On Fri, 22 Mar 2019 at 22:32:02 +0100, William Gathoye wrote:
> It appears the location of these fonts haven't been whitelisted properly
> leading to the Nextcloud client webview (qt5-webengine) to not load them
> to avoid a potential XSS vulnerability.
The CSP violation looks somewhat odd to me:
On Fri, 22 Mar 2019 at 22:15:17 +0100, William Gathoye wrote:
> I'm thus wondering:
>
> * how the system performs the match between branches and how he keeps
> track of the branches (see issue described just above ^)
> * and (more interesting for my use case), how to have a build string
>
Hello there,
I (credit goes to upstream) have been able to narrow down the
authentication issue we were experiencing with the Nextcloud desktop client.
Please note, up to know the latest ownCloud client is still working fine
except it is not keeping the SAML token after the system reboots/user
On 22/03/2019 19:04, Guilhem Moulin wrote:
>
> As the command output suggests, yes. Or at least that's what I propose,
> assuming releng accept to maintain that map.
Let's have our fingers crossed hoping this will be possible to
implement/keep over the long term then :)
>
> Depends what “this”
On Fri, 22 Mar 2019 at 18:12:33 +0100, William Gathoye wrote:
> On 3/22/19 3:55 PM, Guilhem Moulin wrote:
>> Assuming a point in the future where 6.1.5, 6.2.1, and testing/6.2.2 are
>> no longer served by the mirror network, but stable/6.2.2 still is:
>>
>> curl -sw"%{http_code}
On Fri, 22 Mar 2019 at 12:46:12 +0100, William Gathoye wrote:
> On 3/21/19 5:46 PM, Guilhem Moulin wrote:
>> JOIN'ing (ON h.sha256) the result of that query with a similar one (on
>> downloadarchive) we should be able to populate a static map for
>> ngx_http_map_module. That's likely to be more
On 3/21/19 5:46 PM, Guilhem Moulin wrote:
>
> JOIN'ing (ON h.sha256) the result of that query with a similar one (on
> downloadarchive) we should be able to populate a static map for
> ngx_http_map_module. That's likely to be more robust than a map
> obtained by manually scrapping HTML :-P
For