 |
 |
@abh3 commented on this pull request. > @@ -142,7 +143,9 @@ XrdHttpSecXtractor *XrdHttpGetSecXtractor(XrdHttpSecXtractorArgs) // Now return the interface object // - return (XrdHttpSecXtractor *)new XrdVomsHttp(eDest, *vomsFun); + auto base = static_cast<XrdHttpSecXtractor *>(new XrdVomsHttp(eDest, *vomsFun)); + auto wrapper = static_cast<XrdHttpSecXtractor *>(XrdVomsMapfile::Configure(eDest, base)); + return wrapper ? wrapper : base; Two comments: 1) Indeed we did agree to that but that was before issue 1577 which requested we converge the voms mapfile support. That is a good idea but for a clean convergence the code would have to be moved into the plugin. However, that's fairly easy to do given the wrapper which is used for xroot and http. So, that's why I did a turn-around. 2) Actually, you really don't have to change the plugin protocol, just enhance it. So, if you return MAP_FAILED instead of zero then the caller will know there was a fatal error while zero (still used by old plugins) would mean to ignore the plugin for whatever reason. Note that MAP_FAILED is defined in sys/mman.h and specifically is used to indicate that a memory address could not be established due to am error. -- Reply to this email directly or view it on GitHub: https://github.com/xrootd/xrootd/pull/1572#discussion_r780603802 You are receiving this because you are subscribed to this thread. Message ID: <[log in to unmask]> ######################################################################## 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