In an email of 14 September bernard at oracle requested:
For calendars configured "On the Network" in Format
"iCalendar (ICS)" Lightning issue an HTTP GET request
with the following request header:
Could Lightning specify the following value instead?
Taking this. The Accept: header probably should actually be "text/calendar; charset=utf-8", and we should properly set the Accept-Charset: header as well.
Created attachment 281300 [details] [diff] [review]
tested w/ http:// and file:///; I want to test against ftp:// before requesting review.
Bruno, could you please provide an update on the status of this patch/bug?
Bruno, do you want to ask for review for the patch? Do we need the "; charset=utf-8" part if we specify a separate Accept-Charset header?
Actually, you can't specify ";charset=utf-8" in the Accept request header.
The semi-colon is used to separate the "media-range" from the "accept-params".
Created attachment 340799 [details] [diff] [review]
patch rev 2
Comment on attachment 340799 [details] [diff] [review]
patch rev 2
>- channel.loadFlags |= Components.interfaces.nsIRequest.LOAD_BYPASS_CACHE;
>- channel.notificationCallbacks = this;
> var downloader = Components.classes["@mozilla.org/network/downloader;1"]
This will add the etag check to the backup download, but I don't see 304 handled in that code. As far as I see this may lead to empty backups.
Bernard, is this still an issue? Requesting text/calendar might be more correct, but for servers that serve such files as text/plain we break interop.
How about using the following?
Sounds good to me. We'd have to find out what happens if there is no accepted content type and display the right kind of error. Now we just need someone to put up a patch ;-)
Created attachment 8565562 [details] [diff] [review]
Comment on attachment 8565562 [details] [diff] [review]
Review of attachment 8565562 [details] [diff] [review]:
Looks good, r=philipp
Created attachment 8565578 [details] [diff] [review]
Adding patch headers
Comment on attachment 8565578 [details] [diff] [review]
Please mark r- patches as obsolete and patches to be checked in as r+ in the future. That makes handling checkin-needed easier. Thanks!