brunner@bullhead.uucp (08/09/89)
I'm sorry not to have replied to the mail sent earlier, but I both wanted to make one reply to several people. Please don't think I lost or ignored your note (unless you don't see it paraphrased here). Eric Brunner, IBM AWD Palo Alto inet: brunner@monet.berkeley.edu or brunner%ibmsupt@uunet.uu.net uucp: uunet!ibmsupt!brunner (415) 855-4486 ---------------------------begin--------------------------- Steve Dyer (uunet!ursa-major.SPDCC.COM!dyer) asked if the floating point patch, orginally distributed as a diskette, would be distributed via the newsgroup, the answer is yes - I'm waiting on a related fix from HCR to the FORTRAN compiler. Dave Fetrow, Yuval Tamir and Peter Honeyman pointed out that an anonymous (internet) ftp site would be a further improvement. Anonymous ftp would be convient for some, but a) IBM Palo Alto is not directly connected to the Internet for policy reasons, and b) some of what is or will be on IBMSUPT is non-IBM source in non-diff form, which means that we couldn't keep the "anonymous" (ftp) and "known IBM customer" (uucp) archives identical... I'm discussing this with management. Keith Moore (uunet!cs.utk.edu!moore@cygnusx1.cs.utk.edu) asked if the fixes would be posted to comp.sys.ibm.pc.rt, by now he probably knows the answer is yes. Adam Feigin (uunet!GARP.MIT.EDU!rochester!arclight!adam) suggested that the volume and issue number be included in each fix some where near the begining of the message. The keyword in each posting has the volume and issue number, viz. Keywords: V1.1 Keywords: V1.2 Keywords: V1.3 asych multiport driver Keywords: V1.4 quotas Melbourne Keywords: V1.5 resource map panic Keywords: V1.6 profile Keywords: V1.7 gprof Keywords: V1.8 security shell script Keywords: V1.9 chown security Greg Rogers (!uunet!cs.uiuc.edu!grogers) also wrote asking if there were keywords in my postings of 24 July. If it would be helpful that I place them elsewhere in the postings please let me know. Carl Braganza (uscacsc.usc.edu!carl) recommended that the RCS identifier be included in the context diffs, I wrote a long note on this for internal circulation -- not because I am "anti-RCS" (quite the converse!), but because a) RCS was not distributed, b) the diffs would then include some portion of the copyright which I wish to avoid, c) the context diffs, our use of RCS, and your use of patch ensures that there is no ambiguity as to what has or hasn't been fixed or whether there is some "unofficial" fix in place, d) the rcsid actually provides no information to the ACSC about the source unless assumptions are made, and e) anyone using RCS and having revision levels equal to or higher than ours would be clobbered. Still, the jury is out -- the main thing is that we get the fixes out the door. The simpler we can make it the better. Send your comments! Adam (ibmsupt!uunet!GARP.MIT.EDU!rochester!arclight!adam) wrote that patch detected a previously applied patch when he applied V1.7 which fixed a problem with fpglue.s. The fix to fpglue.s was part of the corrective service diskette. Adam also asked about UCBFIX being defined in ufs_syscalls.c, it is a configuration option in /sys/conf/GENERIC. ---------------------------end---------------------------
grogers@m.cs.uiuc.edu (08/10/89)
>/* Written 7:28 pm Aug 8, 1989 by brunner@bullhead.uucp in m.cs.uiuc.edu:comp.sys.ibm.pc.rt */ > > Keywords: V1.1 > Keywords: V1.2 > Keywords: V1.3 asych multiport driver > Keywords: V1.4 quotas Melbourne > Keywords: V1.5 resource map panic > Keywords: V1.6 profile > Keywords: V1.7 gprof > Keywords: V1.8 security shell script > Keywords: V1.9 chown security The ibm.apars notesfile, where bugs are annouced and fixes posted, also included a new version of hc and bind-4.8. Since your list didn't cover these items, I must ask, why not? I'm unsure as to whether I should consider your postings as "official" in any sense. greg