Hi, i found another bug in FTS while parsing quoted-printable/base64-encoded sender's name inside 'From' header (again, i was figuring out why some of messages cannot be searched). Header: From: =?UTF-8?Q?A=2CB?= <test at example.com> Data send to Solr by Dovecot: <field name="from">A@, B &lt;test at example.com&gt;</field> Header: From: =?UTF-8?Q?A_B=2CC?= <test at example.com> Data send to Solr by Dovecot: <field name="from">A@</field> Header: From: =?UTF-8?B?QSBCLEMK?= <test at example.com> Data send to Solr by Dovecot: <field name="from">A@</field> Dovecot 2.2.27. azur
On 9 Apr 2017, at 15.42, azurit at pobox.sk wrote:> > Hi, > > i found another bug in FTS while parsing quoted-printable/base64-encoded sender's name inside 'From' header (again, i was figuring out why some of messages cannot be searched). > > > Header: From: =?UTF-8?Q?A=2CB?= <test at example.com> > Data send to Solr by Dovecot: <field name="from">A@, B &lt;test at example.com&gt;</field> > > > Header: From: =?UTF-8?Q?A_B=2CC?= <test at example.com> > Data send to Solr by Dovecot: <field name="from">A@</field> > > > Header: From: =?UTF-8?B?QSBCLEMK?= <test at example.com> > Data send to Solr by Dovecot: <field name="from">A@</field> > > > Dovecot 2.2.27.This is a bit troublesome to fix and it's already been broken for years before anybody noticed.. I think it's going to take a few releases to get this fixed. I think we'd need to fix message-address parser to support EAI-compatible UTF8 headers and that likely fixes this as well.
Cit?t Timo Sirainen <tss at iki.fi>:> On 9 Apr 2017, at 15.42, azurit at pobox.sk wrote: >> >> Hi, >> >> i found another bug in FTS while parsing >> quoted-printable/base64-encoded sender's name inside 'From' header >> (again, i was figuring out why some of messages cannot be searched). >> >> >> Header: From: =?UTF-8?Q?A=2CB?= <test at example.com> >> Data send to Solr by Dovecot: <field name="from">A@, B >> &lt;test at example.com&gt;</field> >> >> >> Header: From: =?UTF-8?Q?A_B=2CC?= <test at example.com> >> Data send to Solr by Dovecot: <field name="from">A@</field> >> >> >> Header: From: =?UTF-8?B?QSBCLEMK?= <test at example.com> >> Data send to Solr by Dovecot: <field name="from">A@</field> >> >> >> Dovecot 2.2.27. > > This is a bit troublesome to fix and it's already been broken for > years before anybody noticed.. I think it's going to take a few > releases to get this fixed. I think we'd need to fix message-address > parser to support EAI-compatible UTF8 headers and that likely fixes > this as well.Is there any workaround until it's fixed? This bug is little complicated as complete fix includes also reindexing afffected emails (which, probably, means reindexing everything as it would be very hard to identify such e-mails).