[FFmpeg-devel] How to handle compiler warnings

[Mike Melanson]

Roberto Togni wrote:
> On Tue, 29 Jan 2008 11:45:15 +0100
> Diego Biurrun <diego at biurrun.de> wrote:
> 
>> The topic has come up again, it's time to discuss the subject.  I
>> propose to try to avoid compiler warnings as much as possible in order
>> to
>>
>> - have cleaner code,
>> - have important warnings not be drowned out,
>> - make FFmpeg a programming textbook.
>>
>> This does not include warning fixes that slow things down or obfuscate
>> the code, but if in doubt I personally would err on the side of fixing
>> the warning.
>>
> 
> Ok for me as long as:
> - if there are useless warning, they are disabled instead of fixed
> - unclear cases are discussed first so that silencing a warning does
> not create a bug
> - we clearly decide which warnings must be fixed, since a warning-less
> compilation is a moving target (compiler and platform dependant)

I wanted to mention some future developments for the FATE Server (
http://fate.multimedia.cx/ ) regarding warnings:

* Several people have requested that, at the very least, the database
summarize stats about how many lines of stderr text are in each build.
This is planned for the short term.

* As always, pretty graphs are planned to help visualize warning trends.

* I want to build a diffing tool that will make it easy to identify new
warnings that creep in over time.

* Correlate warnings that occur across most or all configurations.

Other ideas welcome.

-- 
	-Mike Melanson