Sorting, Pausing, and other minor things

[Tracy]

1) It would be wonderful to be able to sort the Groups List (the one that holds the list of all subscribed groups, which may or may not be split up into groups of groups). Currently (version 6.82), I believe the program sorts the list alphabetically by group name. Not bad, but it would be nice to be able to change that (perhaps by searching for a specific word or phrase in the list of subscribed groups, and moving those to the top/bottom of the list). I can easily picture that particular thing being difficult, but it is a thought. It would also be nice to be able to sort by other columns (such as "New Files" and "Updated"), thus allowing a quick check of which groups have more things in them (perhaps an indicator of group activity), and when things were last updated (for those of us who only do manual header updates).

2) I would like to suggest the ability to pause header downloads. I know this has been suggested before, and rejected for what seemed good reason (to wit: there being no good way to choose a pausing point). But it seems to me that with the way that Newsbin currently downloads headers (downloading in chunks of ~50,000 headers at a gulp), it would be relatively easy to pause after the most recent group of 50,000 was received and before sending the next XOVER command (since, at least theoretically, you would have the article number of the most recently retrieved header, you would be able to resume downloading headers when "Pause" was lifted, by simply starting from that number and grabbing your next 50,000). Since I realize this might lead to a few lost headers (probably only a small number in a given gulp), it might be just as easy to back the article number up by, say, 100 or so when restarting, thus insuring that you got any missed headers (and since any that were downloaded as duplicates would be caught when importing the headers into the database, it should be a case of no harm, no foul).

3) I would also like to see the ability to "limit" the number of connections used to download headers. For instance, let's assume that there is 1 configured news server, and it allows 10 concurrent connections. In such case, allocating 1, or even 2 connections to downloading headers, and allowing the remaining connections to download other things (or even sit idle, if that is needed or desired by the user) would be a good way to do this. This could be configurable in the news server setup (total connections X, header downloading connections Y, where Y <= X). The case where Y == X would be the current situation, whereas when Y < X, connections would be open for other traffic (downloads, etc) or even just sit idle (for instance, when someone wishes to download headers in the background on a shared internet connection, but doesn't want to hog the entire connection (and doesn't want to put a "speed limit" throttle value in place for Newsbin for whatever reason), so that if no one else happens to be using the internet connection at the moment, it can download in relative peace, but if someone else *is* using the connection, they at least get some bandwidth). In fact, being able to "dynamically" configure these values would be nice as well - so, for instance, when there are things to be downloaded, the number of header connections could be decreased (thus allowing for downloading files, but not completely disallowing header downloading), then increased later when there were no waiting files (or when someone was, say, going to bed for the night and wanted to open up header downloads to the max).

Thoughts?

[Quade]

You can do 3 now by adding a copy of the single server, setting the connections on the copy to 8 then set the initial server to 2. Disable header downloads on the copy server so those 8 connections are only used for files. The primary server will then be used for headers and files once the header download completes. People have asked to use multiple connections to download headers for a single group too. My main concern with adding too many options to header downloads is errors causing gaps in the headers. Once I chop of the header downloads into more discrete units, the chance of a header download error causing a gap in headers downloads increases.

1 and 2 I've noted. The next version is going to have major changes so, they might be possible. I'm thinking the groups list needs a "find" more than sort. So you could just type the name in at the top and only show matching groups. Not saying you won't be able to sort.

[Tracy]

Yes, I have been toying with changing the number of connections (in regard to limiting header downloads), and it does work. Just a tad cumbersome (making sure the right server is handling the right tasks and such). But it is absolutely a good workaround for the time being.

I do agree that in terms of the group list, Find would be very useful (probably a first choice, if it comes to a choice on which things to implement). But the sort - especially for columns other than the group name itself - would be quite useful as well.

Thank you for adding them to your list of things to look into

[Quade]

Yes, I have been toying with changing the number of connections (in regard to limiting header downloads), and it does work. Just a tad cumbersome (making sure the right server is handling the right tasks and such). But it is absolutely a good workaround for the time being.

I guess where we differ is I never want more than 2 connections for header downloads. After these header downloads complete, these two connection go right back to file downloading. As long as both servers are the same priority they'll run in parallel.