kris.c

I have written client/server progs to transfer files . Through this program, I am now able to transfer files larger than 1 MB from my machine to any remote machine where the client program is running without any problem..


The other requirement of the program is to read a file that exists in hte remote machine.. I have used the same concept even for that code and i have traced it many times to find any pit-falls, but I am not able to transfer files greater than 5000 bytes without the data being curropted..
I also noted these things..
1 . the data in the packets that is being sent from hte remote machine is valid, thats not the case after I receive them here for the final few packets
2. also, the numbers returned by send() and recv() match for a large number of packets sent , the mismatch occurs in the last few packets..
By curropted ,I mean some sections of the file getting repeated right at the end , or after the actual file has been transfered, a part of the file also gets added.
The lan is a standard one, I can do my usual ftp and telnet stuff.

I cud have posted the code , but its quite bulky..
Has anyone faced this problem before, please help

kris.c

i would like to add one more thing here :
I was instructed break up large files into chunks of 1000 bytes and then use the send() iteratively ..
only in the in the last packet, the send() and recv() functions return different values..

Salem

Deja-vu again and again?
send() and recv() functions
file transfer across network

1. Are you using a TCP or UDP connection?

2. Do you send the length of the file first, then the contents of the file?

__________________
If you dance barefoot on the broken glass of undefined behaviour, you've got to expect the occasional cut.

kris.c

forget about that ,i cud fix it..

Hunter2

If the answer to (2) is no, perhaps you're outputting the receive-buffer before checking the error status of recv(). Or, maybe you're assuming that (packet_size) bytes have been received instead of checking recv()'s return value. That would lead to (in the first case) the last packet being repeated, or (in the second case) a segment of the second-last packet being repeated.

*EDIT: Actually, the second case would lead to random unpredictable bits of data strewn all throughout your output, and a predictable segment of the second-last packet being appended to the last.

Judging from the previous threads, this is *probably* the problem.

*EDIT2: Actually, the 'predictable' segment will *not* occur if:
(filesize % buffersize) == (unpredictable_bytes % buffersize).
See if you can figure out why.

__________________
Just Google It. √

(\ /)
( . .)
c(")(") This is bunny. Copy and paste bunny into your signature to help him gain world domination.