I must say that I got interested in command line tools in the build process quite recently after having worked for the past 5 months on a j2ee project living and dogfooding latest eclipse wtp builds (I'm a .net-first-person, don't worry). There were things
that just simply weren't in the IDE and it seemed natural to automate things outside the IDE. As ant was already integrated inside eclipse I decided I'd give it a go and although some results were easy to do other were obscure and strange, also the whole extension
model to me is still incompehensive, why not choosing namespaces and a schema is a mistery to me.
Anyhow I had this thing in my faculty to use maven and as I'm one of those get-on-the-bandwagon-first people I decided to go with maven 2 that just got released. Wow, an eye-opener. Although still rough around the edges it gives a plugin-oriented way to
manage the whole project lifecycle, all from a command line.
This brings me to my question. In the begining people were saying msbuild is like ant for .net, is it so or does it cover/will it ever cover aspects that maven/2 handles or will that be entering too much into VSTS terrytory?
I really like MSH and I think it's on the right path to create a really nice programming and management enviroment. However what left me shocked with this video is how much weight did Snover loose I should do some workout myself I think...
If you'd like a thread featured way to communicate there is the ms forum betanews.microsoft.com or you could open one on google groups. Great project, I hope you produce something that we can install and play with (perhaps shared source initiative). Also can you envision a device-driver oriented language?
This sounds as a great research project and it would be great if they'd open it more to the public. Of course having a Monad port for it would be a great benefit for the developers team. I actually dreamed of a thing like this and these guys are
having all the fun doing it I'd have also a question, for instance Apple is doing their drivers development in a
constrained C++. Don't you think it would be nice to have a special purpose language that would be great for device driver writing? That would allow to write the code once and then achieve different machine code in function of the compiler switches or different
compilers. Such code should be more verifiable, perhaps more mantainable in time, have better productivity and a smaller learning curve. Or not? Does it sound convenient or is c,c++ a good language choice anyway?
BTW, I'm talking about release here though, not devel,
I consider myself a developer (that's why I love C9 btw). I have used both rpm and deb based distro's and I probably don't have anything major against them. Except that after x number of months packages stop getting released as often and or time (dev's need
packages yersterday), or packages get built on other glibc or some other library and you start upgrading, sometimes even uncompatible packages and you quickly end up with a unusable system. Then you have to start from scratch. In a way binary based linux is
much more vulnerable to dll hell. In the past I'd be using a GNU/Linux distro for about 4-6 months,then format and repeat with distro x+1. I ran a gentoo distro for a year and a half flawlessly with all init things going OK something that was a real problem
with RedHat (mileage may vary depending on hardware ),even after various updates. I would have had it still if I had more space on that disk.
Package systems are good solutions but aren't free of problems, often are less then flexible (ex. trees compiled for certain architectures and/or not taking all advantages of your hardware). Plus you have to have developers that take care of each tree (think
about commercial vendors that have to support >1 distros, it's a nightmare). And I'm not bashing rpm in any way don't misunderstand me, I liked it when I installed my first RH5, I liked it when I discovered src.rpm's but it's still a feat to keep a system
Good or bad linux does one thing well, it's the GNU toolchain. Almost every linux program gets built with it. If you nail the toolchain you can nail most other problems. I have not seen any other linux distro (or did not care to see at the
time) except gentoo that has such a tight coupling between their tools and development process, a remarkable example in software engineering. Perhaps you could
in extremis distcc binary packages, but then again, the computation cost, the various architectures, use flags, etc would probably make it unmantainable even if automated.
This is a strange discussion to be having on a MS forum , so... let's start fighting back and convince the world and all flamers that MS is commited in doing really great software by investing in .Net on all platforms (linux included, btw is Rotor
My personal wish that I suggested on the MSH newsgroups, why not have MSH on linux as well ?
except our version of windows update will allow us to install pretty much any software with signed packages from secure repositories.
Imagine if windows update also let you download and install a signed cab with adobe photoshop from a secure source, well up2date gimp. Downloads the package and the dependancies. I have extra repos in my sources as do most linux users, dag, freshrpms, fedora-extras,
what have you.
You can do the same with yum or mdk's urpmi.
Of course you all guys know that the best combination is having all the GNU goodness, backed up by gentoo build system and running on the NT kernel and MSH as shell?
And it exists too
Haven't seen the interview yet but I hope that you have gentoo in your labs.
Question does MS have their C,C++ on the linux platform? I bet there are some gentoo'ers (like me) that would like their platform to compile a bit faster
The licencing interview is a real jewl in economies behind software developing. One thing that was clear is that investment has to be recuperated. I think that the investment will be recuperated now or never as I belive this kind of integration will become focus even for free projects. Projects like Eclipse could bring products equivalent to Rational's suite or TS in rather limited amount of time. Many
tools already exist and there's always Ant, Maven, Subversion, Bugzilla, trac, xUnit, ... On the other hand I don't belive MS can loose the investment as they actually made a system that will help them build better software and it's MS's best pay-off.
PS. My own personal question, if I have a versioned project in SVN, how is migration done? Can I keep SVN (as I like it a lot, and have a lot of history in it)? I attended a TS event and the MS guy said that many ISV's are developing plug-ins for the TS studio,
is there some group that could help me use SVN? Doesn't MS think that plug-in based developing made eclipse what it is today (not counting the big companies behind it) and effort should be made to make it more mainstream? That way, we can push the lower bount of the market penetration even further as development
would/could/should happen without cost to MS. All that said, I'm not a Java freak There is a one more thing (but this is just off the wall), what would happen if you set up a TS farm site? (read sourceforge) That way people would not have to buy the server and still be happy with the clients, not to mention the proof of scalability that you'd like to get. Just for Open Source projects of course (or closed if they pay)
I actually heard one guy swore "we really screwed up.." (the 1.0 framework). Anyhow, it seems a part of the video got cut off when he said that "people always enjoy this story". Yup, we take mental notes so whach out what you say
I think that I revised my oppinion on what would the best customers be for this kind of equipment. Graphics are great but you
SHOULD think about the >1'000'000'000 Chinese and the rest of the world that can't have keyboards for all their characters.