Custom Query (420 matches)

OAI download sometimes needs to get two files, but only one is present.

The code (OAIDownloader) goes through and looks for identifier tags, extracts the URL, works out the extension, then saves the file as the same name as the oai file with the new extension. This is fine until you come to a case where there are two identifiers with the same extension, as they get saved to the same filename.

1. Thread problem: random infinite wait for next action in queue when the queue should actually already contain the next action
2. When the username/password typed in at the startup of GLI is wrong, keeps popping up invalid authentication dialog
3. exit condition when user presses cancel, or on connectexception should quit while(true) loop in ActionQueue.java's run() and Gatherer should exit()
4. New bug: GLIapplet used to build and preview a collection, then browser pointed elsewhere before user returns to open the GLI applet and rebuilds the collection. On clicking preview the browser crashes.

• Fixed: Bug in GLIapplet since 2.80 and after. If after using the GLIapplet, the user pointed the browser elsewhere and returned to the GLIapplet, this would cause the reloading of GLIapplet to wait forever.

Fixed by making RemoteGreenstoneServer class into an OOP class rather than being filled with static methods and variables. Now this class is split into 3, the other two being RemoteGreenstoneServerAction.java and ActionQueue.java

Prof Witten identified several problems to do with the Download panel.

Subject: Problems with Download panel Date: Wed, 27 Aug 2008 05:02:36 +1200 From: Ian Witten <ian.witten@…> To: Kathy Don <kjdon@…>, David Bainbridge <davidb@…>

I've been playing with the Greenstone Download panel and encountered a few problems. Maybe it's just my setup here and things have been fixed since I downloaded Greenstone just before I left. But if not, it would be great if these were fixed before the upcoming release, though maybe they're not top priority -- I'll leave the decision up to you. The *'d ones are probably enhancements that you shouldn't consider for the next release, the ?? one is probably not our problem at all.

1. I just rebuilt the OAI demo collection (oai-e), and the extracted

metadata in the Enrich panel looked like ex..Title, etc (with two dots). Although the collection built, the classifiers etc were wrong of course.

2. When using the Download panel to download websites, the Get

Information button produces a popup panel that is completely blank.

3. When downloading from ​http://englishhistory.net/tudor/citizens with

both the "below this URL" and "within the same site" options switched on and a download depth of 2, it downloads about 450 files including ones like ​http://englishhistory.net/tudor/*. But the "below this URL" option should have prevented this, no?

4. *? When you download from the same place with a download depth of 1,

you get about 35 files including a few (4, I think) with names like ​http://englishhistory.net/tudor/citizens/?N=D. Can this be prevented?

5. ?? If you then download a second time, it re-downloads some of the

files even though they haven't changed. (some message about the timestamp not being set, maybe it should be set by the webserver and there's nothing we can do about this).

6. * The log file accumulates, whereas in the interface it seems clear

that the log file pertains only to the download operation that you have just done. It's really confusing.

7. I couldn't get the Z39.50 or SRW/SRU parts of the Download panel to

work at all. E.g. I tried with

• Host: z3950.loc.gov
• Port: 7090
• Database: Voyager

(see www.loc.gov/z3950/lcserver.html#addr).