Re: Make segfaults with attached sources

[John Pye]

Hi Steve,

There certainly were some bugs in that makefile of mine, the problem is
that Make is failing to notice on my bad syntax.

In particular, I was attempting to do pattern substitution on an unnamed
variable, and to concatenate that invalid variable substitution with
another string. Perhaps that is the problem. I'm not about to use beta
build of make though, as that would be too much fiddling to take on
board. I will just have to accept that certain syntax errors in my
makefile will cause make to crash, and try not to make those errors.

I'll look forward to the next release though! [more hinting!]

Cheers
JP

Steve James wrote:

>Hello John,
>
>I concur: that Makefile segfaults with make 3.80. However, you may like to 
>know that it doesn't upset 3.81beta3 (linux and MinGW builds). I expected 
>this might be the case because I've been through the same experience: one 
>day, I added another innocuous line to my make script, and it just blew up. 
>So I don't know what the specific bug is, just that it's gone away now in 
>3.81beta3. We now have no choice but to use a beta build of make in 
>production. I'm looking forward to the next stable release [hint, hint].
>
>Cheers,
>Steve
>
>On Friday 19 August 2005 07:05, John Pye wrote:
>  
>
>>Hi there
>>
>>I've just had my first ever segfault from GNU Make, with the attached
>>sources which I've been writing. I just added the bits about 'TESTS' and
>>'TESTOBJS' etc in the make file, so it's simeothing which I've done
>>there which is crashing perhaps it's my attempt to use
>>$(VAR:%.test:%.test.o) type patterns more than once in a line.
>>
>>I've reproduced this bug on FC3 and cygwin with version 3.80 in both
>>cases. Hope the following will help a bit with tracking down the problem.
>>
>>    
>>
>
>--snip--
>
>  
>