El diumenge, 24 de juliol de 2016, a les 8:09:36 CEST, Michael Joost va escriure: > For me, any use of pdfsig results in just a segfault. I am using > poppler-0.45 with nss-3.25.
Can not reproduce, can you please open a bug in bugs.freedesktop.org and attach one of the files you're using to have this crash? Cheers, Albert > As far as I can tell, the signature verification itself run's ok (valid > signature result), but there is some kind of double-delete error caused > from the SignatureHandler destructor. > > The destructor first calls NSS_CMSSignerInfo_Destroy(CMSSignerInfo), > which destroys CMSSignerInfo->cert, hence, fills it with dada pattern. > > in NSS_CMSSignerInfo_Destroy(): > -------------------------------- > print si->cert > $74 = (CERTCertificate *) 0x6e8010 > > Subsequently, the SignatureHandler destructor calls > NSS_CMSSignedData_Destroy(CMSSignedData), which attempts to destroy > CMSSignedData->tempCerts[*] > > But one of those tempCerts (in this case the one-and-only, but not > necessarily the first) is the one previously destroyed with > NSS_CMSSignerInfo_Destroy(), thus having the dada pattern cause the > segfault. > > in SignatureHandler::~SignatureHandler(): > ----------------------------------------- > print (CMSSignedData->tempCerts[0]) > $76 = (CERTCertificate *) 0x6e8010 > > > There seems to be no obvious change in the destroy code in recent > versions. Not sure why I'm the only one running into this problem. > > > br, > > Michael Joost > sonet at michael-joost dot de > _______________________________________________ > poppler mailing list > [email protected] > https://lists.freedesktop.org/mailman/listinfo/poppler _______________________________________________ poppler mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/poppler
