Ubuntu
remotefs package

remotefs fails to build karmic, gcc 4.4 strict aliasing

Bug #439682 reported by David Sugar
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
remotefs (Ubuntu)
Fix Released
Undecided
Fabrice Coutadeur

Bug Description

This is a common problem in legacy coding patterns written around handling socket addressing and gcc 4.4's more strict aliasing checks. This bug exists to post a fix and close building remotefs for Karmic.

Related branches

lp:ubuntu/karmic/remotefs
David Sugar (dyfet-deactivatedaccount)
Changed in remotefs (Ubuntu):
assignee: nobody → David Sugar (dyfet)
status: New → In Progress
Revision history for this message
David Sugar (dyfet-deactivatedaccount) wrote :

fix for gcc 4.4/karmic attached as a debdiff. Should also be pushed to upstream...

Changed in remotefs (Ubuntu):
status: In Progress → Confirmed
Revision history for this message
Fabrice Coutadeur (fabricesp) wrote :

Hi,

Don't you think that deleting the cast on line 240 is enough to fix the FTBFS? This change would be less intrusive, and we won't be casting const to non const structure, just to get a value.
I agree with your second change with ipv6.

If you agree with my comment, please update the debdiff with the new change, and also update the maintainer field (you can use the update-maintainer script for that) :-)

Unsubscribing u-u-s for the moment. Please resubscribe them when the debdiff is ready.

Thanks,
Fabrice

Changed in remotefs (Ubuntu):
status: Confirmed → Incomplete
Revision history for this message
David Sugar (dyfet-deactivatedaccount) wrote :

I had thought about that, but the actual type of client_address is conditionally determined from an #ifdef for IPV6, hence unless it is cast to a known type in advance, you do have to do the cast in line 240 because it might either be a struct sockaddr_storage * or a struct sockaddr_in *. Hence, I did not see a less intrusive way to solve this at the time, at least for all possible compile paths. If you suggest we keep the code change and I simply correct the Maintainer field, I could re-upload for that.

Revision history for this message
Fabrice Coutadeur (fabricesp) wrote :

Make sense.
 I'll upload your debdiff (and take care of the maintainer change :-) ).

Thanks,
Fabrice

Changed in remotefs (Ubuntu):
assignee: David Sugar (dyfet) → fabrice_sp (fabricesp)
status: Incomplete → In Progress
Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package remotefs - 0.11+debian-1ubuntu1

---------------
remotefs (0.11+debian-1ubuntu1) karmic; urgency=low

  * src/server.c: Fix for gcc 4.4 strict aliasing requirements (LP: #439682)

 -- David Sugar <email address hidden> Wed, 30 Sep 2009 22:08:21 +0000

Changed in remotefs (Ubuntu):
status: In Progress → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.
Subscribing...

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.