Meeting minutes 2010-09-19

[Agon]

For mono stuff to run on linux you need mono installed in system. But that should not be a problem, ubuntu for example comes with it installed already.

You need a bit more than just Mono. I know the Mono devs are proud to need less space than Java.
Also not all Linux distributions are supported. And some Linux guys don't like Mono on there system.

Also you need to check out how easy it is to install Mono on Mac OS.

And for a protocol, utility library you don't need to "depend on lots of other libs".

I guess all developers who don't use Mono in there application would prefer a C library.

[hoijui]

indeed i did not look at it, but as it is called PlasmaDownloader.dll, i guessed it only supports plasma (yeah i know, am crazy, sorry).

i suggested python in a command line driven approach, or a C library.
using a mono library is obviously not simple through C (which is what most languages support) on linux, compared to using a C library.

it is not that I do not want to cooperate with existing systems (remember, i am not maintaining a lobby). thing is, that other lobbies will not use your lib, if they are not windows only.
if you can prove me wrong (as in, all non-windows only lobbies want to use your lib), the problem is solved.
if not, we need an other solution, even if you think you have the perfect one already.

[Licho]

Well I just said that no existing lobby currently in use needs it.

For experimenting, its simple to just use mono one which already exists. Its also possible to extend Tobi's rapid one, but python implementation is unlikely to support all features of mono one, especially new file registration, because it would be complex to code/maintain.

For future lobbies, its of course best if its in portable native library because that has best direct interoperability.

I just think such lib is not in demand atm and its existence could slow down development cycle of the system.

[hoijui]

i see.. think i get you now
i just mentioned python as we already have rapid there.

the way a native lib would work in practice, in my eyes:
native lib is developed, and slowly brought up to par with eg your C# system, or SLs download system. if it is kind of stable, and SL team likes it (i guess, they would be heavily influencing the design decisions anyway), they would start to use it. if it gets pretty bug free on windows, SD and/or TASClient could use it too.
i imagine some engine devs and all lobby devs would have commit access to the main repo.

but yeah, if none of the existing lobby projects has interest in this idea at all, it would probably not be a good idea to even start with it.

[AF]

Are there examples of how to use it and an API doc anywhere we can see?

[Licho]

Outdated and source codes (like unitsync :)
If there is serious demand for it I can update it.

[abma]

just a few notes:

i startet to write the download-app in c++, because the current lobbys still can't download all files needed to play. Currently it's also difficult to explain someone, who is new to spring, to install the current engine + mod + map.
Installing the current engine is no problem, as there is an installer avaiable, but to install a mod + map to play it... there are many steps needed. I also wanted to improve my c++ skills

We need something like an one-click installer that installs the current top5 maps + mods and this has to be cross-platform and aviable as console-programm to (for the ai-ladder for example).

The biggest issue currently is, that springfiles isn't used to search for files, also the Plasma-Interface doesn't use all avaiable mirrors from springfiles. This often causes to fail the download of a map.

Also, TASClient + Springlobby doesn't use all avaiable download-systems (rapid-support for example is missing).

The best solution to solve this problems would be: create an easy-to-use search-api on springfiles.com. IMHO using the plasma-interface for downloading is easy (for example, gsoap can generate c-code), to use it for uploading is quiet difficult, as it has there many depends too.

To create an search-api i contacted already jj...

If you interested in what i already did: http://github.com/abma/pr-downloader/
Please keep in mind, that it was written quick and dirty + i had only c-skills at the beginning.