Re: [Gimp-developer] Example: Vala as code generator
2011/5/3 Simon Budig si...@budig.de: Martin Nordholts (ense...@gmail.com) wrote: I'm trying hard to find time hacking on GIMP, and not having to waste time on GObject C boiler plate means a lot to me. At first I was thinking what the hell, I'll just come up with the the damn boilerplate code manually then. But right after I began doing that I started to feel like I was wasting my time, and I can't stand that feeling. Hm. This paragraphs leaves me a bit perplexed, because it gives the impression that the most important thing about including vala is to make you more comfortable with our codebase. You blame mitch for a blunt dismissal, but this reads a lot like bluntly forcing down something through mitchs throat. Not sure if that is any better. You are right, that isn't any better. I should just give up on these patches, I clearly don't have the support for them I hoped for. Obviously, in my opinion we increase the quality of our codebase by using Vala for this helper class mostly because the number of readable and documented version controlled lines of code is less than if we would also version control the GObject C boiler plate. That is not the only measurement of code quality however and we are simply weighting the pros and cons differently. / Martin -- My GIMP Blog: http://www.chromecode.com/ GIMP 2.8 schedule on tasktaste.com ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
Re: [Gimp-developer] Example: Vala as code generator
On Sun, 2011-05-01 at 17:15 +0200, ense...@gmail.com wrote: Hi all, We have discussed the possibility to use Vala as a code generator. What follows is a set of patches so we all can see what that would look like. Any objections to me pushing these commits? Yes, and I'm currently on vacation and can't respond in detail, but IIRC I have made myself very clear about vala. I want a helper class in the implementation of single-window mode session management, and this is in my opinion a perfect example of when to use Vala: a small utility class at the bottom of the architecture with only GLib as a dependency. These patches do not add any additional tarball build time dependencies, valac is only required when building from git. / Martin ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
Re: [Gimp-developer] Example: Vala as code generator
The following statement of Mich's unfortunately describes most of nowadays C.S. students [1] And this is *exactly* the problem. We would end up with programmers that quickly learnt vala, having no clue about GObject. That's absolutely horrible. Just like the people who only know how to write java or #C code. They know how to use all the fancy classes, but they have never implemented a list or anything lowlevel themselves. I don't want people who know vala, but don't had to learn GObject. One way of viewing the issue is: Do you want to cater to the programmers you want, or the programmers you have? (With the hope of turning the latter into the former.) Nicolas Robidoux [1] research.cs.queensu.ca/~cisc422/2010f/readings/papers/whereAreTheSoftwareEngineersOfTomorrow.pdf ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
Re: [Gimp-developer] Example: Vala as code generator
On Mon, May 2, 2011 at 9:15 AM, Nicolas Robidoux nicolas.robid...@gmail.com wrote: One way of viewing the issue is: Do you want to cater to the programmers you want, or the programmers you have? (With the hope of turning the latter into the former.) It's worse than that, even. This is a perpetual argument. People were railing against C in the 70s because people who weren't programming in assembler weren't intimate with the stack, heap, or registers. They hated the subroutine call inefficiency. They hated losing the ability to dynamically change executing code, or freely use instructions as data. Software development - all engineering, really - is about trades. C imposed a base level of inefficiency for a higher level of productivity. Today, C is a hindrance to even greater productivity (GObject itself is evidence of this). It's less useful to know how to write a hash table interface with its own unique API than it is to know how or when to use one. If a hash table is part of the language, then it becomes the One True Implementation and Everybody knows it and uses it, and does so mostly-correctly. FWIW, to me, Vala seems like a reasonable choice for GObject based systems. I don't think it's going to expand the developer base now because Vala prefers an understanding of GObject, which is little used outside OSS systems. The syntax should be familiar to C/C++/C# users, but the trade here is for productivity. While another language (C++, C# or Java) may have more developers, that language would require an object system be built around GObject, which just raises the question: Why not Vala? One would assume that a stable Vala compiler exists or is currently stable enough to use on all target platforms. Chris ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
Re: [Gimp-developer] Example: Vala as code generator
2011/5/2 Michael Natterer mi...@gimp.org: On Sun, 2011-05-01 at 17:15 +0200, ense...@gmail.com wrote: Hi all, We have discussed the possibility to use Vala as a code generator. What follows is a set of patches so we all can see what that would look like. Any objections to me pushing these commits? Yes, and I'm currently on vacation and can't respond in detail, but IIRC I have made myself very clear about vala. Hi Michael, I'm trying hard to find time hacking on GIMP, and not having to waste time on GObject C boiler plate means a lot to me. At first I was thinking what the hell, I'll just come up with the the damn boilerplate code manually then. But right after I began doing that I started to feel like I was wasting my time, and I can't stand that feeling. I find your blunt dismissal of these two patches really discouraging. Can't we at least push them to master and have them in for a while and see if we can discover some real problems? If it really doesn't work out, we can just reformat the generated .h and .c files and discard the .vala file. If you are on vacation and don't have time to properly review or test these patches, please take your time to do so. I'm not going to push these patches without your OK, and if you're busy for a few weeks, then I'll have to wait. I can work on other items on the GIMP 2.8 schedule in the meantime... Regards, Martin -- My GIMP Blog: http://www.chromecode.com/ GIMP 2.8 schedule on tasktaste.com ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
Re: [Gimp-developer] Example: Vala as code generator
Martin Nordholts (ense...@gmail.com) wrote: I'm trying hard to find time hacking on GIMP, and not having to waste time on GObject C boiler plate means a lot to me. At first I was thinking what the hell, I'll just come up with the the damn boilerplate code manually then. But right after I began doing that I started to feel like I was wasting my time, and I can't stand that feeling. Hm. This paragraphs leaves me a bit perplexed, because it gives the impression that the most important thing about including vala is to make you more comfortable with our codebase. You blame mitch for a blunt dismissal, but this reads a lot like bluntly forcing down something through mitchs throat. Not sure if that is any better. I must say that - while I have my share of frustrations with the gobject boilerplate code - I don't think that adding vala helps the quality of the Gimp codebase. And this is IMHO what should be in our focus. From reading a lot of code in a lot of different free software projects I have to say, that Gimp still is one of the shining examples regarding consistency and quality of the source code. This has a lot to do with mainly Sven and Mitch investing countless hours into refactoring and restructuring the code, enforcing a common code structure and generally fighting against sloppiness. I don't see how mixing this code with code written in a new language will help the quality of the source. Vala code will inherently develop a set of patterns that differs from the C code (and don't underestimate the value of grepping through the code base). Also new contributors will not only have to learn about GObject and GTK+, they have to learn vala as well. And adding vala *is* adding a new entry barrier, vala is by no means a lingua franca. Bye, Simon -- si...@budig.de http://simon.budig.de/ ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
