D9033: man ioslave: asserts trying to display pam(8)
Jonathan Marten
noreply at phabricator.kde.org
Tue Nov 28 13:00:58 UTC 2017
marten created this revision.
marten added a reviewer: Plasma.
Restricted Application added a project: Plasma.
Restricted Application added a subscriber: plasma-devel.
REVISION SUMMARY
That page, and many others, links to another page with just a plain name:
.so PAM.8
This is read by man2html and passed to MANProtocol::readManPage() with a filename of "../PAM.8". This is then resolved against the containing directory "/usr/share/man/man8", becoming "/usr/share/man/PAM.8". No such file exists and the mandir.entryList().first() asserts because the list is empty.
The relative filename is correct in the case of other pages. e.g. telinit(8) which uses:
.so man8/init.8
So it is not enough to simply ignore the "..". The change to man2html here attempts to correct the reference by detecting whether it is of the form "../page.sect" and is not "../manN/page.sect", and removing the ".." component only in the first case.
While investigating this I found it useful to have a better diagnostic if the .so page could not be found. The change here to kio_man.cpp detects that case and displays an error page instead of asserting. This should only happen if there is a bug in the man page, but it is useful to have a pointer so that the upstream man page can be fixed.
TEST PLAN
Built kio_extras with this change. Verified correct display of the man pages:
man:/pam(8) - simple page name
man:/service(8)
man:/telinit(8) - relative page name with section
man:/vigr(8)
man:/mq_open(2) - relative page name to another section
REPOSITORY
R320 KIO Extras
REVISION DETAIL
https://phabricator.kde.org/D9033
AFFECTED FILES
man/kio_man.cpp
man/man2html.cpp
To: marten, #plasma
Cc: plasma-devel, ZrenBot, progwolff, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, apol, mart
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/plasma-devel/attachments/20171128/503c24a2/attachment.html>
More information about the Plasma-devel
mailing list