Sundar,

A few more comments below.

Sue


Sundar Yamunachari wrote:
> Sue,
> 
>    Thanks for the review. See my response below. I have updated the 
> webrev at http://cr.opensolaris.org/~ysundar/3976/ with the changes 
> based on the comments I accepted.
> 
> Thanks,
> Sundar
> 
> Susan Sohn wrote:
>>
>> Sundar Yamunachari wrote:
>> > Hi Jan and Sue:
>> >
>> >     Please review the code for bug fixes.
>> >
>> >     The webrev is at: http://cr.opensolaris.org/~ysundar/3976/
>> >
>> >     3976 AI DHCP lease is short and not being extended
>> >     http://defect.opensolaris.org/bz/show_bug.cgi?id=3976
>> >
>> >     4070 Client booted with AI image cannot access outside the subnet
>> >     http://defect.opensolaris.org/bz/show_bug.cgi?id=4070
>> >
>> >     3968 installadm does not create directory
>> >     http://defect.opensolaris.org/bz/show_bug.cgi?id=3968
>> >
>> >     3945 installadm delete-service doesn't remove the target 
>> directories
>> >     http://defect.opensolaris.org/bz/show_bug.cgi?id=3945
>> >
>> > Thanks,
>> > Sundar
>>
>>
>>
>> Hi Sundar,
>>
>> installadm.c
>> ------------
>> General: bug 3945 says the target directory isn't being deleted, but 
>> the code changes at 566 look like it's the service that wasn't being 
>> deleted?
> The problem happened because, there were issues with 
> remove_service_data() code. When I asked to remove one entry, it was 
> corrupting everything because of the way I used strtok (see the change 
> made in 1289).
> The change in 566 is to make sure that the service record is deleted 
> when delete-service is called
>>
>> 1259 - not in your changes, but should there be an error message here?
> If the file doesn't exists that means we don't need to remove the entry.

So there are valid cases that remove_service_data would be called and that file 
wouldn't exist? Maybe add a comment?

>> 1269 - if you intended for this file to be in /var/tmp rather than 
>> /var, you're missing the final "/" in "/var/tmp/". And maybe call it 
>> /var/tmp/ai.<pid> or /var/tmp/installadm.<pid> ?
> okay. I will change it to /var/tmp/installadm.<pid>
>>
>> 1281 - before you return, you need to cleanup tmp_fp and associated file
> okay. Actually tmp_file is a local variable and so it should not be 
> freed in 1324. I will use fclose().

1281 - Don't you want fclose(tmp_fp) rather than fclose(tmp_file)?

>> 1299 - Once you get service_name and directory, why not do the check 
>> on 1312? Why do you look for file and data? In other words, it seems 
>> like 1299-1306 could be deleted?
> Right now I am using only service_name and directory because I am 
> deleting only directory now. We will be deleting boot file and directory 
> corresponding to the web server after November. So I don't want to 
> delete the code now.
>>
>> 1309 This comment needs to be rewritten to reflect what code is 
>> actually doing (i.e., only writing line if it isn't a match to existng 
>> service/image)
> okay.
>>
>> setup-dhcp.sh
>> -------------
>> 165 - should the error message be expanded here so that if you can't 
>> get the default route, you tell the user that they won't be able to 
>> access outside the subnet or whatever the implications would be? 
>> And/or say "continuing anyway", since you don't abort?
> okay. It is kind of difficult to tell them that the client is the one 
> that is affected. I will change it to "Cannot get default router. Please 
> check whether default route is configured". We don't abort for any DHCP 
> errors. I don't want to add "continuing anyway" just for this message.

ok, but you have a typo: fdefault->default

Sue

> Thanks,
> Sundar
>>
>>
>> Sue
>>
> 


Reply via email to