 |
.. (לתיקייה המכילה) | |
Python's sqlite (sqlite3) doesn't support asynchronous operations, what should I do? | |
|
In general you should use the "aiosqlite" library instead of the regular "sqlite3" because accessed to the DB are long and costly. In our assignment you *may* use it (and we prefer if you do), but we *won't* check if you did (i.e. you can use the default one). Packages for async DB and file access are : "aiosqlite", "aiofiles". |
What should I do with the 2nd part of the HTTP Basic authentication, because the connection is non-persistent and I don't have the previous request? | |
|
That shouldn't matter, the 2nd epoch of the scheme contains all the needed information. The first response is to make sure the client sends the auth info, it can also send it beforehand. |
Should I recheck authentication in sub-directories after accessing the protected parent directory? | |
|
Yes, for every sub-directory under a protected path you should recheck authorization each time. The protocol specifies that the *client* should assume that the above happens, and preemptively send the authorization information. |
What should I do when there's no MIME type matching the file's type? | |
| Just leave the "Content-Type" header empty when building a response, it will automatically add the default one. |
In Windows, the file system doesn't differ between "Path" and "path" (caps), what should I do? | |
| You should just ignore the caps (just turn everything to small letters). |
Can the realms be contained\nested in each other? | |
| No, the realms are disjoint - for a path there is only one matching realm. |
Does every path has a corresponding realm and vice versa? | |
|
No, not every path should have a realm, but each realm should have a path. The mapping (in the DB) between path and realm are one-to-one. |