Thanks Hamish, I suspected there would be good reasons.<br>Before getting this email, I've gone through and replaced all "wget" with "wget -c". Feel free to change back if you think it needed. If not, I'll probably get to it eventually.<br>
<br>Of note, I'm keeping a local copy of the /tmp directory for quick synchronising and rebuilding the scripts.<br><br><div class="gmail_quote">On Thu, Sep 10, 2009 at 7:44 AM, Hamish <span dir="ltr"><<a href="mailto:hamish_b@yahoo.com">hamish_b@yahoo.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div class="im">Cameron wrote:<br>
> I notice that you have used "wget -c" and "wget -r" in different<br>
> places in the install scripts.<br>
<br>
</div>Right, this is intentional.<br>
<br>
-c to continue aborted download instead of starting over as file.1,<br>
file.2, file.3, etc. This has been used for large downloads (10s of MB)<br>
which are not expected to change. i.e. versioned sample datasets. This<br>
is not safe for files which may change on the server, it simply compares<br>
the number of bytes to decide if it has the whole file. If the file is<br>
already downloaded wget will say that and then exit with a EXIT_SUCCESS<br>
exit status.<br>
<br>
<br>
-r is really for recursive mode, but when used on a single file it will<br>
replace the existing one. This is also to help the scripts be a bit more<br>
robust to the"file.1" vs. "file" problem, especially for files downloaded<br>
directly into the filesystem & not the build dir.<br>
<br>
I've used this for downloading patches and other small text files which<br>
may change with time. When -r is used in this way --output-document= or<br>
-O must be used as well, otherwise the recursive side of it tries to<br>
reconstruct the dir structure. Maybe a "rm" before the wget could ensure<br>
that the downloaded version is fresh without needing -r. Shrug.<br>
<br>
<br>
Finally for trivial downloads I don't expect to change (icons) I used<br>
"wget -nv" which is non-verbose mode. Using that for a long download<br>
could give the impression things have stalled, ^C, then file.1, file.2,<br>
cruft. "wget --quiet" masks error messages so is to be avoided.<br>
<br>
<br>
I think wget already checks if stdout is a tty before deciding to use<br>
=====> vs. |.....|.....|.... style progress meters. (ie logfile friendly)<br>
Otherwise you could switch that on by hand as well.<br>
<div class="im"><br>
<br>
> Should we go through the scripts and convert all "wget" to "wget -c" ?<br>
<br>
</div>I don't think so; one size doesn't fit all..<br>
<br>
as long as everyone builds and downloads to /tmp/ dirs and cleans up<br>
every time, the flags will be mostly redundant overkill.<br>
<br>
I'm happy with ad hoc and folks using or not using those flags as they<br>
see fit.<br>
<br>
<br>
regards,<br>
Hamish<br>
<br>
<br>
<br>
<br>
_______________________________________________<br>
Live-demo mailing list<br>
<a href="mailto:Live-demo@lists.osgeo.org">Live-demo@lists.osgeo.org</a><br>
<a href="http://lists.osgeo.org/mailman/listinfo/live-demo" target="_blank">http://lists.osgeo.org/mailman/listinfo/live-demo</a><br>
</blockquote></div><br><br clear="all"><br>-- <br>Cameron Shorter<br>Geospatial Systems Architect<br>Tel: +61 (0)2 8570 5050<br>Mob: +61 (0)419 142 254<br><br>Think Globally, Fix Locally<br>Commercial Support for Geospatial Open Source Solutions<br>
<a href="http://www.lisasoft.com/LISAsoft/SupportedProducts.html">http://www.lisasoft.com/LISAsoft/SupportedProducts.html</a><br>