In a recent upgrade the terminal said there were about 600 updated documents, and dhelp ran for a long time, perhaps 2 hours (might have been more), but did eventually complete. This was on fairly decent hardware, though a few years old.
Perhaps in earlier cases the index process would have run to competion if I'd given it enough time. When I first checked, indexing was hung up running gs on a particular document. I later tried to figure out which document, but the process had finished. Is there any logic in what dhelp does to kill a process if it takes too long? Otherwise, that particular process either finished or had an error. After that I noticed that new processes kept appearing, apparently on new documents, so I let it run. I know there used to be some bugs in gs in which it hung up forever on some documents (some of which were on my system). I was under the impression the bugs had been fixed. I see at least these possibilities: 1) The indexing process really does take a very long time. 2) One or more of the index programs being invoked is messed up, resulting in the long times. Likely only some documents trigger the problem. 3) Something in dhelp (or whatever is controlling the indexing) is messed up. The initial discussion on this bug seemed to focus on 1); I suspect something else is going on. But I really have no idea what the internals are. If 1) is true, it would probably be good if the index process did not hold up the installations, either by backgrounding it or relying on cron. But I'm not sure what the implications of that are for people who want to use the docs before they are reindexed. Ross -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
