vjs@calcite.UUCP (Vernon Schryver) (08/07/89)
Am I the only person in the world who thinks that make(1) should be compiled
so that '-d' works? Does no one else have enough memory or commit obscure
errors that would be illuminated by `make -nd`? `Make -p` just is not enough
for my mistakes.
I guessed history was a good enough reason make(1) was built small in Microport
3.0e. Finding the same restriction in ISC 2.0.2 has made me wonder if someone
at AT&T is asleep at the helm. (Presumably, ISC just uses cmd/make/make.mk
as it comes from New Jersey.)
Vernon Schryver
vjs@calcite.uucp or ...{pyramid,sgi}!sgi!calcite!vjsjohna@haddock.ima.isc.com (John Adams) (08/07/89)
In article <58@calcite.UUCP> vjs@calcite.UUCP (Vernon Schryver) writes:
-Am I the only person in the world who thinks that make(1) should be compiled
-so that '-d' works?
-I guessed history was a good enough reason make(1) was built small in Microport
-3.0e. Finding the same restriction in ISC 2.0.2 has made me wonder if someone
-at AT&T is asleep at the helm.
Let's see. The latest documentation I could find about -d was an old Sys III
manual. But I tried it anyway under ISC 2.0.2. It worked! What's your
complaint? Have you tried to use "make -nd"? Do you have some axe to grind
with Interactive?
John Adams (home) johna@grumpy.ima.isc.com (617) 646-6491
(work) johna@haddock.ima.isc.com (617) 661-7474vjs@calcite.UUCP (Vernon Schryver) (08/08/89)
I wrote: > Am I the only person in the world who thinks that make(1) should be compiled > so that '-d' works? Does no one else have enough memory or commit obscure > errors that would be illuminated by `make -nd`? `Make -p` just is not enough > for my mistakes. > >I guessed history was a good enough reason make(1) was built small in Microport > 3.0e. Finding the same restriction in ISC 2.0.2 has made me wonder if someone > at AT&T is asleep at the helm. (Presumably, ISC just uses cmd/make/make.mk > as it comes from New Jersey.) In article <14206@haddock.ima.isc.com>, johna@haddock.ima.isc.com (John Adams) writes: > Let's see. The latest documentation I could find about -d was an old Sys III > manual. But I tried it anyway under ISC 2.0.2. It worked! What's your > complaint? Have you tried to use "make -nd"? Do you have some axe to grind > with Interactive? > > John Adams (home) johna@grumpy.ima.isc.com (617) 646-6491 > (work) johna@haddock.ima.isc.com (617) 661-7474 The following is the first thing my 9-day-old copy of 2.0.2 says to `make -nd`. make: no debugging information available. Make does procede to do normal stuff, but that's not what I'm grumbling about. It should have babbled thousands of lines as it computed timestamps and considered rules. The source one finds on the SVR3.[012] tapes supports -d, if you compile it that way. It can be very useful for debugging hard makefiles, as well as proving bugs are in the makefile not make itself. In an NFS environment, I've found it invaluable. Vendors in this lowest end of the market should build the "big" version of make(1), as they now build large UUCP. In the high end of parallelizing make's, I don't know about nmake, but believe smake/pmake supports such debugging. I'm concerned by the attitude suggested by your posting. I have no more "ax to grind" with ISC than with any other concern that has accepted my money. My clumsy words, included above in their entirety unlike your partial quote, tried to cast blame on AT&T rather than ISC. Isn't haddock.ima.isc.com where the ISC customer support people work? Is your attitude an official representation of the concern Interactive Systems Corportation has for the happiness of its customers? Even Microport and Bell Tech never jumped on customers, no matter how stupid and offensive we are. All Other ISC employees whom I've recently met on the telephone or by email, as well as others now at ISC I've known in the business up to 20 years, have been friendly, helpful, and polite. They have provided very useful leads for solving other difficulties I've had with 2.0.2. Vernon Schryver vjs@calcite.uucp
johnk@opel.uu.net (John Kennedy) (09/14/89)
I have a very large makefile with many .SUFFIXES and inference rules.
About three quarters of the way through the make, make(1) bombs
with "hash table overflow"..
1) is there any way to tell make to use more space for its tables?
2) is there any aftermarket make product that has more capacity and
will run on Microport Sys V/AT, rel 2.4?
Thanks in advance,
John
--
John Kennedy johnk@opel.uu.uunet
Second Source, Inc.
Annapolis, MDtoy@ecse.rpi.edu (Ray Toy) (09/16/89)
In article <276@opel.uu.net> johnk@opel.UUCP (John Kennedy) writes: >I have a very large makefile with many .SUFFIXES and inference rules. > >2) is there any aftermarket make product that has more capacity and > will run on Microport Sys V/AT, rel 2.4? > You might want to try using GNU make. I have gnumake ver 3.48 (?) running and it works most of the time. It does core dump on occasion and sometimes incorrectly executes the commands, but it is usable. You can get a copy from prep.mit.ai.edu or tut.cis.ohio-state.edu, I think. Hope this helps. Ray -- Ray ----> Toy toy@ecse.rpi.edu