[Ns-developers] ns-3.7.1 Release Candidate

[Faker Moatamri]

Tom Henderson wrote:
> On 2/23/10 4:18 AM, Faker Moatamri wrote:
>> Hi,
>>
>> This is a bug fix only release, for now it includes the above bug fixes.
>>
>> If you have more changes to be done to it, please don't hesitate to
>> suggest them to me and give me the changeset to be included in ns-3.7.1
>> release.
>>
>> Ns-3.7.1-RC is available on:
>> http://www.nsnam.org/releases/ns-allinone-3.7.1-RC.tar.bz2
>> please test and report errors if any.
>>
>> The ns-3.7.1 release is scheduled on Monday March 1st.
>
> Faker, first, thank you for putting this release together.  Sorry for 
> the somewhat late comments, but I suggest that you postpone this 
> release until the following:
>
> 1) I realize that bug fixing is ongoing but there are a few bugs that 
> I think could be considered, unless there is planned an ns-3.7.2 
> release (note that I am assuming no plans for such a release, unless 
> critical stuff were to arise)
>
>  - bug 816 fixes some problems for the emulation/virtual machine users 
> (patch committed today)
>  - bug 602, fix merged today
>  - bug 706 on wifi backoff counting
>  - Pavel's forthcoming patch for AODV bug 772, which users have 
> complained is blocking them
>  - bug 820 (test.py logging)
>  - bug 821 (AODV logging)
>
> (Note:  796 was fixed by the patch to bug 806, and could be listed below)
>
> Do any maintainers have other bugs that they feel should be in a minor 
> bugfix release for 3.7?  Do maintainers want to _not_ include any of 
> the above because they are destabilizing at this time?  Note that I am 
> probably most concerned about the AODV patch and the 816 patch, but I 
> don't know whether the Wifi ones are considered also important fixes.
I can postpone the release of ns-3.7, the thing is that the bug fixes 
are causing errors in automatic builders, so may be we should wait more 
until the builds are stable enough to commit this to ns-3.7.1
>
> 2) RELEASE_NOTES should have an itemized list of all bugs fixed, so 
> ns-3.7 users can decide whether they care about this release
>
I will take care of that.
> 3) Do we have any test experience with this release yet?  Can we get 
> some sign-off or verification that the eventual RC passes all testing 
> (static builds, debug, optimized) on our supported platforms?
I made ns-regression buildbot run ns-3.7 and it looks ok for the moment.
>
> 4) we discussed before releasing a patch from 3.7 to 3.7.1 but there 
> wasn't any strong consensus.  What advice do we give ns-3.7 users 
> about this release?  What I had in mind was:
>   - prepare also a patch from 3.7 to 3.7.1  - create a wiki page "How 
> to upgrade your code"
>     -- first, decide by reading the RELEASE_NOTES if you should care 
> about this patch
>     -- if so, cp -r your directory to a scratch place
>     -- apply the patch, and make sure you are happy with how the code 
> now behaves before abandoning your old code
>
> We could also dispense with releasing/distributing patches and create 
> on this wiki page a step "how to make an ns-3.n to ns-3.n+1 patch" if 
> people do not want to get into the business of releasing patches
>
The first idea would be easier for users if they didn't interfere with 
files involved in the patch. Otherwise I think they can just pull and 
merge using mercurial after saving their work (just to be on the safe side).
Best regards
Faker Moatamri
> Thanks,
> Tom
>