Tim wrote: > Mikkel L. Ellertson: >> One way would be to have a utility that would go through the files, >> determine the mime-type, and set the attributes. Run it as a nightly >> cron job to index new files. > > Ugh, rather than store data in a sane way in the first place, do more > work on top of it, repeatedly. I'm reminded of mail programs that > continually re-index their mail folders, rather than update their > indexes when things get changed, leaving them alone in the meantime. > No - this was in a way to handle files that were added, but were missing the data, not files created properly. IN other words, how to handle files that would otherwise fall through the cracks. If we could bet all the file formats changed to include the mime-type data in the file header, that would be a better fix. But somehow I do not see that happening with existing formats. also, you would not be checking every file - just the ones missing the mime-type attributes. Unless you told the system to rebuild the attributes for every file, then cron job would just check to see if the file had the mime-type set. If not, try and determine the type and set the attribute. If that fails, set it to type unknown, so you do not have to try it again next time, and generate a message for the logs. That way, you can grab the list from the logs, (or search for the type unknown attribute,) and set the type manually. As more programs are updated to properly handle the mime-type attributes, then nightly cron job would have less to do. I am not sure if file managers should have the job of creating the attributes when bring in files form file systems that do not support them, or if that should be left to the nightly cron job... Mikkel -- Do not meddle in the affairs of dragons, for thou art crunchy and taste good with Ketchup!
