On 29 April 2014 12:30, Martin Decky <[email protected]> wrote:

> For generated code, insofar as it is used as a sort of intermediate
>> representation and not for programmers to look at, coding standards are
>> pretty much irrelevant.
>> Coding standards are for people, generated code is for compiler to parse.
>>
>
> I see your point. But let me add two of my points:
>
> a) If the code is generated, it should not be a major problem to
>    generate it in such a way that the compiler does not throw thousands
>    of warnings on it and that it is not a single module that takes
>    minutes to compile. I certainly don't care about the placement of
>    braces in generated code, but I do care about "debugability".
>
> b) The package of Jainja that Guillaume linked does not contain the
>    original Java code and the Java-to-C translator. It contains only the
>    generated C code, thus it is akin to a binary blob (it is an
>    intermediate code blob). This is perfectly fine for the coastline,
>    but not so fine for the mainline.
>
> This should be in no way a discouragement for Guillaume. It is only an
> explanation why I am not jumping the gun and integrating Jainja into the
> HelenOS mainline.
>
>
>

I agree on both points. I did not inspect the tarball too closely or try to
build it.
_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/listinfo/helenos-devel

Reply via email to