On 29/03/20 at 23:52 +0200, Adam Borowski wrote: > On Sun, Mar 29, 2020 at 05:53:11PM +0200, Xavier wrote: > > >>> IIRC with an access token the limit is 5000 queries per hour. Without a > > >>> token, much, much less. > > >> > > >> The UDD code calls uscan. Is there a way for uscan to use an access > > >> token? > > > > > > not for now, but I can add this feature. Just to insert an additional > > > header ? > > > > This little diff adds --http-header option. To use it: > > > > uscan --http-header=Access-Token=qwertyuiop > > > > Sounds good for you ?
Yes, thanks! > Tokens hardly ever are portable between sites, thus I guess they'd need to > be used only for matching URLs. I'm not sure if udd already parses target > servers or calls uscan as-is, letting it do all the work. > > Second issue is, tokens are (initially) an user-"password" pair that > requires further processing. For example, GitHub uses Basic http auth > (https://developer.github.com/v3/auth/). If I read the RFC > (https://www.ietf.org/rfc/rfc2617.txt) correctly, Basic auth gets mangled > into a header and is sent as-is; no idea if other token-using webservices > use fancier auth schemes. Even if just Basic is used, it might be simpler > to not require the user how to do the mangling. In the UDD case, I think that's enough to fix the present issue. I'm not sure something more general should be done? Lucas