Re: [giFTcurs-devel] Auto-search-more idea

[Starling]

Christian Häggström <address@hidden> writes:

> What do others think? Maybe this question should be directed to the
> plugin writers?

Not much plugin writers can do about the Fasttrack network: once you
send a query to another server, you can't make that server work any
more intelligently, most servers don't.  Same goes for Gnutella.  
And heaven forbid you submit a feature request to OpenFT, they can get
kind of nasty.  All of the plugins "give up" on a search anyway,
returning with a message on how many results were found.  There is no
way to determine however, whether the network really does not contain
those files, or whether you just haven't reached them yet.  As well,
the files that match a search query could be offered to the network at
any time in the future.

Therefore I agree that for certain search queries a plugin should
never give up unless explicitly asked to do so, but would probably
meet with swift resistence if I suggested it to the community.  It
would also mean a change in the giFT protocol, sending a
'auto-search-more' or not flag in the initial search packet.  Just
seemed to me like the burden of initiating and recording a search was
deliberately placed in the client's hands.

> Yes, I had problem separate it from the auto-search stuff. I will
> rewrite it myself with help from your patch. The documentation was good
> however. It will probably go in as it stands.

Thank you for looking at my code!  (The gboolean pointer return was
indeed a typo.)  Feel free to use it however you like; I'm just happy
to have a nice filesharing client that works on Linux.  As for my real
name, I'd rather not give it up, go ahead and omit me from any AUTHORS
files.  At any rate, if all I'm committing is that little "Delete
search" option I added as an afterthought, it's no matter: I give you
full rights to copy, modify, rewrite and incorporate it.


Starling