No, it was not obvious from Mirars text that this was the kind he had in mind. Only that it would involve `sprintf("%f",exp(511))' (and possibly `sprintf("%e",exp(511))' and/or `"8.404e+221"') somewhere.

/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)

Previous text:

2003-01-05 13:59: Subject: exp(511)

---

The test_true(stringp(sprintf("%f",exp(200)))) kind, obviously.

/ Martin Nilsson (bygger parser

Hmm. My data point:

exp(101);

(1) Result: 73070599793680668906492145854235224565612544.000000

exp(102);

Segmentation fault

Pike 7.4.13 on Linux 2.4.17 i686, glibc-2.3.1

The segfault is due to sprintf("%O",...); write("%e", exp(x)) is OK up to x=710 (at which point it turns "inf" of course).

Aah, nearly forgot: --with-long-double-precision.

/ rjb

Previous text:

2003-01-05 21:37: Subject: exp(511)

---

I'm happy as long as it doesn't segfault. It probably has a problem because I'm using double precision (up to around e^180 is ok).

I added tests to check that it returns strings for medium (e^200) and large (e^65536) for all five float formats.

/ Mirar