Tweet
Grateful for any insight/advice on resolving the following. I have a standard index job that runs through a couple of thousand .pdfs and matching .descs, spider mode but starting from an html file which explicitly lists the files to index. This job completes correctly with Zoom 5.1.1.1003, but fails with 5.1.1006 - Zoom crashes close to completion, the index log file is truncated in mid line with no error message, and the new index files all have .tmp names in the relevant directory. There are no errors in the system error log, so it is not a problem with the web server as far as I can see.
-
-
We found a new bug today. In version 5.1.1006 released last week we fixed a bug with indexing very long words (over the 35 character limit) which could sometimes cause Zoom to abort indexing the rest of the page, but still continue with subsequent pages.
Bug in fixing this bug, we created a new bug. Also related to long words over the 35 character limit. In particular circumstances this can cause a crash. Opps.
We have a solution for this new bug, but it will probably be next week before we can make it available. A workaround for most sites will be not to use comma's and dots as join characters. This reduces the word length.
So this might be your problem. Or it might be totally unrelated. If it always crashes on the same PDF, can you E-mail us the PDF in question? -
Further tests
I repeated the run with 5.1.1006 with the full set of files, and the usual join settings (dot and hyphen), and it crashed towards the end of the 2000+ files.
I then tried 5.1.1006 with just the last 100 files in the indexing job, which is where the crash ocurred. Everything ran ok.
I then tried unchecking all the join settings, and reran the job. It ran to completion.
So although I can't pin down the file that's triggering this crash, it does seem to be related to your suggestion. I'll wait for the fix and try again.
thanksComment
Comment