[pacman-dev] Maelstrom(Pacman in C#) project brief

Bozhidar Batsov lordbad at e-card.bg
Mon Jul 23 10:32:48 EDT 2007


Any feedback is most welcome and I read it all the way no matter how 
long it is. BTW I personally enjoy python very much, but I'm a hardcore 
GNOME user and dislike QT(although I confess that KDE is the better DE 
for the moment). I'm just so used to GTK apps that I can't look at QT 
apps for a long time without my head starting to hurt. It seems to me 
that most people pay more attention to Whirlwind, but I want to assure 
all the command-line addicts(like me) that maelstrom will have a very 
capable console client. Actually I work on it far more than on the 
gui(for now at least).

Best Regards,
Bozhidar Batsov

Georg Grabler wrote:
> So go for it.
> I bet a couple of users will like it, and use it, as they use other 
> projects. I personally also would like it, for the simple reason that 
> i'd have it easier to get people to use arch.
>
> I'm thinking about a friend of mine (as well as some guys at work 
> here). He plays one computer game, which has a native linux 
> distribution of the software, which runs perfectly on arch. He'd like 
> to use Linux on his desktop. He's the typical office user (MSN, ICQ, 
> Office, Printing, E-Mail and this Game).
>
> Thus, he is a typical windows user, and telling him how the 
> application works is by far easier than giving him the hard truth 
> about linux and commandline.
>
> I'll hopefully soon be working again on pyalpm and a pyQT client, 
> hopefully Cx will help me some (or i can help them with their pyQT 
> client project). Havn't had much contact, maybe he's lost in the few 
> lines i wrote, since a lot of them are replaceable by new features the 
> alpm gained in the past month.
>
> I'd never say that i wouldn't use the project myself. Maybe i like it 
> in the end, and to be true, i already liked the interface you provided 
> (screenshots).
>
> I just want to tell you, of course, a native one always was "cleaner" 
> "nicer" and "sexy", but it definitely re invents the wheel, for the 
> reason that the code will be written twice. Also, maybe you come up 
> with ideas how things could be implemented in pacman, since you seem 
> to have plenty experience, also in c/c++ development.
>
> I've chosen pyrex for the same reason as you the native re-write: i 
> didn't like the procedural API interface of alpm - but it gets better 
> every day. Even when i'm currently not working on my projects, i 
> follow the development lists carefully.
>
> If you want that project - do it. As i said, a lot of people will 
> appreciate it.
>
> Also, why i considered this memory leak things is, that i may would 
> like some background application searching for updates, providing a 
> systray entry if updates are available (as other tools do). Though, 
> this would need a different / faster / less system consuming operation 
> for checking updates (as taking the provided rss feeds), but it would 
> be possible, and a very nice feature.
>
> Yours,
> Georg
>
> PS: i know i write too much, hope you even want to read it ;)
>
> On 7/23/07, *Bozhidar Batsov* < lordbad at e-card.bg 
> <mailto:lordbad at e-card.bg>> wrote:
>
>     Georg Grabler wrote:
>     > That's a fine point, Gabriel.
>     > I just thought the same about my servers, by reading your message.
>     >
>     > Especially virtual servers it would take enough space then.
>     >
>     > Arch isn't a focused desktop distribution. Anyway, the team is
>     working
>     > hard to make pacman / alpm better (providing a better interface). In
>     > the current state, they work towards to get by default
>     compatible swig
>     > bindings (at least the approach seems to be quite clear and clean,
>     > following the git), what means you can also build your application
>     > upon the alpm binding for C#.
>     >
>     > I also think that re-inventing the wheel isn't what should be done,
>     > especially not in such a "small" project. Of course, desktop
>     > environments and applications also often re-invent features.
>     >
>     > What i really appreciate are the features your client seems to have.
>     > That's a huge piece of work you've done, and basically what i'd
>     like
>     > to see for pacman. GTK/QT interfaces for the original ALPM, for
>     every
>     > kind of desktop user.
>     >
>     > I've been working some time now on pyalpm to make a compatible
>     client.
>     > I just don't get along, since my work still eats me up, and the
>     > critsit won't end, as it seems.
>     >
>     > Yours,
>     > Georg
>     >
>     > On 7/23/07, *Gabriel C* <nix.or.die at googlemail.com
>     <mailto:nix.or.die at googlemail.com>
>     > <mailto: nix.or.die at googlemail.com
>     <mailto:nix.or.die at googlemail.com>>> wrote:
>     >
>     >     Bozhidar Batsov wrote:
>     >     > VMiklos wrote:
>     >     >> Hello,
>     >     >>
>     >     >> Na Fri, Jul 20, 2007 at 05:40:40PM +0300, Bozhidar Batsov
>     >     <lordbad at e-card.bg <mailto:lordbad at e-card.bg>
>     <mailto:lordbad at e-card.bg <mailto:lordbad at e-card.bg>>> pisal(a):
>     >     >>
>     >     >>> Well I have no intention to fork Arch. I strive to make a
>     >     product 100%
>     >     >>> compatible with the existing, but I want to offer
>     through it a few
>     >     >>> things that are missing in pacman and a couple of newer
>     >     technologies. I
>     >     >>> don't think that mono is a bad thing just because .NET is a
>     >     Microsoft
>     >     >>> product. After all Miguel de Icaza has stated many times
>     that
>     >     if he had
>     >     >>> mono 8 years ago there wouldn't be one line of C code in
>     GNOME. I
>     >     >>> personally consider it to be a much better framework than
>     >     java. Style
>     >     >>> and consistency are almost perfect here. Pascal notation
>     for
>     >     methods,
>     >     >>> camel for vars, great generics, great datatypes, security...
>     >     >>>
>     >     >> let's say you would write this in python or perl, we
>     would have
>     >     the same
>     >     >> problem: pacman is a lowlevel tool, it should be fast and
>     have
>     >     as less
>     >     >> deps as possible. mono can be a great tool but are you sure
>     >     it's nice to
>     >     >> have the whole mono framework in an install cd?
>     >     >>
>     >     >> - VMiklos
>     >     >>
>     >
>     >     >>
>     >     > Not exactly so. For one thing there are many popular programs
>     >     build with
>     >     > mono today - beagle, f-spot, tomboy, muine and others so
>     you are
>     >     likely
>     >     > having the framework installed anyway. There are talks
>     that soon
>     >     mono
>     >     > will be accepted as an official dependency in GNOME.
>     >
>     >     All the programs you are talking about are almost broken but is
>     >     not the point here.
>     >
>     >     You think to much about 'installations with an DESKTOP/X'
>     that is.
>     >
>     >     *Why* do you think I would install _mono_ ( about 70MB bloat
>     ) on
>     >     my _servers_ ?
>     >
>     >     Just to run an PM ?
>     >
>     >
>     >     Gabriel
>     >
>     >     _______________________________________________
>     >     pacman-dev mailing list
>     >     pacman-dev at archlinux.org <mailto:pacman-dev at archlinux.org>
>     <mailto:pacman-dev at archlinux.org <mailto:pacman-dev at archlinux.org>>
>     >     http://archlinux.org/mailman/listinfo/pacman-dev
>     >     <http://archlinux.org/mailman/listinfo/pacman-dev>
>     >
>     >
>     >
>     ------------------------------------------------------------------------
>
>     >
>     > _______________________________________________
>     > pacman-dev mailing list
>     > pacman-dev at archlinux.org <mailto:pacman-dev at archlinux.org>
>     > http://archlinux.org/mailman/listinfo/pacman-dev
>     >
>     I don't see how 70MB are so crucial, but everyone has different
>     opinion
>     on every subject. I have 370GB HDD on my home PC, and don't
>     imagine that
>     there are many PC's left with 5GB drives around so 70MB doesn't
>     seen so
>     much. As I previously stated bindings do not reflect the true
>     nature of
>     C# and produce fairly ugly code, so I prefer to do stuff natively. On
>     the subject of possible memory leaks on behalf of mono, I'd like
>     to say
>     that this is not an app that runs 24/7(Maelstrom or Whirlwind). A
>     typical session is over for a couple of minutes so there is little
>     danger of leaks as they tend to manifest in longer running apps.
>     And of
>     course the goal of the project is not to reinvent the wheel, but to
>     build a better wheel so out car could run faster and be equiped with
>     more features - for example maelstrom has support for package
>     categories, something which seems to be ignored in arch - they are
>     implemented with filtering regular expressions and are very helpful
>     especially for a GUI client. Color support in the terminal is
>     native, no
>     patches required, and the conf file is scheduled to be enhanced as
>     well
>     with additional options... So these huge 70MB may not go to waste
>     completely in the end...
>
>     _______________________________________________
>     pacman-dev mailing list
>     pacman-dev at archlinux.org <mailto:pacman-dev at archlinux.org>
>     http://archlinux.org/mailman/listinfo/pacman-dev
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> pacman-dev mailing list
> pacman-dev at archlinux.org
> http://archlinux.org/mailman/listinfo/pacman-dev
>   





More information about the pacman-dev mailing list