April 09, 2004
Website monitoring?
I guess I might post this on LazyWeb, or just take the time and write the tool, but figured I'd ask first :)
There have been many times in the past few years when I've wished for a simple tool that could help me in monitoring the progress of software projects that I am interested in - and particularly, the release of new versions of tools that I maintain FreeBSD ports for. Of course, some projects have announcement mailing lists, others have the good grace to provide RSS feeds... but there are still some that are not quite up to snuff, and that require periodic polling with my browser.
So.. does anybody know of a tool that does something like this:
- fetch the contents of a list of web pages;
- grep the contents for a specific regexp (something like 'The latest version is (\w+)';
- compare the result with the last-stored one (this will cover the failure modes: changed text or even moved webpage);
- issue an alert for those that do not match;
- let me acknowledge the change and store the new version as the compare base when I've seen the alert and dealt with the new version.
Note that this does not compare the whole webpage against the last-stored value - I am not interested in changed layout or added links or news items or similar. The basic premise here is that the comparison is done on only the part of the webpage that interests me.
Also, 'webpage' could easily be extended to 'distribution site' - tools like wget and curl somewhat blur the distinction between HTTP and FTP as transfer media, and there are distribution sites which have dummy files named LATEST-VERSION-IS-1.0.15 for just this purpose.
Yes, I know it is trivial to write such a tool, but I've still not done it, since to write it properly would take about an hour or two.. and I *am* a procrastinator ;) Still... has anybody heard of anything similar that I could use ready-made? :)
Posted by roam at April 9, 2004 01:39 PM
The TrackBack URL for this entry is http://extelligence.ringlet.net/cgi-bin/mt/mt-tb.cgi/84
