On 13 Sep 2004, at 00:32, Alex Tweedly wrote:Looks like a bug. There should be no difference between http and ftp downloads to file, but it seems I managed to mess up here. (At some point, I added some cute stuff for closing files when doing uploads and it looks like I overlooked that this might interfere with downloads.)
I don't know if this is a bug, or if this difference between http:// and ftp:// forms is intentional.
Thanks Dave - I'll cut it down to a simple example stack and submit.
By the way, libUrlDownloadToFile was only added for dealing with huge files that might not fit in available memory. If dealing with "normal" files, I'd recommend using simple gets and puts.
get url "ftp://whatever"; if the result is empty then put it into url "binfile: whatever" else ##error stuff end if
I'm using libURLDownloadToFile to asynchronously load a disk cache, where the total space could be large, even though each file is small (50-250K). (I want each one stored on local disk, both for caching between sessions and for off-line usage).
I could (perhaps should) have done it with load URL and cachedURLs and unload URL - but it seemed simpler to just transfer them direct to disk. I'll take a look at changing over.
Is there a limit to the number of simultaneous loads ?
If I issue say 20 load commands, are they queued so that only N happen at a time ?
If so - what is N ? And, of course, can it be changed ?
Thanks -- Alex.
_______________________________________________ use-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/use-revolution
