This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Re: Networking Help


>>>>> "Huge" == Hugo Tyson <hmt@redhat.com> writes:

    Huge> Fabrice Gautier <Fabrice_Gautier@sdesigns.com> writes:
    >> > > After that I get a conflict which states:
    >> > > 
    >> > > Requires constraint not satisfied: CYGBLD_ISO_BSDTYPES_HEADER ==
    >> > > "<sys/bsdtypes.h>"
    >> > > 
    >> > > What do I need to change to correct this?
    >> > 
    >> > Click Continue? If that doesn't work, just update the value 
    >> > of the option
    >> > by hand.
    >> 
    >> I got the same problem in several occasions. It seems that the
    >> configuration tools is not abble to resolve conflicts which
    >> involve a string value.

    Huge> That's true AFAIK, it's a bug.

    Huge> Here's a tip: in the conflicts window, you can click on a
    Huge> symbol - maybe with the other button - to get a menu
    Huge> including "locate"; hit that and it'll take you to the item
    Huge> so you can enter the value by hand. Saves pain searching for
    Huge> it.

    Huge> I believe the CLUI tool *can* resolve this correctly.

Correct, ecosconfig gets this sort of thing right. The problem is that
the GUI tool invokes libcdl's inference engine in slightly different
ways from the command line tools. There are reasons for this, e.g. in
an interactive tool it makes sense to let users confirm or reject the
decisions made by the inference engine. Unfortunately in some
circumstances you end up with different results, e.g. the GUI tool
cannot resolve a conflict that ecosconfig takes in its stride.
Investigating why this happens is on my TODO list.

Bart

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]