In the shell script, you would add -nc to the second wget command. BTW, http protocol doesn't support a directory listing. Where you see these from a browser it is the web server making an index of the directory and presenting it as html. That is expected and proper.
On Tue, Nov 18, 2014 at 10:25 AM, Larry Brigman <[email protected]> wrote: > You didn't miss understand -nc. You missed the -O option where that > expects the > next argument to be the output file retrieved from the URL. > > > On Tue, Nov 18, 2014 at 10:11 AM, John Jason Jordan <[email protected]> > wrote: > >> At the Clinic Wes composed a lovely command (that I converted to a >> shell script) to download files from a Flashair SD+wifi card that lives >> in a CPAP machine upstairs in my house. The connection is a bit wobbly >> because the advertised range of the Flashair is 30 feet, and that is >> about the distance from the machine to my laptop downstairs, plus there >> is a floor in between. But it does work if I am patient. Here is the >> command: >> >> wget -qO - http://192.168.0.1/DATALOG | grep 'DATALOG.*fname' | sed >> -e "s/^.*fname\"\:\"//" -e "s/\", \"fsize.*//" | while read line; do >> wget "http://192.168.0.1/DATALOG/$line";done >> >> The reason for the grep and filename search is that Toshiba, in its >> alleged wisdom, decided to bury the files in the html rather than just >> use a directory structure like ftp. >> >> But there is a problem: Every night the CPAP machine writes eight more >> small files to the DATALOG folder, where the name of each file starts >> with the date, e.g.: >> >> 20141116_235932_BRP.crc (from Sunday night) >> 20141117_235932_BRP.crc (from last night) >> >> The normal behavior of wget is to re-download files already downloaded >> and append .n to the additional copies, so running the script this >> morning would result in the following files in the folder: >> >> 20141116_235932_BRP.crc (from Sunday night) >> 20141116_235932_BRP.crc.1 (from Sunday night) >> 20141117_235932_BRP.crc (from last night) >> >> After a while the folder where I store these files on my computer is >> going to get horribly cluttered. I could add a line to the script to >> delete all files ending in .1, but considering the time it takes for >> the script to run (due to the poor connection) it would be far >> preferable for wget not to download copies in the first place. >> According to the man page adding -nc (no-clobber) is supposed to do >> this. So I added -nc like this: >> >> wget -qO -nc - http://192.168.0.1/DATALOG ... >> >> When I ran the script it downloaded no new files, but created a file >> 'nc' of zero bytes. WTH? Clearly I have failed to grasp how the -nc >> option is supposed to work. >> >> Are there any wget experts here who can lead me to the light? >> _______________________________________________ >> PLUG mailing list >> [email protected] >> http://lists.pdxlinux.org/mailman/listinfo/plug >> > > _______________________________________________ PLUG mailing list [email protected] http://lists.pdxlinux.org/mailman/listinfo/plug
