Re: -fdump-translation-unit considered harmful

[Stefan Strasser <sstrasser at systemhaus-gruppe dot de>]

Pjotr Kourzanov schrieb:
>   Hence my question: would GPL cover an executable that runs in a 
> separate address space, albeit having some sections /readonly pairwise 
> shared/ with GCC's cc1(plus)?

the problem some people may see isn't some technical way to link to gcc 
while circumventing the GPL. if this isn't covered by GPL by law an 
interface in gcc is the smallest problem of free software in general.
it is that you could easily export the data from the interface to a file 
and e.g. write a proprietary backend which only needs the file. this 
isn't covered by the GPL(compare loading an image by gimp into your 
proprierty app).
but I think this shouldn't be a reason to make gcc less useful for free 
software.

>   BTW, the GCC's backend(s) need not undergo dramatic patching, since 
> all they need to do is to pause when the tree has been built and wait 
> until the source-analyzing code has done its job 

as has been pointed out by robert using a seperate address space because 
of licensing isn't necassary.
so, I don't know the possible problems with gcc's current architecture, 
but a source code analysis tool invoking gcc and reading the output pipe 
obviously isn't the best design, and invoking gcc with argument which 
loads a dynamic library by the analysis tool, what you're trying to 
avoid, isn't either.

is there a reason for not making the front ends dynamic libraries which 
could be linked by any program that wants to parse source code?
gcc backend would be one of those programs.


btw it would be great if this interface is two-way, for translations. so 
you could link the frontend, parse code, modify the tree, and pass it to 
the backend. without the long way round of writing modified source code 
and invoking gcc.


--
Stefan Strasser