tinycc-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Tinycc-devel] Request code review


From: jiang
Subject: Re: [Tinycc-devel] Request code review
Date: Wed, 04 Jun 2014 22:10:12 +0800
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.5.0

Thank you for your help, you really let me say these ideological emancipation.
However, I do have a lot to improve and optimize tcc patches, I hope you publish tcc before 0.9.27 join. I will commit as much as possible into smaller finer. These are my thoughts now.


By the way, I forgot to ask you about me add two assert () reasonable?


jiang

Hey jiang,

I think there's a broader issue at play here involving appropriate communication. I looked for some commentary about this on the internet but could not find it. I know I encountered it somewhere once, but I don't remember. I'll try to state it clearly.

Distributed version systems such as git are meant to minimize unneeded communication between developers. If I come up with a good idea for tcc and others agree, I don't have to send an email with the actual patch to the mailing list. I don't have to provide instructions for which patch sets you need to apply in order to apply my patch. This sort of communication---how do you use my patch---can be automated away. Git is very good at automating away this unnecessary communication.

However, communication is important on projects with multiple developers such as tcc. You are working on a community effort, and you want your code to help the community. It is important, therefore, to discuss your ideas with the community before moving forward with them. Maybe grishka already has some work on a feature you want. Maybe Thomas has some implementation ideas. Maybe Daniel could explain why none of tcc's code base uses assert. These ideas are worth discussing, and moving forward with your ideas without a discussion could potentially waste a great deal of your time. Until you've made a large number of contributions to the project, you should run any and all ideas by the community before implementing them.

So your first step is not to write any code yet. Your first step is to pick one (and for now, only one) idea that you would like to implement. Perhaps its the 64-bit register usage. Propose the idea. Summarize your idea in an email. Ask for feedback. If others think it's a good idea, implement it and ask for a code review, being sure to indicate where you began your work (so it's clear where the code review should begin).

In short, try to start off being less ambitious. :-)

David


reply via email to

[Prev in Thread] Current Thread [Next in Thread]