Logs of specification changes about ircv3-specifications HOT 9 OPEN

IMHO, 2 > 1 > 3 for the options.

I don't think public logging of the channel is that much of an issue - it's a public channel and the vast majority of things said in there is IRCv3 related.

from ircv3-specifications.

👍 for logging. Not sure what exactly you mean by grep though. Search box? Sure. Who needs plain txt these days? But then where is its “con” coming from?

from ircv3-specifications.

The idea for the second one was to do something like:

1. Log everything from #IRCv3 and then have a script that would do grep -A50 -B50 'term' where 'term' would be an issue.
2. Tag all of the lines returned with the 'term' that was grepped for
3. Have a website with all the 'terms' we have and allow users to click on them to read all the lines relating to the term they clicked.

But yes, it would likely be a bit better if the website just had a search function, I think that would get resource intensive for long logs though.

from ircv3-specifications.

1. Put it on a root (e.g. logs.ircv3.org, assuming we can arrange for that)
2. Split it by day.
3. Use Google or another search engine which supports site: searches.

No need for us to write a log-search feature when search engines exist.

from ircv3-specifications.

@Aerdan Mostly true, but Google's output is formatted for "it's in this page somewhere" being good enough: wanting to find the page rather than the exact line. This may or may not be an issue.

@TheSaunders I can actually think of a way to reduce it to the cost of formatting a slightly augmented logfile and additional disk space for the results of each prebaked search, plus a cron job, but avoiding both races and the need to reread the entire result file each time is tricky.

from ircv3-specifications.

@kythyria All browsers worth using have the ability to search within a page, so IMO this is a nonissue.

from ircv3-specifications.

@Aerdan Attempting to load months of text just so you can ctrl-f to search for text regarding an issue would not only be a waste of bandwidth but it would lag the browser considerably.

from ircv3-specifications.

Combining Google's site: search with a reasonable split length (say, one day = one file) makes it not really that big of a deal.

Putting months and months of logs in one file would be dumb, and I'm pretty sure they know that.

from ircv3-specifications.

When updating an issue, add a link to logs of discussion => the website hosting logs needs the ability to make a link to line number, similar to github

from ircv3-specifications.