[llvm-dev] [GSoC 2016] Proposal: CFL-AA by default

[John Criswell via llvm-dev]

On 3/25/16 1:08 PM, Jia Chen wrote:
> Oops thanks for the reminder. I did use another email address and am 
> terribly sorry for it.

Are you working in Calvin Lin's group at Austin?

>
> Let me just post the link:
> https://docs.google.com/document/d/1Kvepb-v5Ta8ug_lLK1kZeexPNlpvj62K5iIF0fMuLyM/edit?usp=sharing 
>

There are a few suggestions that I would make:

 1. Quantify the number of cfl-aa bugs that exist, provided that doing
    so is possible and relatively a low amount of work.  It's not clear
    how substantial the project is without knowing just how broken
    cfl-aa is.
 2. Mention that you will either add more functionality tests or
    regression tests as part of your work.
 3. Can you expand on your involvement with the two projects listed in
    your biography?  What was your role in the Anderson's points-to
    analysis?  I thought Ben Hardekopf did most of the work on that
    project.  Was this follow-up work, a separate implementation, or
    something else?  As for the taint tracker, you should clarify
    whether it was static or dynamic (or a hybrid) and explain for what
    it was used.


Regards,

John Criswell

>
> On 03/25/2016 11:44 AM, John Criswell wrote:
>> Dear Jia,
>>
>> I don't see your proposal in the system.  What is the title of your 
>> proposal, and under what email address did you submit it?
>>
>> Regards,
>>
>> John Criswell
>>
>> On 3/23/16 4:09 PM, Jia Chen via llvm-dev wrote:
>>> Dear llvm devs,
>>>
>>> Based on an earlier discussion about existing pointer analyses in 
>>> LLVM, I quickly hacked up a GSoC proposal on enabling cfl-aa by 
>>> default.
>>>
>>> The decision to write it was made two days before the application 
>>> deadline, hence the writing quality may not be very satisfactory: 
>>> the background section could be less verbose, and the implementation 
>>> section could be more formal. Also it doesn't seem that I am able to 
>>> come up with a tentative schedule since at this point, I have no 
>>> idea how long each of the objective is going to take.
>>>
>>> I've shared my draft through the GSoC website. Any ideas on how I 
>>> could improve are welcomed. Thanks in advance to everyone who is 
>>> willing to take his/her time to read it.
>>>
>>
>>
> - Jia


-- 
John Criswell
Assistant Professor
Department of Computer Science, University of Rochester
http://www.cs.rochester.edu/u/criswell

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20160325/2ec77c6a/attachment-0001.html>