Make extraction of ETag header independent of capitalisation

[PatrickJosh]

Fixes #7703.

Without this pull request, the extraction of the ETag header in a reply from the server is dependent on the capitalisation of the name of the field being ETag because rawHeaderList().contains() is not case-insensitive. However, according to the standard, header field names are case-insensitive. This pull request makes the extraction of the ETag header case-insensitive by using the header function instead of rawHeaderList().contains(). Internally, Qt performs a case-insensitive comparison of the field names, as can be seen here.

I have tested this with my server and can confirm that I see notifications again.

[PatrickJosh]

Actually, when first proposing this pull request, I thought that raw headers in Qt were not case-insensitive at all; this is not the case. Raw headers in Qt are case-insensitive, but using rawHeaderList().contains() is not because rawHeaderList() only gives you a list of the header names in their original capitalisation and does not make them all lowercase, for example. So, a smaller fix would have been the following:

--- a/src/libsync/networkjobs.cpp
+++ b/src/libsync/networkjobs.cpp
@@ -999,7 +999,7 @@ bool JsonApiJob::finished()
     }
 
     // save new ETag value
-    if(reply()->rawHeaderList().contains("ETag"))
+    if(reply()->hasRawHeader("ETag"))
         emit etagResponseHeaderReceived(reply()->rawHeader("ETag"), statusCode);
 
     QJsonParseError error{};

Still, my original patch works and does, in my opinion, not have any downsides compared to this one.

[PatrickJosh]

I have now also fixed the same code in gui/ocsjob.cpp although that did not cause any issues regarding #7703 for me.

[claucambra]

A couple of nitpicks but LGTM

[claucambra]

Awesome, thanks!

[PatrickJosh]

Hi, thanks for merging! My I quickly ask, why is this not included in the 3.15.3 milestone? It is a rather annoying bug in my opinion. If there's nothing that keeps this from being included in 3.15.3, I'd like to kindly ask you to reconsider this for 3.15.3. Thanks!