hidden bugs somewhere. So there are a few questions that I'd like answered before I send a pull request. 1. layout of the files. I see that lzma is currently using C folder to house the C.h and deimos to house the .d files. Is this the standard that all of the repos should follow? I have simply put the .d files in the top level. Which would be the preferred method?The latter (.d files at the top level).
Nov 15 2011
On 2011-11-13 04:00, Walter Bright wrote:On 11/5/2011 2:28 AM, Johannes Pfau wrote:I have bindings for Ruby and Clang ready. I'm going to wait with the Clang bindings until they make a final release of 3.0, they're currently at release candidate 3. -- /Jacob CarlborgNow that deimos is an organization and we can't use pull requests, how do we get new bindings into deimos? I'd still like to get my liblzma bindings in there: https://github.com/jpf91/deimos/tree/liblzmahttps://github.com/D-Programming-Deimos/liblzma Next step: create a pull request for it.
Nov 13 2011
will the bindings project from dsource be added?
Nov 13 2011