Date: Thu, 12 Sep 1996 16:59:47 +0200 (IST) From: Eli Zaretskii To: Robert Hoehne Cc: djgpp-workers AT delorie DOT com Subject: Re: Make 3.74 diffs In-Reply-To: Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII On Thu, 12 Sep 1996, Robert Hoehne wrote: > 1) You are using the external variables 'dos_status' and 'dos_command_running', > but 'dos_command_running' is nowhere defined and 'dos_status' is defined > only as static. These are only minor bugs, which can be solved with my > patch for 'job.c' or you can do it in an other way. > > 2) The next bug in job.c is much bigger which has to do with 'dos_bname', > please read my comments in the diff. These are not bugs, but just consequences of me being in a hurry. I'm almost sure I ran diff on a wrong set of sources. In particular, the entire fragment that deals with `dos_bname' and friends is nuked, because Make doesn't work this way anymore: it never produces a batch file. I'll most probably upload the final alpha version of the sources early next week. > 3) And the last thing has to do with your implementation of setting $SHELL. > I'm not aggree with you, to set this variable automatically every time. > I think, it should be set only in your way, when it is set explicit in > the makefile, because if you set it every time, all the makefiles are > broken, which do not set this variable and assume, that $SHELL points > to the default shell and most commands are executed without the shell. > But if you set this variable, it is in most case different from the default > shell and so every command will be executed by the shell. This is a very important point, but unfortunately I don't understand exactly what are you saying. Can you please elaborate, or give an example of a Makefile that is broken by the way Make works? I built quite a few GNU packages with that Make and never saw any problem. Maybe there is something specific in your system setup that causes problems? And of course, thanks for testing the new version.