<html><head><meta http-equiv="Content-Type" content="text/html charset=us-ascii"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;"><br><div><div>On Jun 15, 2013, at 8:46 AM, Christopher Armstrong <<a href="mailto:radix@twistedmatrix.com">radix@twistedmatrix.com</a>> wrote:</div><div><br></div><blockquote type="cite"><div style="letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;">Yeah... I feel like a lot of the things you say in this email aren't<br>related to the point I was trying to get across, which was the<br>interface and effect, not the implementation of it.<br></div></blockquote><div dir="auto"><br></div><div dir="auto">Even if it's not really what you'd intended, I think it's pretty good to keep discussing these issues (avoiding unnecessary formatting, putting the right responsibilities in the right place).</div><br><blockquote type="cite"><div style="letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;">I also assume that developers will be able to hook up additional<br>observers to save out .concatenated-json files, or stream to some<br>network protocol, or whatever. <br></div></blockquote><div><br></div>To reiterate a little from my previous reply, I think we should be, as much as possible, moving to make that the default. Now, <i>actually</i> making it the default is a pretty disruptive change, but I think we could get as far as "implement the feature at all" before anyone would start complaining ;).</div><div><br></div><div>In other words, every developer should not have to realize they should write a JSON(ish) emitter and figure out how to get that to work on their own; we should just provide one that does the right thing.</div><div><br></div><div>To dial this back to what we actually need to talk about in the proposed log system, the issue is mostly to do with </div><div><br><blockquote type="cite"><div style="letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;"><blockquote type="cite">I can implement the UUID thing as a completely separate ticket and we can<br>discuss it on its own merits; meanwhile, you can happily add log IDs to<br>everything in your applications. I'd be quite happy to table this part of<br>the conversation for a separate ticket; I just thought it would be a nice<br>thing that could fall out of some of the other parts of a new logging<br>system.<br></blockquote><br>Yeah, I'm sorry that this has gotten so much scope-creep.</div></blockquote><div><br></div><div>No problem. This is all good stuff to talk about; we don't discuss operational concerns of managing Twisted services (either server *or* client side) nearly often enough on this list. I wish we had more discussions like this!</div><br><blockquote type="cite"><div style="letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;">I'll reiterate that I think redoing what I've been speaking about as the<br>"system" key is most important (but maybe it should be called<br>something other than "system", since you've clarified what its intent<br>was in another email).<br></div></blockquote></div><div><br></div><div>Well, "system" was a pretty horrible name for the thing satisfying that intent. We have "log_namespace" in the new logging system, and I think that (in combination with an idiomatic "log_id", that only needs to be unique within that namespace) you could get more or less what you want.</div><div><br></div><div>I think we're going to drop the "system" key entirely (except in the contexts where it's required for compatibility, of course). Jean-Paul's recent comment in the quote file is apropos.</div><br><div>-glyph</div><div><br></div></body></html>