| |||||||||||||
GCC and static analysisGCC and static analysisPosted Apr 21, 2012 21:15 UTC (Sat) by mitchskin (subscriber, #32405)Parent article: GCC and static analysis
Is this really a surprise? RMS's resistance to a GCC plugin system was intense and long-lived. LLVM/clang, on the other hand, tried to be modular and pluggable from the very beginning. One might naturally expect that LLVM would be a better platform for this sort of thing, even if GCC did finally start to soften its stance on plugins; GCC's architecture isn't going to change wholesale overnight.
(Log in to post comments)
GCC and static analysis Posted Apr 21, 2012 22:59 UTC (Sat) by jpnp (subscriber, #63341) [Link]
Google have also shown some resistance to GPL projects. When GCC was the only game in town they'll use it, but recently CLang has gotten to the stage of being a feasible alternative. I'd imagine that, even if they considered both GCC & LLVM technically equal to the task, they'd favour the non-GPL project.
GCC and static analysis Posted Apr 22, 2012 15:15 UTC (Sun) by mbligh (subscriber, #7720) [Link]
Google is a company run by its engineers. You'll find a widely diverse set of opinions and attitudes and licensing - it depends more on the individuals involved than some corporate stance.
GCC and static analysis Posted Apr 23, 2012 3:42 UTC (Mon) by dberlin (subscriber, #24694) [Link]
Yes, google has no official corporate stance, except that we support both projects.
GCC and static analysis Posted Apr 23, 2012 23:35 UTC (Mon) by marcH (subscriber, #57642) [Link]
Any big company has a widely diverse set of opinions and attitudes. The only difference between companies is how much of this diversity is hidden/controlled.
GCC and static analysis Posted Apr 25, 2012 10:15 UTC (Wed) by wahern (subscriber, #37304) [Link]
If you read the actual explanation of the issues (follow the link in the article), you'd realize that your analysis bares no relationship whatsoever to the real reasons they switched. He lists several requirements, and GCC by his estimation scored better than LLVM/Clang for most of them. The only issue issue is that the areas where LLVM/Clang excelled happened to be far more important for those particular engineers wrt to their particular project.
|
|||||||||||||