Update and improve myth.rebuilddatabase.pl

[gnassas@…]

I tried to use myth.rebuilddatabase.pl for the first time today and noticed that it's stuck on recordings ending in .nuv and also that they should conform to the 0.18 format for filename (channel_starttime_endtime.nuv) so I thought I'd fix that up. Then I thought up a few features that I'd like to see so I started hacking...

The idea around --file and --answer are that you could write a (possibly spreadsheet-driven) script that would import tons of recordings without intervention. Like perhaps if you were importing all the episodes from one of those sitcom dvds...

Although this looks like a huge change (the entire body of the script has been rewritten) it's actually pretty much the same as before. I mostly reduced the two paths of execution (nuv vs. non-nuv) into one and added a helper routine for user queries.

• George

[gnassas@…]

the new script

[gnassas@…]

just noticed that commflag takes different arguments depending on the video file's origin

[Isaac Richards]

(In [9011]) myth.rebuilddatabase script update, closes #1263.

[gnassas@…]

Sorry to reopen this.

Nathan Stratton wrote me off list to point out that matching recordings by filename *and* hostname works badly when you have multiple backends -- everything not recorded on the BE you're running on looks like an unknown recording and triggers a "Do you wish to import" prompt. I removed hostname from the matching query.

While I was in there I noticed that exec is being used to invoke mythcommflag. Obviously this isn't healthy inside a loop. Fixed.

Also, open up the default extensions to include the usual suspects.

Sundry other tweaks.

[gnassas@…]

this is the last of it, I promise!

[Isaac Richards]

(In [9019]) New patch fixes some issues. Closes #1263 again. =)