At present (and, it appears, for as long as the feature has been implemented), there is a discrepancy between how the documentation tells you to configure "composite ids" for authdb files, and how the code validates them.

In the documentation (ie: https://xrootd.slac.stanford.edu/doc/dev50/sec_config.htm#_Toc42289936 )
The example says:

Example

= atlddm o atlas r production u ddm

= atlprod o atlas r production

x atlddm /atlas ld

x atlprod /atlas lrw

o atlas /atlas lr

A client whose name is ddm and is a member of the atlas organization and has a role of production is allowed to lookup and delete files in /atlas. A client who is a member of the atlas organization and has a role of production is allowed to lookup, read and write files in /atlas. Otherwise, client’s who are members of the atlas organization can only lookup and read file in /atlas.

And further up, the description of the idspec values is:

idspec

is a single letter indicating the type of entity identifier that follows. You must specify one or more of them.

This will not work, however, as the code is looking for

// Id's are of the form 'c:', make sure we have that (don't validate it)
//
if (strlen(pp) != 2 || !index("ghoru", *pp))

That is, it wants something that looks like "o:" not "o".

Changing the provided specifiers in the Example to

= atlddm o: atlas r: production u: ddm
= atlprod o: atlas r: production

(and not altering the x lines)

successfully removes the error.

Could either the documentation be changed to agree with the code, or the code be changed to agree with the documentation, whichever is out of sync with the design intent?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or unsubscribe.

Use REPLY-ALL to reply to list

To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1