Paul Wise writes:
On Tue, Apr 16, 2013 at 4:08 AM, Florian Fuchs wrote:
This could be something for a HTML(5)/JS/CSS savvy person (to make it *really* convenient). I'm thinking scrollable log views, that interactively append/prepend entries to the view, depending on your position in the log file. Or commenting on log snippets and saving them for further review.
As long these are based on progressive enhancement, sounds good.
I think "no thanks to interactive convenience", unless the candidate is *really* good at Zen Garden stuff (ie, won't need to spend any time thinking about progressive enhancement, let alone learning any CSS or JS). Convenience can only be realized if logs can be read and effective queries are easy for users to generate, and the replies are easy to understand, too. Let's leave Web2.0 stuff for next summer, please.
Remember, one of the top reasons for doing this is so that the admin can forward the log(s) to mailman-users for analysis. Besides interactive access, I'd also like to be able to script (curl, wget) the queries, so as far as I'm concerned the only web design needed for the reply is
text = slurp_log(log, filter)
html = """\<html> <head> <title>Log: {0}</title> </head> <body> <h1>Log: {0}</h1> <pre> {1} </pre> </body> </html> """.format(str(filter), test)
<wink/>
The *hard* and *interesting* part to me is the query API (ie, what the REST looks like, and that it be flexible enough to support enhanced queries in the future) and UI (what you'll use from Postorius).
Even getting access to the files is going to be non-trivial if we go after MTA logs too. We're going to need to convince admins that Mailman should be allowed to read those logs, so we're going to need to think about preventing users from writing queries that allow them to find out who's getting mail from pr0N sites and Kim Jong Un. I'd also like the code to do the MTA message ID threading to get the whole MTA history of a message.
Steve