| Title: | DIGITAL UNIX (FORMERLY KNOWN AS DEC OSF/1) |
| Notice: | Welcome to the Digital UNIX Conference |
| Moderator: | SMURF::DENHAM |
| Created: | Thu Mar 16 1995 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 10068 |
| Total number of notes: | 35879 |
Well, it finally happened - I ran across a case where perl's Configure script actually failed, and failed spectacularly at that. On a newly-installed 4.0B system (a default installation just yesterday), all went fine until: Run make depend now? [y] sh ./makedepend make depend MAKEDEPEND= sh ./makedepend make depend MAKEDEPEND= sh ./makedepend make depend MAKEDEPEND= sh ./makedepend make depend MAKEDEPEND= sh ./makedepend And so on. If I let it run, it eventually fills the entire process table. It's clearly a recursive loop. I've done some digging around into ./makedepend, and so far haven't found a clue. As I dig some more, I thought that just maybe some other poor soul might have already solved the problem, so ... Has anyone gotten this to build on 4.0B? I also tried the 5.002 kit that I had lying about, and it failed the same way on 4.0B. On the other hand, both 5.002 and 5.003 build just fine on various 3.2[GH] systems that are lying about, and when I ftp the resulting /usr/local/* stuff over to the 4.0B system, it all runs without any apparent problems. But it'd be nice if I could build it on a 4.0B system (so we don't have to keep a 3.2 system around just for building perl. ;-) What the failures have in common is that they are running ./makedepend on a 4.0B system. In particular, they seem to be trying to make the $(FIRSTMAKEFILE) entry in the Makefile, and for some reason this produces infinite recursion with 4.0B's make, but not with earlier versions of make. What's going wrong isn't at all obvious. If I can figure out how to get to comp.lang.perl.misc from behind the firewalls, maybe I'll also ask there ...
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 9074.1 | perl 5.003 "makes" with no problem on 4.0B | SMURF::SEAGRAVES | Jim, Digital UNIX Tech. Partners. Eng. Grp.,381-6199 | Tue Mar 11 1997 14:37 | 7 |
the make depend works for me - which make are you using ? /bin/make is the one that is being used to make depend perl for me. | |||||
| 9074.2 | Binary files /bin/make and /sbin/make differ | APACHE::CHAMBERS | Wed Mar 12 1997 14:05 | 13 | |
Well, whaddaya know, I typed "which make" and it said "/sbin/make". I then typed "diff /*bin/make", and it gave me the above title message. So I said "set path = (/bin $path)" and typed a "make" command, and it works. The perl 5.003 release now seems to be happily compiling (and of course it'll pass all its tests ;-). So the version of make that is in /sbin is the culprit. It differs from /bin/make (which is actually /usr/ccs/bin/make) in some subtle way that causes it to fail for the perl Makefile. Both programs are the same size, but have different sums, so they are indeed different somehow. I wonder if it's worth reporting, and if so, to who[m]? | |||||
| 9074.3 | QUARRY::neth | Craig Neth | Wed Mar 12 1997 14:39 | 14 | |
As near as I can tell, they are intended to be the same (/sbin/make is built from the same sources as /usr/ccs/bin/make). A QAR would appear to be in order. FWIW, there are actually 4 different make executables shipped on the base system: /sbin/make /usr/bin/make aka /usr/ccs/bin/make /usr/bin/posix/make /usr/opt/ultrix/usr/bin/make You can always try another one if you don't like the one you've got... :-) | |||||