The egcs vs. gcc fiasco comes to mind, but IIRC there have been a number of major reworkings. Certainly, there is an unbroken lineage from the original release to the present day - that's indisputable. Equally, it's indisputable that GCC is one of the most popular and powerful compilers out there. By these metrics, the claims are entirely correct.
However, having said that, the modern GCC wouldn't pass the "heraldry test" and there have been more than a few occasions when politics have delayed progress or disrupted true openness. The first of these is really a non-issue unless GCC applies for a coat of arms, but the second is more problematic. As GCC grows and matures, the more politics interferes, the more likely we are to see splintering.
Indeed, rival FLOSS compiler projects are taking off already, suggesting that the splintering has become enough of a problem for other projects to be able to reach critical mass.
Personally, I'd like to see GCC celebrate a 50 year anniversary as the top compiler. Language frontend developers can barely keep up with GCC, they won't be able to keep up with other compilers as well. Maximum language richness means you want as few core engine APIs as possible where the APIs have everything needed to support the languages out there. GCC can do that and has done for some time, which makes it a great choice.
But the GCC team (and the GLibC team) could do with being less provincial and more open. Those will be key to the next 25 years.