From transfire at gmail.com Sat Feb 23 20:31:14 2008 From: transfire at gmail.com (Trans) Date: Sat, 23 Feb 2008 20:31:14 -0500 Subject: [Rake-devel] Rakefile optional when using a rake directory Message-ID: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> Hi-- I'd like to suggest a simple feature addition to Rake. I would like to define rake tasks in a default subdirectory (preferably rake/) and have rake automatically load any files in it having a .rake extension when there is no Rakefile. If the Rakefile is present it would defer to it and not automatically load them. So this would strictly be an additional feature, and not cause any backward compatibility issues. I would very much prefer this for keeping my project folders as concise as possible. Right now nearly all my Rakefiles have but a single line in them telling them to load from the subdirectory, something which seems not all that uncommon these days. So it would be very nice not to have to bother with the Rakefile at all in these cases. Thanks, T. From assaf at labnotes.org Sat Feb 23 20:47:28 2008 From: assaf at labnotes.org (Assaf Arkin) Date: Sat, 23 Feb 2008 17:47:28 -0800 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> Message-ID: <8B0337DD-AE18-41AD-88D2-1E1250E28BDA@labnotes.org> On Feb 23, 2008, at 5:31 PM, Trans wrote: > Hi-- > > I'd like to suggest a simple feature addition to Rake. I would like to > define rake tasks in a default subdirectory (preferably rake/) and > have rake automatically load any files in it having a .rake extension > when there is no Rakefile. If the Rakefile is present it would defer > to it and not automatically load them. So this would strictly be an > additional feature, and not cause any backward compatibility issues. > > I would very much prefer this for keeping my project folders as > concise as possible. Right now nearly all my Rakefiles have but a > single line in them telling them to load from the subdirectory, > something which seems not all that uncommon these days. So it would be > very nice not to have to bother with the Rakefile at all in these > cases. I'm not sure if this is documented anywhere, but if you place any file with the .rake extension inside the rakelib sub-directory, relative to the Rakefile, it will be automatically imported after the Rakefile is loaded. Assaf http://labnotes.org > > > Thanks, > T. > _______________________________________________ > Rake-devel mailing list > Rake-devel at rubyforge.org > http://rubyforge.org/mailman/listinfo/rake-devel From adam.q.salter at gmail.com Sat Feb 23 20:55:25 2008 From: adam.q.salter at gmail.com (Adam Salter) Date: Sun, 24 Feb 2008 12:55:25 +1100 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> Message-ID: <52CBB945-0468-47E0-ADEE-BA3442B32F88@gmail.com> Ummm, a suggestion without a patch is pretty useless. I'm not saying that the suggestion would definately go into the public release anyway, but really... if you have something you want to see, code it and let the admins decide if they want to implement it. On 24/02/2008, at 12:31 PM, Trans wrote: > Hi-- > > I'd like to suggest a simple feature addition to Rake. I would like to > define rake tasks in a default subdirectory (preferably rake/) and > have rake automatically load any files in it having a .rake extension > when there is no Rakefile. If the Rakefile is present it would defer > to it and not automatically load them. So this would strictly be an > additional feature, and not cause any backward compatibility issues. > > I would very much prefer this for keeping my project folders as > concise as possible. Right now nearly all my Rakefiles have but a > single line in them telling them to load from the subdirectory, > something which seems not all that uncommon these days. So it would be > very nice not to have to bother with the Rakefile at all in these > cases. > > Thanks, > T. > _______________________________________________ > Rake-devel mailing list > Rake-devel at rubyforge.org > http://rubyforge.org/mailman/listinfo/rake-devel From transfire at gmail.com Sat Feb 23 21:19:02 2008 From: transfire at gmail.com (Trans) Date: Sat, 23 Feb 2008 21:19:02 -0500 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: <8B0337DD-AE18-41AD-88D2-1E1250E28BDA@labnotes.org> References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <8B0337DD-AE18-41AD-88D2-1E1250E28BDA@labnotes.org> Message-ID: <4b6f054f0802231819v32867d20ge3ab6d2de925659c@mail.gmail.com> On Sat, Feb 23, 2008 at 8:47 PM, Assaf Arkin wrote: > > On Feb 23, 2008, at 5:31 PM, Trans wrote: > > > Hi-- > > > > I'd like to suggest a simple feature addition to Rake. I would like to > > define rake tasks in a default subdirectory (preferably rake/) and > > have rake automatically load any files in it having a .rake extension > > when there is no Rakefile. If the Rakefile is present it would defer > > to it and not automatically load them. So this would strictly be an > > additional feature, and not cause any backward compatibility issues. > > > > I would very much prefer this for keeping my project folders as > > concise as possible. Right now nearly all my Rakefiles have but a > > single line in them telling them to load from the subdirectory, > > something which seems not all that uncommon these days. So it would be > > very nice not to have to bother with the Rakefile at all in these > > cases. > > I'm not sure if this is documented anywhere, but if you place any file > with the .rake extension inside the rakelib sub-directory, relative to > the Rakefile, it will be automatically imported after the Rakefile is > loaded. Ah. Okay, I didn't recollect that was automatic. But of course that's not what I'd like to have either. It's isn't as useful that way --if I have a Rakefile it's a fairly simple mater of loading other rake scripts by hand, and I can conditionally control that if I want too. I'd like it so the Rakefile itself can be done without. (Plus I don't like the name rakelib --all my other directories are 3-4 charchters long --picky I know ;-) T. From transfire at gmail.com Sat Feb 23 21:52:01 2008 From: transfire at gmail.com (Trans) Date: Sat, 23 Feb 2008 21:52:01 -0500 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: <52CBB945-0468-47E0-ADEE-BA3442B32F88@gmail.com> References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <52CBB945-0468-47E0-ADEE-BA3442B32F88@gmail.com> Message-ID: <4b6f054f0802231852o7e105a28j2b6d2c189592ba41@mail.gmail.com> On Sat, Feb 23, 2008 at 8:55 PM, Adam Salter wrote: > Ummm, a suggestion without a patch is pretty useless. > I'm not saying that the suggestion would definately go into the public > release anyway, but really... if you have something you want to see, > code it and let the admins decide if they want to implement it. Jim would probably do a better job of it, but just the same this should do, in rake.rb: class Application # ADD 'rake' to end of list DEFAULT_RAKEFILES = ['rakefile', 'Rakefile', 'rakefile.rb', 'Rakefile.rb', 'rake'].freeze def raw_load_rakefile # :nodoc: here = Dir.pwd while ! have_rakefile Dir.chdir("..") if Dir.pwd == here || options.nosearch fail "No Rakefile found (looking for: #{@rakefiles.join(', ')})" end here = Dir.pwd end puts "(in #{Dir.pwd})" unless options.silent $rakefile = @rakefile # -- NEW CODE if File.directory?(@rakefile) options.rakelib |= @rakefile @rakefile = '' end # -- END NEW CODE load File.expand_path(@rakefile) if @rakefile != '' options.rakelib.each do |rlib| Dir["#{rlib}/*.rake"].each do |name| add_import name end end load_imports end I'm not sure why the possibility for @rakefile == '' even exists, but I utilized it for my patch. If for some reason that's not appropriate, then extending the if-clause to include the loading of the rakefile in an else-clause would suffice. T. From jim.weirich at gmail.com Sat Feb 23 22:57:20 2008 From: jim.weirich at gmail.com (Jim Weirich) Date: Sat, 23 Feb 2008 22:57:20 -0500 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: <4b6f054f0802231819v32867d20ge3ab6d2de925659c@mail.gmail.com> References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <8B0337DD-AE18-41AD-88D2-1E1250E28BDA@labnotes.org> <4b6f054f0802231819v32867d20ge3ab6d2de925659c@mail.gmail.com> Message-ID: On Feb 23, 2008, at 9:19 PM, Trans wrote: > But of course that's not what I'd like to have either. It's isn't as > useful that way --if I have a Rakefile it's a fairly simple mater of > loading other rake scripts by hand, and I can conditionally control > that if I want too. I'd like it so the Rakefile itself can be done > without. (Plus I don't like the name rakelib --all my other > directories are 3-4 charchters long --picky I know ;-) Use -R to change the name of rakelib to whatever you want. If you want to get rid of the Rakefile itself, create an empty file somewhere convenient and use the -f option to specify it. And since you probably don't won't to type all that stuff out, create a small alias or batch file for it. All the tools are there. -- -- Jim Weirich -- jim.weirich at gmail.com From transfire at gmail.com Sun Feb 24 01:17:30 2008 From: transfire at gmail.com (Trans) Date: Sun, 24 Feb 2008 01:17:30 -0500 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <8B0337DD-AE18-41AD-88D2-1E1250E28BDA@labnotes.org> <4b6f054f0802231819v32867d20ge3ab6d2de925659c@mail.gmail.com> Message-ID: <4b6f054f0802232217n70217c41r20a5f32234890e06@mail.gmail.com> On Sat, Feb 23, 2008 at 10:57 PM, Jim Weirich wrote: > On Feb 23, 2008, at 9:19 PM, Trans wrote: > > But of course that's not what I'd like to have either. It's isn't as > > useful that way --if I have a Rakefile it's a fairly simple mater of > > loading other rake scripts by hand, and I can conditionally control > > that if I want too. I'd like it so the Rakefile itself can be done > > without. (Plus I don't like the name rakelib --all my other > > directories are 3-4 charchters long --picky I know ;-) > > Use -R to change the name of rakelib to whatever you want. If you > want to get rid of the Rakefile itself, create an empty file somewhere > convenient and use the -f option to specify it. And since you > probably don't won't to type all that stuff out, create a small alias > or batch file for it. All the tools are there. Sure, but that's not a distributable solution. T. From transfire at gmail.com Mon Feb 25 06:54:14 2008 From: transfire at gmail.com (Trans) Date: Mon, 25 Feb 2008 06:54:14 -0500 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <8B0337DD-AE18-41AD-88D2-1E1250E28BDA@labnotes.org> <4b6f054f0802231819v32867d20ge3ab6d2de925659c@mail.gmail.com> Message-ID: <4b6f054f0802250354u4ac971t9bf466365738160e@mail.gmail.com> On Sat, Feb 23, 2008 at 10:57 PM, Jim Weirich wrote: > On Feb 23, 2008, at 9:19 PM, Trans wrote: > > But of course that's not what I'd like to have either. It's isn't as > > useful that way --if I have a Rakefile it's a fairly simple mater of > > loading other rake scripts by hand, and I can conditionally control > > that if I want too. I'd like it so the Rakefile itself can be done > > without. (Plus I don't like the name rakelib --all my other > > directories are 3-4 charchters long --picky I know ;-) > > Use -R to change the name of rakelib to whatever you want. If you > want to get rid of the Rakefile itself, create an empty file somewhere > convenient and use the -f option to specify it. And since you > probably don't won't to type all that stuff out, create a small alias > or batch file for it. All the tools are there. Would it help if I said, "Please"? Please... with sugar on top. T. From hgs at dmu.ac.uk Mon Feb 25 08:09:53 2008 From: hgs at dmu.ac.uk (Hugh Sasse) Date: Mon, 25 Feb 2008 13:09:53 +0000 (WET) Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: <52CBB945-0468-47E0-ADEE-BA3442B32F88@gmail.com> References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <52CBB945-0468-47E0-ADEE-BA3442B32F88@gmail.com> Message-ID: On Sun, 24 Feb 2008, Adam Salter wrote: > Ummm, a suggestion without a patch is pretty useless. Are you saying that discussing an idea with people who know the project inside out, who could tell you whether another approach would meet your needs (better, perhaps), who could save you writing and debugging code, and who could save themselves having to read it would be a waste of time? Are you saying that you should not contribute ideas to one project if your commitments to contribute code for other projects means you have no available time? Hugh From hgs at dmu.ac.uk Mon Feb 25 08:20:43 2008 From: hgs at dmu.ac.uk (Hugh Sasse) Date: Mon, 25 Feb 2008 13:20:43 +0000 (WET) Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: <4b6f054f0802250354u4ac971t9bf466365738160e@mail.gmail.com> References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <8B0337DD-AE18-41AD-88D2-1E1250E28BDA@labnotes.org> <4b6f054f0802231819v32867d20ge3ab6d2de925659c@mail.gmail.com> <4b6f054f0802250354u4ac971t9bf466365738160e@mail.gmail.com> Message-ID: On Mon, 25 Feb 2008, Trans wrote: > On Sat, Feb 23, 2008 at 10:57 PM, Jim Weirich wrote: > > On Feb 23, 2008, at 9:19 PM, Trans wrote: > > > But of course that's not what I'd like to have either. It's isn't as > > > useful that way --if I have a Rakefile it's a fairly simple mater of > > > loading other rake scripts by hand, and I can conditionally control > > > that if I want too. I'd like it so the Rakefile itself can be done > > > without. (Plus I don't like the name rakelib --all my other > > > directories are 3-4 charchters long --picky I know ;-) > > > > Use -R to change the name of rakelib to whatever you want. If you > > want to get rid of the Rakefile itself, create an empty file somewhere > > convenient and use the -f option to specify it. And since you > > probably don't won't to type all that stuff out, create a small alias > > or batch file for it. All the tools are there. > > Would it help if I said, "Please"? > > Please... with sugar on top. Can I suggest you explain your goal a bit more? Why do you want things to work this different way? People come to rake with the expectation that it is like a Ruby Make, and has a Rakefile as a result of that heritage. Why is it good to break that metaphor? You've explained how you would like things to work, but not why, or how this would help Rake users in general in the future. At the moment I know to look for a ./configure, a Makefile, a Rakefile, or a setup.rb in a downloaded ruby project. This approach would give me more things to consider to understand what is going on. > > T. Hugh From jason.lapier at gmail.com Mon Feb 25 15:55:35 2008 From: jason.lapier at gmail.com (Jason LaPier) Date: Mon, 25 Feb 2008 12:55:35 -0800 Subject: [Rake-devel] truncated task descriptions Message-ID: <93b44c1f0802251255t62ca5eb8gb4aca18e18c6aa01@mail.gmail.com> Hey all - This is an extremely trivial point, so don't spend too much time on it :) I understand the rationale behind the truncation of task descriptions in 'rake -T' but I find the 80 column limit a little too condensed for my personal tastes. I whipped up a rake task that displays tasks and allows me to specify my column limit with an environment variable. Something like: $ export RAKE_DESC_COLS=180 $ rake t $ rake t[some-pattern] The rake task is detailed here: http://offtheline.net/2008/2/24/rake-and-truncated-task-descriptions If anyone thinks they might find this useful, I'd be willing to create a patch (to work with '-T' rather than a task named 't'). Let me know if you'd prefer using an environment variable or another command-line switch. Personally, I like the environment variable because I can stick it in my .bashrc and have different column width depending on what computer I'm in front of. - Jason L. -- My Rails and Linux Blog: http://offtheline.net From jim.weirich at gmail.com Mon Feb 25 16:08:25 2008 From: jim.weirich at gmail.com (Jim Weirich) Date: Mon, 25 Feb 2008 16:08:25 -0500 Subject: [Rake-devel] truncated task descriptions In-Reply-To: <93b44c1f0802251255t62ca5eb8gb4aca18e18c6aa01@mail.gmail.com> References: <93b44c1f0802251255t62ca5eb8gb4aca18e18c6aa01@mail.gmail.com> Message-ID: On Feb 25, 2008, at 3:55 PM, Jason LaPier wrote: > This is an extremely trivial point, so don't spend too much time on it > :) I understand the rationale behind the truncation of task > descriptions in 'rake -T' but I find the 80 column limit a little too > condensed for my personal tastes. I whipped up a rake task that > displays tasks and allows me to specify my column limit with an > environment variable. Something like: > $ export RAKE_DESC_COLS=180 > $ rake t > $ rake t[some-pattern] > > The rake task is detailed here: > http://offtheline.net/2008/2/24/rake-and-truncated-task-descriptions > If anyone thinks they might find this useful, I'd be willing to create > a patch (to work with '-T' rather than a task named 't'). Let me know > if you'd prefer using an environment variable or another command-line > switch. Personally, I like the environment variable because I can > stick it in my .bashrc and have different column width depending on > what computer I'm in front of. I would accept a patch that allowed an environment variable to set the column width for -T. Don't forget unit tests. -- -- Jim Weirich -- jim.weirich at gmail.com From jason.lapier at gmail.com Mon Feb 25 19:54:57 2008 From: jason.lapier at gmail.com (Jason LaPier) Date: Mon, 25 Feb 2008 16:54:57 -0800 Subject: [Rake-devel] truncated task descriptions In-Reply-To: References: <93b44c1f0802251255t62ca5eb8gb4aca18e18c6aa01@mail.gmail.com> Message-ID: <93b44c1f0802251654l73151e1bj2940715de8b1dca3@mail.gmail.com> On Mon, Feb 25, 2008 at 1:08 PM, Jim Weirich wrote: > > On Feb 25, 2008, at 3:55 PM, Jason LaPier wrote: > > This is an extremely trivial point, so don't spend too much time on it > > :) I understand the rationale behind the truncation of task > > descriptions in 'rake -T' but I find the 80 column limit a little too > > condensed for my personal tastes. I whipped up a rake task that > > displays tasks and allows me to specify my column limit with an > > environment variable. Something like: > > $ export RAKE_DESC_COLS=180 > > $ rake t > > $ rake t[some-pattern] > > > > The rake task is detailed here: > > http://offtheline.net/2008/2/24/rake-and-truncated-task-descriptions > > If anyone thinks they might find this useful, I'd be willing to create > > a patch (to work with '-T' rather than a task named 't'). Let me know > > if you'd prefer using an environment variable or another command-line > > switch. Personally, I like the environment variable because I can > > stick it in my .bashrc and have different column width depending on > > what computer I'm in front of. > > I would accept a patch that allowed an environment variable to set the > column width for -T. Don't forget unit tests. > > -- > -- Jim Weirich > -- jim.weirich at gmail.com > Cool, thanks. Patch file attached (test included). - Jason L. -- My Rails and Linux Blog: http://offtheline.net -------------- next part -------------- A non-text attachment was scrubbed... Name: rake_add_col_width.diff Type: text/x-patch Size: 1567 bytes Desc: not available Url : http://rubyforge.org/pipermail/rake-devel/attachments/20080225/e3bec838/attachment-0001.bin From adam.q.salter at gmail.com Tue Feb 26 07:04:27 2008 From: adam.q.salter at gmail.com (Adam Salter) Date: Tue, 26 Feb 2008 23:04:27 +1100 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <52CBB945-0468-47E0-ADEE-BA3442B32F88@gmail.com> Message-ID: <3DE1F991-C869-4DE6-92FF-0793DFE794E2@gmail.com> Hugh, I suspect you understand the actual tone with which my comment was expressed (albeit a little hurriedly)... Of course it's good to make a request. But it's only a request. A request and a patch is better. Generally in the "open source" world it is necessary to be "the change you want to see". In other words if you really want to see something done do it yourself. Submit a patch, _and_ discuss/request. We are all busy people and I'm sure the principle maintainer, Jim, has his own things to think about. As it happens Transfire was quick to supply a patch of sorts, but I'm not sure I like loading all files in a directory as rakefiles or running without an explicit rakefile. There is sake if you need it (http://errtheblog.com/posts/60-sake-bomb). And Transfire could always patch his own version (or even fork rake) if he really wants his functionality. On 26/02/2008, at 12:09 AM, Hugh Sasse wrote: > On Sun, 24 Feb 2008, Adam Salter wrote: > >> Ummm, a suggestion without a patch is pretty useless. > > Are you saying that discussing an idea with people who know the > project inside out, who could tell you whether another approach > would meet your needs (better, perhaps), who could save you writing > and debugging code, and who could save themselves having to read it > would be a waste of time? Are you saying that you should not > contribute ideas to one project if your commitments to contribute > code for other projects means you have no available time? > > Hugh > > _______________________________________________ > Rake-devel mailing list > Rake-devel at rubyforge.org > http://rubyforge.org/mailman/listinfo/rake-devel From transfire at gmail.com Tue Feb 26 07:52:03 2008 From: transfire at gmail.com (Trans) Date: Tue, 26 Feb 2008 07:52:03 -0500 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <8B0337DD-AE18-41AD-88D2-1E1250E28BDA@labnotes.org> <4b6f054f0802231819v32867d20ge3ab6d2de925659c@mail.gmail.com> <4b6f054f0802250354u4ac971t9bf466365738160e@mail.gmail.com> Message-ID: <4b6f054f0802260452g1b8b21e9n295c1d36a44824a5@mail.gmail.com> On Mon, Feb 25, 2008 at 8:20 AM, Hugh Sasse wrote: > Can I suggest you explain your goal a bit more? Why do you want > things to work this different way? People come to rake with the > expectation that it is like a Ruby Make, and has a Rakefile as a > result of that heritage. Why is it good to break that metaphor? > You've explained how you would like things to work, but not why, or > how this would help Rake users in general in the future. At the > moment I know to look for a ./configure, a Makefile, a Rakefile, or > a setup.rb in a downloaded ruby project. This approach would give > me more things to consider to understand what is going on. You make a good point. There is reason to think of Rake as Ruby Make, it is where Rake has it's roots. But should we let that connection hold Rake back from advancement? It doesn't seem the case since we already have auto-loading of rakelib/ and now task arguments, two features that Make (AFAIK) has no equivalent. My reasoning for the suggestion is to further the principle of DRY. Yes, it would mean adding one more item to your list of things to expect, namely a rake/ directoy. But on three counts I suspect that's not much to ask. Most of the time we already have a good hunch of when a project will support Rake (almost any Ruby projects for instance); documentation will likely inform us in anycase; and scanning for "rake" is rather isomorphic to scanning for "rakefile". The upshot is a cleaner project directory. Instead of, for instance, README Rakefile bin doc lib rakelib test it's README bin doc lib rake test Saving us the unecessary repetition, being more concise, and fitting in more neatly with the other common project directories. T. From hgs at dmu.ac.uk Tue Feb 26 07:54:46 2008 From: hgs at dmu.ac.uk (Hugh Sasse) Date: Tue, 26 Feb 2008 12:54:46 +0000 (WET) Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: <3DE1F991-C869-4DE6-92FF-0793DFE794E2@gmail.com> References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <52CBB945-0468-47E0-ADEE-BA3442B32F88@gmail.com> <3DE1F991-C869-4DE6-92FF-0793DFE794E2@gmail.com> Message-ID: On Tue, 26 Feb 2008, Adam Salter wrote: > Hugh, > I suspect you understand the actual tone with which my comment was > expressed (albeit a little hurriedly)... Tone never comes across well in email, alas. > Of course it's good to make a request. But it's only a request. A > request and a patch is better. Agreed. That is the ideal case. > Generally in the "open source" world it is necessary to be "the change > you want to see". In other words if you really want to see something > done do it yourself. Submit a patch, _and_ discuss/request. We are all This is what I'd really like to challenge. Why && how? There is a desire that open source spread because of its positive benefits. If the only RFEs we will accept are from "customers" who can, are willing to, and have sufficient time to code, then this desire is "toast". Clearly, too much, too far in the other direction wastes time, So the open source world needs to encourage people to make *intelligent* suggestions, which doesn't just mean code; evidence of use cases is an example of something that would be intelligent. A significant class of talented programmers don't interact well with other people. Not universally true of course, and too much a stereotype, but there is truth in this. So we need input from people with more user-centric skills to make open source more usable, if it is to become truly mainstream. This won't happen while "We only accept patches as change suggestions" is a refrain. > busy people and I'm sure the principle maintainer, Jim, has his own > things to think about. And this is another one. There seems to be an idea in the open source world that developers are the only really busy people. Isn't the time wasted by users of the program of any importance? Never mind the problems of the poor souls who have to install some package with a huge dependency graph, on N different architectures. I'm constantly seeing people told this, and if we want open source to prosper then finding ways to drive interested people away is not going to help. Would it kill developers to put sensible suggestions on a prioritized wish list? Something may only get implemented in 2037, but at least the person making the request feels they have been heard. > > As it happens Transfire was quick to supply a patch of sorts, but I'm That's the problem with sending a patch first without allowing discussion. It gets killed for some implementation detail which could have been prevented by discussion. > not sure I like loading all files in a directory as rakefiles or > running without an explicit rakefile. So I asked him for more details of why it was needed in case his case is really compelling, but I (for one) haven't "got it". > There is sake if you need it (http://errtheblog.com/posts/60-sake-bomb). > And Transfire could always patch his own version (or even fork rake) > if he really wants his functionality. Yes, of course. There are disadvantages here, which we need not discuss. > Hugh From adam.q.salter at gmail.com Wed Feb 27 04:57:44 2008 From: adam.q.salter at gmail.com (Adam Salter) Date: Wed, 27 Feb 2008 20:57:44 +1100 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <52CBB945-0468-47E0-ADEE-BA3442B32F88@gmail.com> <3DE1F991-C869-4DE6-92FF-0793DFE794E2@gmail.com> Message-ID: <15474463-1F67-46DD-A8BF-50FE65985E33@gmail.com> On 26/02/2008, at 11:54 PM, Hugh Sasse wrote: > This is what I'd really like to challenge. Why && how? There is a > desire that open source spread because of its positive benefits. If > the only RFEs we will accept are from "customers" who can, are > willing to, and have sufficient time to code, then this desire is > "toast". Clearly, too much, too far in the other direction wastes > time, So the open source world needs to encourage people to make > *intelligent* suggestions, which doesn't just mean code; evidence of > use cases is an example of something that would be intelligent. I will try not to jump so quickly on anybody who makes a request next time... I agree it can be daunting for those new to the scene. I still think a patch is the best way to get your changes made... and having said that Rake is a developer tool, so requests are likely to come from people who should be able to make a patch... ;) and often preparation of a patch also helps the requester better understand what they are asking for (I know it does for me)... Best, -Adam From jim.weirich at gmail.com Wed Feb 27 09:28:06 2008 From: jim.weirich at gmail.com (Jim Weirich) Date: Wed, 27 Feb 2008 09:28:06 -0500 Subject: [Rake-devel] Suggestions, Patches and Unit Tests Message-ID: <74A5D427-580C-43D6-8466-02199BA61381@gmail.com> Since the topic has been discussed somewhat on the mailing list, I thought I would put forth my own preferences in these matters. First of all, suggestions are always welcome. They may or may not be accepted, but I always love hearing suggestions on improvements to Rake (or any other library I work on). Patches supplied with suggestions are nice gesture. If you can supply a patch, feel free to do so. If you can include unit tests with that patch, then that is even better. In fact, if you wish to supply unit tests for a new feature, but no implementation, that is fine too. In general, the lack of a patch (or unit tests) rarely have an impact on the acceptance (or rejection) of a particular suggestion. What they *do* impact is how quickly you might see a particular suggestion implemented. Thanks to everyone who participates in the Rake discussions. -- -- Jim Weirich -- jim.weirich at gmail.com From jim.weirich at gmail.com Wed Feb 27 09:38:26 2008 From: jim.weirich at gmail.com (Jim Weirich) Date: Wed, 27 Feb 2008 09:38:26 -0500 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: <4b6f054f0802250354u4ac971t9bf466365738160e@mail.gmail.com> References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <8B0337DD-AE18-41AD-88D2-1E1250E28BDA@labnotes.org> <4b6f054f0802231819v32867d20ge3ab6d2de925659c@mail.gmail.com> <4b6f054f0802250354u4ac971t9bf466365738160e@mail.gmail.com> Message-ID: On Feb 25, 2008, at 6:54 AM, Trans wrote: > On Sat, Feb 23, 2008 at 10:57 PM, Jim Weirich > wrote: >> On Feb 23, 2008, at 9:19 PM, Trans wrote: >>> But of course that's not what I'd like to have either. It's isn't as >>> useful that way --if I have a Rakefile it's a fairly simple mater of >>> loading other rake scripts by hand, and I can conditionally control >>> that if I want too. I'd like it so the Rakefile itself can be done >>> without. (Plus I don't like the name rakelib --all my other >>> directories are 3-4 charchters long --picky I know ;-) >> >> Use -R to change the name of rakelib to whatever you want. If you >> want to get rid of the Rakefile itself, create an empty file >> somewhere >> convenient and use the -f option to specify it. And since you >> probably don't won't to type all that stuff out, create a small alias >> or batch file for it. All the tools are there. > > Would it help if I said, "Please"? As I understand it, there are two requests here. (1) Change the name of rakelib to rake (or maybe just added rake in addition to rakelib?) (2) Allow the omission of the Rakefile if rake (or rakelib?) is present. On (1), I am disinclined to change the name of rakelib (or add additional redundant functionality) in the absence of a compelling reason to do so. I believe the reason you assert if merely a preference for the shorter name, correct? On (2), this is a more interesting proposal. I'm not sure I would use the feature myself, but if there is a wider desire amoung rake users for this, I might consider it. Fair 'nuff? -- -- Jim Weirich -- jim.weirich at gmail.com From assaf at labnotes.org Wed Feb 27 12:57:20 2008 From: assaf at labnotes.org (Assaf Arkin) Date: Wed, 27 Feb 2008 09:57:20 -0800 Subject: [Rake-devel] Rakefile optional when using a rake directory In-Reply-To: <3DE1F991-C869-4DE6-92FF-0793DFE794E2@gmail.com> References: <4b6f054f0802231731h3b739480y106e76252b74739c@mail.gmail.com> <52CBB945-0468-47E0-ADEE-BA3442B32F88@gmail.com> <3DE1F991-C869-4DE6-92FF-0793DFE794E2@gmail.com> Message-ID: <5037b6e40802270957n62dcc336wc60ade425de281a@mail.gmail.com> On Tue, Feb 26, 2008 at 4:04 AM, Adam Salter wrote: > Hugh, > I suspect you understand the actual tone with which my comment was > expressed (albeit a little hurriedly)... > Of course it's good to make a request. But it's only a request. A > request and a patch is better. I beg to differ. If you need something, ask for it. There's a high chance that it already exists just not apparent to you, someone else is working on it or motivated to make it happen by your request, that it was attempted before but didn't work, doesn't fit the goals of the project, or that there's a better way of doing things. Asking costs nothing, and you get more people involved in figuring out the solution. The only rule is, don't expect anything to happen without putting effort into it. The effort to make a code change happens involves a patch, but the right way to start is often enough to ask first. Assaf > Generally in the "open source" world it is necessary to be "the change > you want to see". In other words if you really want to see something > done do it yourself. Submit a patch, _and_ discuss/request. We are all > busy people and I'm sure the principle maintainer, Jim, has his own > things to think about. > > As it happens Transfire was quick to supply a patch of sorts, but I'm > not sure I like loading all files in a directory as rakefiles or > running without an explicit rakefile. > There is sake if you need it (http://errtheblog.com/posts/60-sake-bomb). > And Transfire could always patch his own version (or even fork rake) > if he really wants his functionality. > > > > On 26/02/2008, at 12:09 AM, Hugh Sasse wrote: > > > On Sun, 24 Feb 2008, Adam Salter wrote: > > > >> Ummm, a suggestion without a patch is pretty useless. > > > > Are you saying that discussing an idea with people who know the > > project inside out, who could tell you whether another approach > > would meet your needs (better, perhaps), who could save you writing > > and debugging code, and who could save themselves having to read it > > would be a waste of time? Are you saying that you should not > > contribute ideas to one project if your commitments to contribute > > code for other projects means you have no available time? > > > > Hugh > > > > _______________________________________________ > > Rake-devel mailing list > > Rake-devel at rubyforge.org > > http://rubyforge.org/mailman/listinfo/rake-devel > _______________________________________________ > Rake-devel mailing list > Rake-devel at rubyforge.org > http://rubyforge.org/mailman/listinfo/rake-devel >