egghelp.org community

[whz]

Im working with three bots, two on different computers on the same network, and a third on a seperate network. The two on the same network are linked, sharing, and transferring the user file fine. When I link the bot from the seperate network to the hub on the other network, it links, but I get an error when they try to share&transfer the userfile.
This is the error I see from the bot on the seperate network

Downloading user file from hub
[bot2(dcc)] [16:24] Ending sharing with hub (Can't send userfile to you
(internal error)).

This is what I see from the Hub

[20:29] Sending user file send request to bot2
[20:30] BAD FILE -- can't send userfile

Transfer and share mods are loaded, flags are right, so Im not sure really whats going on

[TCL_no_TK]

what eggdrop version is the bot that you recived the message from? Please
_________________
TCL the misunderstood

[whz]

eggdrop v1.6.18

[whz]

The other thing it might be is the firewall on the two bots on the other network, both of their telnet ports are open and have no problems linking with the third bot on the other network, but do they use another port for transfering user files? The third bot on the seperate nework is on the dmz computer.

[nml375]

It would seem your hub is unable to create or open the copy of your userfile that's to be sent. Checked system quotas, write permissions, and such?

Userfile transfers are done on random ports (pretty much like a dcc file transfer - in fact, you can use the same settings to restrict which ports are used for the userfile transfer as with dcc's).
_________________
NML_375, idling at #eggdrop@IrcNET

[darthanubis]

https://bugs.launchpad.net/ubuntu/+source/eggdrop/+bug/220327

I not sure if this is related. Sound like whats happening to me, except, once the userfile can't be transferred the hub crashes HARD. Then after that, once it queries the nonhub bot, it crashes hard again.

This behavior is with the .18 ubuntu packages as well as the .19 I compiled myself with the ssl patch.

The text files of the outputs from the crashes as well as backtraces are there at the above link. I did not want to drop a massive paste here.

[Alchera]

Did you install eggdrop from source or did you use apt?

If installed via apt then start from scratch using source and the instructions on the main page.
_________________
Add [SOLVED] to the thread title if your issue has been.
Search | FAQ | RTM

[darthanubis]

[TCL_no_TK]

was the source also from ubuntu? since the reason i asked the version, was i couldn't find the given error message in eggdrop1.6.18 source to check why this happens. looks like this might be down to ubuntu's changes to eggdrop source code.
_________________
TCL the misunderstood

[nml375]

The error-message is present in the "standard" source..
It's raised whenever the userfile copy, that is to be "dcc'd", can't be opened. Or rather, whenever "raw_dcc_send(share_file, "*users", "(users)", share_file))" returns anything greater than 0 (in 1.6.18, see src/mod/share.mod/share.c:1913, and src/mod/transfer.mod/transfer.c:972/1241). More speciffic, DCCSEND_BADFN in this case.

That in turn is caused whenever the "fopen(filename, "r")" instruction in raw_dcc_send returns NULL. Unfortunately, the current code lacks means of getting any further information (some logdumping of errno would've been helpful).

My best suggestions would be to check tmp-paths, file system permissions, and such (are you starting it from the right directory, do you have write-permissions in current directory, etc)
_________________
NML_375, idling at #eggdrop@IrcNET

[darthanubis]

[darthanubis]