ioFTPD Next

[darkone]

Time to list plans for next versions.

- New core
- NT Service
- Unicode support
- Rewritten module support
- PHP trigger for scripts (EXEC/TCL/PHP)

Major update(1):
- Dynamic user and group structures?
- AD module with public source codes
- CGI capability to HTTPd
- VHOST capability to HTTPd

Major update(2)
- Add missing HTTPd functionality
- Rename daemon?

...... we will release some test versions before the releases.

[odd]

Sounds good.

Nice to hear your working on the ioftpd again

[icefuzion]

NT Service support nice..

Thanks for updating us. Were all looking forward to the speed of the new core as well.

[durulink]

*Unicode support*

GOOooooooooooooooooooooD~~ ^^

[Pu$u]

jo the new version would be very cool (i'm looking forward for it) , but wich limits you want for the unregistred version ?

[davibou]

Nice work dude !!!!
Yeah Rename daemon !!!!

[SnypeTEST]

- PHP trigger for scripts (EXEC/TCL/PHP)

I less than 3 you.

[Stardog]

If you rename the deamon, then we have to rename all the scripts.

[darkone]

Originally posted by jron
Also I hope you don't over extend yourself by adding full on HTTPd support, like when you took months to add HTTP features that few people use =( This is YOUR project no doubt, but we are all in love with the FTP so, FTP features would be nice. Raiden is always adding really nice ideas to their FTPd they are just poorly implemented, its a shame really. =(

Moving to httpd doesn't take resources off the ftpd developement. Current httpd code is ~1200lines & took less than a week to write and optimize (prior version of httpd took much longer to do, as it had stream mime parser - which I learnt later on, isn't even a task that httpd should do => and which resulted to the rewrite). What I did spend some time on, was php module. However, it was done so, that it's possible to use it in the future with both ftp & telnet.

In the future, just like in the past, I will not focus primarily on adding features. ioFTPD wasn't ever designed to include all the whistles & bells - I think you all should know it by now.

Yes, there's several good things in raiden that are currently missing from io. First few that come to my mind are: Database (ODBC) support, NT service support, Unicode support, MLST.

However, I do have to set certain priorities for tasks. It makes no sense to add database support, before I've added dynamic user/groupfiles & unicode support. (with dynamic user/groupfile, I mean that it has database entry where modules can store data) And again, there is no point of adding dynamic user/group structures before unicode support is in. Which leads to why I'm writing the new core now: imo there was no point of rewriting socket readline function (~150lines) for the old core to support utf-8. (Point of rewriting is to make something better, not to make something bloated)

[darkone]

Btw, I'm working on SSL part now (designing the data structure). Hopefully tomorrow, I can get to do some actual code again

After SSL I'm going to work on speedlimits. (propably late this week / early next week) Then comes the module support & unicode... and finally php.

[darkone]

SSL read, and accept functions have now been implement. Read is not yet fully optimized, but good enough for now. I'll propably have connect & write routines done by end of week, and I can then give some performance figures on loopback transfer.

I expect it to be able to max nearly 100mb/sec. Transfering 32mb file and parsing new lines from it took nearly 5seconds & consumed less than 7% of cpu time of one processor (according to task manager, which isn't really that reliable )

[darkone]

Another update.. I'm now working on adding throttling, algorithm is all new. It's more efficient, more accurate and allows bandwidth limits on harddrives (which could provide to be useful on slow network shares, cd-rom/dvd, etc.).

[Luke]

hi d1

any eta when you gonna let us smell/feel the new ver ?

[neoxed]

When it's done?

[darkone]

Right, I want to do everything right before posting next release. This means, that there will be several internal testing phases before the actual release (however I'm quite sure, that my code is stable I think I've been playing with pointers long enough to produce quality code when sleeping)