|
| From: | John N. Armstrong |
| Subject: | Warning messages not understood |
| Date: | Fri, 24 Nov 2000 12:24:44 -0600 |
|
I’m not sure that this is a bug (in
fact, I’m certain it’s not), but I can’t understand the behavior I’m getting
from GNU make 3.79 running on an NT 4.0 SP6 system. I have a “typical” recursive make
file structure that I’m using to build a Java application. The top-level make
invokes make in each of the sub-directories, which invoke make in their
sub-directories, and so on. In every case, I’m invoking make using the “$(MAKE)
–C <sub-directory> <target>” syntax. The makefile in each sub-directory
includes a global makefile (by means of the include directive) with common variable
definition and a default rule for making Java class files from Java source files. Not surprisingly, all the makefiles are
very similar in structure; for example, each makefile has a pseudo-target
named, “clean:”. As each makefile is processed, I get
the following type of warning messages: makefile:<line-no> warning: overriding commands for target
‘clean’ makefile:<line-no> warning: ignoring old commands for
target ‘clean’ I receive a pair of these messages
for each and every pseudo-target that happens to have the same name as a
pseudo-target in a higher-level makefile. I don’t understand this. I’ve not
seen this behavior on other systems running GNU make (same release). I thought
each invocation of make (i.e. for a sub-directory) was essentially independent
of other (higher-level) invocations. However, these warning messages give me
the impression that a common “rule space” is being used, and when make
encounters a pseudo-target previously seen, its commands temporarily “overlay”
any previous commands associated with that pseudo-target. This happens irrespective
of whether I run make in UNIX mode (using the –unix command line option) or
Win32 mode (using the –win32 command line option). Other than the fact that I get lots
of these warning messages each time I run make, everything appears to work OK –
that is, the end result is correct. I can find no information in the
documentation as to why this happens, or how this behavior can be suppressed. Can you enlighten me on what’s going
on here? Thanks! John John N. Armstrong, Sr. Partner Mobile 972-529-0195 Armstrong, Sileven & Associates, L.L.P. |
| [Prev in Thread] | Current Thread | [Next in Thread] |