Show quoted text 12 lines
On Sun, 22 Jun 2003 01:11:56 +0200 (CEST)
(Marco van de Voort) wrote:
> shell already, which is enough to run standard scripts). And if they do,
> it is somewhere else, not in /bin.
>
> linux.shell() takes care of this IIRC.
You are right.
But if he really wanted an OS/Distribution independent solution, he would
use FindFirst/Next instead of 'ls'. Probably this is for the theseus
project, which is in fact Linux only.
One shouldn't do things Linux only if not absolutely needed so. Do like you develop for Unix, and leave the "getting things to run on *BSD" to the BSD porter if it really happens.
Believe me, 90% of the maintenance on the FPC BSD ports is just changing ifdef linux to ifdef Unix, and changing other basic things. (like hardcoded paths, assumptions that are dangerous under Linux too with all those distro's)
P.s. Lazarus runs under NetBSD/x86. It loads awfully slow, but after 45 secs of starting up it is usable and feels "quick" on this P-III 450 with 384M. I ran a project and it started in the debugger but terminated with SIGBUS, which was presented to me by the IDE (so the gdb interaction is ok).
The problem seems to be that libintl and libgmodule weren't auto linked in for some reason, however I must yet confirm this with 100% certainty by doing some more test builds.
The SIGBUS above is probably the same problem for the compiled demo, as described above for lazarus.