libicui18n.so.36 not found, required by "evolution"

[E. J. Cerejo]

Dominic Fandrey wrote:
> E. J. Cerejo wrote:
>> Dominic Fandrey wrote:
>>> E. J. Cerejo wrote:
>>>> Gerard wrote:
>>>>> On Sun, 10 Feb 2008 10:23:36 -0500
>>>>> "E. J. Cerejo" <ejcerejo at optonline.net> wrote:
>>>>>
>>>>> [snip]
>>>>>  
>>>>>> Can portmanager work in conjection with portupgrade?
>>>>>>     
>>>>>
>>>>> Yes, I use it all the time.
>>>>>
>>>>>  
>>>>>> Why has the ports tree be up to date?
>>>>>>     
>>>>>
>>>>> What conceivable reason would you have for using an outdated ports 
>>>>> tree?
>>>>>
>>>>>  
>>>>>> Will portmanager download anything from the cvsup repos?
>>>>>>     
>>>>>
>>>>> It will only fetch programs that need updating, just like portupgrade.
>>>>>
>>>>> Would you be so kind as to explain your reluctance to update your
>>>>> system? The number of potential programs that need updating seems
>>>>> rather immaterial when compared to the potential system wide
>>>>> improvement in its overall performance. You could simply start the
>>>>> upgrade in the evening when you are through using the PC. Depending on
>>>>> the speed of your machine, it might very well be done by the next
>>>>> morning, if not sooner.
>>>>>   
>>>> My system was updated yesterday and I'm trying to resolve the issues 
>>>> that arose from the updating.  I can't update my system everyday I 
>>>> just don't have time for it.
>>>
>>> If you don't want to rebuild the stuff, just add
>>>
>>> libicui18n.so.36    libicui18n.so
>>>
>>> to your /etc/libmap.conf file. This solution works if no functions 
>>> have been removed from the library interface, which only very rarely 
>>> happens.
>>
>> I just ran pkg_libchk -m piped to a file but looks pretty confusing, 
>> is there a way to get only the output for the lines containing 
>> /usr/local/lib/libicui18n.so.38?
> 
> I can reccomend you the -q option, which will give you a very clean 
> output simply listing all the packages that have problems.
> Without the -q option the program shows its status, which is not well 
> readable in a file, because it doesn't use \n but \r.
> 
> That you get output with the -m/--mean flag, but not without means that 
> the false positive detection is erroneous. For the time being I'd 
> recommend you to use the libmap.conf solution.
> Later this day I will add some debugging functionality to the script and 
> send that version to you, if you are willing to spend your time helping 
> me finding this problem.
> 

Well I'm glad to say that the output problem was just a minor problem, 
the cat and grep worked fine when sending the output to the terminal and 
I just copied it and pasted it into gedit.  And indeed pkg_libchk is by 
far the best solution, if I was to update every package that depended on 
icu I would have to update 239 packages which discouraged me right away, 
by running pkg_libchk the amount of packages to be updated was reduced 
to 71, wrote a quick script and left it running overnight and those 
packages got updated and by god I don't have any more problems with 
libicui18n.so.38.  Nice little tool.

One more question.  I see that there is no man page for pkg_libchk, how 
do I find out more about its options and switches?  What command or 
commands does it rely upon?