IIRC, wget is for pulling files from remote sites. There's a utility wput that does the reverse. This is available from sourceforge & is part of the cygwin suite.
Bruce wput man page http://wput.sourceforge.net/wput.1.html DESCRIPTION Wput is a free utility that is able to upload files to a ftp-server. Wput is non-interactive and background-capable. It can upload files or whole directories and is meant to be a robust client even for unstable connections and will therefore retry to upload a file, if the connection broke. Wput supports resuming, so it automatically continues uploading from the point where the previous upload stopped, meaning that you can kill Wput anytime and it will (if the remote ftp-server supports this, being most likely the case) finish the partial uploaded file. Wput supports connections through proxies, allowing you to use it in an environment that can access the internet only via a proxy or to provide anonymity by hiding your ip-address to the server. For SOCKSv5-proxies Wput supports also listening mode, allowing you to use port-mode ftp through a proxy (useful if the remote ftp is behind a firewall or a gateway). Wput supports timestamping, so it will (in the ideal case and if timestamping is enabled) only upload files, that are newer than the remote-file. The upload-rate of Wput can be restricted, so that Wput won't eat all available bandwidth. URL-Input-Handling URLs are recognized by the ftp://-prefix Wput first reads the URLs from command-line, and associates the first file with the first URL, the second file with the second URL etc. It then transmits the file/URL combinations that are already complete. In situations where more URLs than files are specified, Wput tries to guess the local filename from the URL. Afterwards, Wput uses the --input-file (if any) and reads the URLs using the same sheme as above. If there are still files remaining, but no URL has been specified for them, Wput uses the last known URL for each of the files. So you can specify e.g. one URL and read all filenames from a file. Or use *wput *.txt ftp://host*, to transfer all *.txt-files. So *EXAMPLES* for further examples. To be on the safe side, it is recommended to supply the files before the URLs. Guessing Local File If Wput has an URL without a corresponding filename, Wput tries to guess the local file's location. e.g. using wput ftp://host/directory/path/file, Wput will look out for /directory/path/file. If not found, Wput looks for ./directory/path/file, ./path/file and ./file. OPTIONS *-V* *--version* Display the version of wput. *-h* *--help* Print a help screen, with a short description of wput's command-line options. *-b* *--background* Go to background immediately after startup. If no output file is given, wput will redirect its output to "./wputlog" *-o* *logfile* *--output-file=**logfile* Log all messages to *logfile*. *-a* *logfile* *--append-output=**logfile* Append all logged messages to *logfile*. *-q* *--quiet* Turn off Wput's output. *-v* *--verbose* Turn on verbose output. This gives some more information about what Wput does. If you specify this flag twice, you get debug output. *-nv* *--less-verbose* Be less verbose. That means reducing Wput's output to a minimun. Specifiing this flag more often is equal to the --quiet flag. Some people also like combining the -v and -nv flags, being quite senseless. *-i* *file* *--input-file=**file* Reads URLs and filenames from *file*. If there are URLs on the command-line too, these will be retrieved first, unless sorting is enabled. See also the URL-Input-Handling section. If *file* is -, the URLs will be read from stdin. If you want to pipe the contents of the file that shall be uploaded to stdin, this cannot be done (yet). But you can use the --input-pipe flag and read the contents a) from a named pipe -I "cat named.pipe; echo > /dev/null" or b) directly from the command, that outputs the data. (See --input-pipe) Do *not* do things like *find | wput ftp://host/ -i -*! Wput would upload all files from the current directory (since the first output of find will be '.') and afterwards each file again (since find postes its name to Wput. And further problematic is that Wput will upload each directory that is given by find and since find itself recurses all directories, the files would be uploaded three times (or even more often for further subdirectories). Use *wput ftp://host/* to upload everything from the local directory. Or use *find ! -type d | wput ftp://host/ -i -* to tell find, not to output directories. *-s* *--sort* If sorting is enabled Wput first reads all URLs from any input-devices available and will sort them before transmitting each file. The sorting order is: ip/hostname, port, username, password, directory, filename. Sorting requires a bit more memory since all data needs to be hold there. *--basename=**path* This option causes Wput to snip *path* from all input-files when they are connected to the URL. wput /usr/share/doc.tgz ftp://host/ would create ftp://host//usr/share/doc.tgz, whereas specifing /usr/share/ as basename will result in ftp://host/doc.tgz being created. *-I* *command* *--input-pipe=**command* If no file/directory can be "guessed" (see "Guessing Local File") from the URL, the output of *command* *is taken as file-input. command is invoked as follows:* command ftp "username" "ip/hostname" port "remote_directory" "remote_filename" The hostname is only supplied if the ip cannot be resolved. If you do not want these parameters to confuse the programm from which you read the contents, use something like '-I "cat file; echo > /dev/null"' so that these parameters are passed to echo and to /dev/null afterwards. Since the progressbar is not capable of handling unknown filesizes, the filesize is set to 1 GiB. Therefore the ETA shows a wrong value. *-R* *--remove-source-files* Unlinks/deletes files that have been successfully transmitted. *--bind-address=* *ADDRESS* When making client TCP/IP connections, bind() to *ADDRESS* on the local machine. *ADDRESS* may be specified as a hostname or IP address. This option can be useful if your machine is bound to multiple IPs. *-t* *number* *--tries=**number* Set number of retries to *number*. Specify -1 for infinite retrying. The default is to retry 3 times, with some exceptions on fatal errors. *-nc* *--dont-continue* If this flag is specified, resuming will be turned off, meaning that a remote file being smaller than the local one will be overwritten. To skip this file, you have to enable --skip-existing. See also * USAGE.resumehandling* *-u* *--reupload* If this flag is specified, a remote file having the same size as the local one is to be uploaded. Skipping is default. *--skip-larger* If this flag is specified, a remote file being larger than the local one will be skipped. Default is reuploading it. *--skip-existing* If this flag is specified, the upload of a file will be skipped if the remote file already exists. *-N* *--timestamping* If timestamping is enabled, Wput will retrieve a directory list and parse it to determine the remote file-date. If the local file is newer than the remote one (there is a default allowed timevariance of 5 seconds, which can be adjusted in the *wputrc*-file) it is uploaded, otherwise skipped. The local date is dermined by the mtime (time of last modification), using the current time-zone. This should be equal to the output of ls -l. Since you usually do not want to resume existing files, you should employ the --reupload --dont-continue flags as well. *-l* *RATE* *--limit-rate=**RATE* If you don't want Wput to eat up all available bandwidth, specify this flag. *RATE* is a numeric value. The units 'K' (for KiB) and 'M' (for MiB) are understood. The upload rate is limited on average, meaning that if you limit the rate to 10K and Wput was just able to send with 5K for the first seconds, it will send (if possible) afterwards more than 10K until the average rate of 10K is fulfilled. *--no-directories* If Wput is unable to CWD into a directory, it will try to create it. If this is not the desired behaviour specify this flag to force Wput not to create any directories. *-Y* *MODE* *--proxy=**MODE* MODE can be either *http* for http-based proxies (such as SQUID), *socks*for SOCKSv5 proxies or *off* to disable the proxy. *--proxy-user=**NAME* If the proxy-server requires authentication, use *NAME* as user-name. You need to specify --proxy-pass too. These information can also be stored in the wputrc-file. *--proxy-pass=**PASS* Specifies the password to use for the proxy. *-p* *--port-mode* Per default, Wput uses passive mode ftp, which works well for most configurations. If passive mode fails, Wput automatically falls back to port mode. If you want Wput to start using port mode ftp, specify this flag. *-A* *--ascii* Wput automatically determines which transfer-format to use, by looking at the file-extensions. Certain files are recognized as ASCII. These are: txt, c, java, cpp, sh, f, f90, f77, f95, bas", pro, csh, ksh, conf, htm, html, php, pl, cgi, inf, js, asp, bat, cfm, css, dhtml, diz, h, hpp, ini, mak, nfo, shtml, shtm, tcl, pas Specifying this flag forces Wput to use ASCII mode file transfers. *-B* *--binary* Specifiing this flag forces Wput to use BINARY mode file transfers. *--force-tls* If this flag is specified and Wput is linked with the OpenSSL-library, the flag enforces the usage of TLS: If no TLS-connection can be established the process will cancel and not try to go on with an unencrypted connection. DIAGNOSTICS Normally, the exit status is 0 if either everything went fine or there was nothing to do. If some files were skipped during the upload (due to timestamping or resume-rules) the exit status is set to 1. If some files failed to be transmitted due to an remote error, exit status is 2. If some files failed and some others were skipped, exit status is 3. For general problems like failure of some system-functions the exit status is 4. BUGS You are welcome to send bug reports and suggestions about Wput to the sourceforge.net Bugtracker on sf.net/projects/wput Please send all available information that might concern this bug (e.g. Operating System and what can be done to reproduce the error). Supply also the debug-output (but remove confidential data if any), which helps a lot analysing the problem. If you use a wputrc file, it might also be useful to provide the relevant parts of it. If there is a crash due to a segfault or similar, try to run it in a debugger, e.g. gdb 'which wput' core and type where to get the backtrace. It would also be great help if you could recompile wput with memory-debugging support (make clean; make memdbg; [make install]) and use this debug-dump. SEE ALSO Many options can be set in a wputrc file. For its documentation consult the sample file provided by Wput. There are some USAGE.* files in the doc/ directory of Wput. These contain further information and samples on how to use Wput. AUTHOR Wput is written by Hagen Fritsch <[email protected]<fritsch%[email protected]> > 2009/11/25 ctx2002 <[email protected]> > curl , wget can do you job. > > i rember wget will automatically retry ftp connection, > curl probably has a retry option too. > > they all windows version. > > regards, > > anru > > On Nov 25, 12:18 am, Olwen Williams <[email protected]> wrote: > > I have a site which gets files transferred from an in-house application > > using a batch file and the windows built-in command line FTP several > > times a day. At least in theory it does, but for days on end this > > process can fail. The connection seems to be failing > > > > I was wondering if someone can recommend an ftp client which can be run > > unattended, initiated from a batch file in a windows environment, and is > > sensible enough to retry if it loses the connection part way through. A > > choice of active and passive may help as well. > > > > The PC support people are blaming the web host, and the webhost says > > there is nothing funny there. The problems may have started to occur > > when the server at the host failed, and the site was moved to another > > server. The hosting firm say that with the exception of php and apache > > which were differently configured that everything else is the same on > > the new server. > > -- > NZ PHP Users Group: http://groups.google.com/group/nzphpug > To post, send email to [email protected] > To unsubscribe, send email to > [email protected]<nzphpug%[email protected]> > -- Bruce Clement -- NZ PHP Users Group: http://groups.google.com/group/nzphpug To post, send email to [email protected] To unsubscribe, send email to [email protected]
