Re: gps(1)'s use of pseudo-ttys
Ludovic Brenta <[email protected]> Tue, 10 May 2005 15:42:24 +0200
| Newsgroups | gmane.comp.ide.gps.user |
|---|---|
| Message-ID | <[email protected]> |
writes: > Arnaud Charlet" wrote: >> This issue has been resolved in CVS some time ago already and in the >> future 3.0 release. >> >> If you're a GAP user, watch for the next GAP release. >> >>> Is there >>> some good reason why gps doesn't just fork/exec and front-end gdb? >> >> "just" ? and communicate with gdb how ? If the answer is "pipes", then >> pipes are not powerful enough to interact with gdb properly in some cases. >> >> Arno > > Yes, I meant - and should have written - "fork/exec/pipe". Thanks for the > answer, Arno. > > John A. Murdie I note that GPS has been changed in this department since the last academic release; now it uses /dev/ptmx if available. You may want to build GPS from the latest sources and see if it solves your security concerns. The GPS 2.1.0 on Debian does not exhibit the behaviour you described however. It still uses /dev/ptyz0 et al, though. -- Ludovic Brenta.