%% "Pomerleau, Francois" <[EMAIL PROTECTED]> writes:
pf> Hello, I am developing applications in Ada and I use make for my
pf> compilation. As part of my build process, I generate an Ada file
pf> (version.adb) that contains the build date and the release
pf> version. I was able to do so us
Definitely the next version of make should resolve all the EINTR
problems.
--
---
Paul D. Smith <[EMAIL PROTECTED]> Find some GNU make tips at:
http://www.gnu.org http://www.paulandlesley.
Fails also with 3.79.1. The code in that area didn't change, so I'm
not really surprised.
Some local folks have noted that this is an exposure on NFS filesystems
mounted with the "intr" flag. I assume this is a valid way to do mounts
so it's a real problem, but maybe explains why it isn't seen
Yes, I meant to mention that this was seen under NFS. And only on
solaris, though we do an equivalent amount of banging on NFS
from Linux and Alpha clients.
Kevin
Howard Chu wrote:
>
> I've seen this kind of problem before in other programs, but usually only on
> NFS-mounted filesystems. G
I've seen this kind of problem before in other programs, but usually only on
NFS-mounted filesystems. Generally on local UFS partitions the system calls
are atomic. It would be simpler if we could use sigaction() and set the
SA_RESTART flag for these signals, but the Solaris man pages don't mentio
Please get a copy of GNU make 3.79.1 (the latest version) and see if
that works any better.
I'm just finishing up a new release 3.79.2 which changes the way
jobserver works so EINTR is not a problem.
--
---
Paul D. Smi
%% Robert Bogomip <[EMAIL PROTECTED]> writes:
rb> Consider this makefile:
rb> t: m+=2
rb> t:
rb> @echo "[${m}]"
rb> Then the following sort of works:
rb> $ make t
rb> [ 2]
This has been fixed in CVS (it just gives "[2]" again).
rb> (Interestingly, make 3.78 produces
%% Jonathan Thornburg <[EMAIL PROTECTED]> writes:
jt> This is either a bug in how the patsubst function works in GNU Make
jt> 3.79, or a bug in the documentation for same:
jt>files := $(patsubst %, ../%/%.gfa, $(dirs))
jt> The GNU Make 3.79 info file says
>> `$(patsubst PATTE
%% "Albert L. Ting" <[EMAIL PROTECTED]> writes:
alt> Not sure if this is a solaris bug or a make bug, but it works
alt> fine in make-3.78.1. It appears make now treats targets with the
alt> same timestamp as the prerequisite as out of date. That didn't
alt> happen with make-3.78.1.
Actually, further investigation yields the following (what I consider to
be) inconsistency:
$ echo 'all: ;' | make -q -f- ; echo $?
0
$ echo 'all: ; $X' | make -q -f- ; echo $?
1
To me, that's wrong; you should get the same result with a completely
empty string as you do with a variable
%% "Albert L. Ting" <[EMAIL PROTECTED]> writes:
alt> Doing a "make -q" with the following makefile returns a 1 status
alt> where it should really return a 0 status. Can this be fixed? I
alt> know I could wrap an ifndef expression around the ALL commands,
alt> but I use this a lot.
al
%% Regarding make 3.79 status problem; you wrote:
alt> Doing a "make -q" with the following makefile returns a 1 status
alt> where it should really return a 0 status. Can this be fixed? I
alt> know I could wrap an ifndef expression around the ALL commands,
alt> but I use this a lot.
Ca
These are known problems with 3.79; I'm working on a 3.79.1 for these
and other minor glitches. You can find patches to fix this in the GNU
make bug database (see the README for a URL), PR/1700.
--
---
Paul D. Smith <[
%% Johnny Wales <[EMAIL PROTECTED]> writes:
jw> What about a 'make No_I_Really_Really_Really_Mean_It-install'
jw> option, or something along those lines that prints a warning, asks
jw> 'Are you sure you wish to do this here zany thing?' then installs
jw> to the -proper- (indicated by whic
What about a 'make No_I_Really_Really_Really_Mean_It-install' option, or
something along those lines that prints a warning, asks 'Are you sure you
wish to do this here zany thing?' then installs to the -proper- (indicated
by which) location? That'd be good. Reason being: I'm a programmer, so I
can
%% Johnny Wales <[EMAIL PROTECTED]> writes:
jw> Well, looks like the 3.79 install didn't quite work.
Yes, it did.
jw> I did a configure,
jw> make, make install, and everything went off without a hitch. Ah! I've just
jw> found the bug. Here's my uname -a:
jw> Linux vegas 2.2.13 #1 Tue
On Mon, 29 May 2000, Johnny Wales wrote:
> Well, looks like the 3.79 install didn't quite work. I did a configure,
> make, make install, and everything went off without a hitch. Ah! I've just
> found the bug. Here's my uname -a:
> Linux vegas 2.2.13 #1 Tue Nov 9 00:09:41 GMT 1999 i586 unknown
>
%% Darren Hiebert <[EMAIL PROTECTED]> writes:
dh> I did this and confirmed that cp -p does not preserve the nanosecond
dh> field in the copy.
Bummer! :(
Thanks for working this, though. I'll make a note of it somewhere. I
may have to make a configure option to disable nanosecond timestamp
Paul D. Smith wrote:
> Another thing you can do is write a little test program (you'll have to
> use C as I don't think perl, etc. support this field either) that shows
> the values of st_mtime and st_mtim.tv_nsec...returned by stat(2), then
> run it on a file, and then on a copy created with "cp
%% Darren Hiebert <[EMAIL PROTECTED]> writes:
dh> Paul D. Smith wrote:
>> Hmm. Are you using Irix "cp" to do this? If so, this may be a bug in
>> the Irix "cp" command. Maybe.
>> If you're using another cp, then that's probably your problem...
dh> Good idea. I checked and found, by
Paul D. Smith wrote:
> Hmm. Are you using Irix "cp" to do this? If so, this may be a bug in
> the Irix "cp" command. Maybe.
>
> If you're using another cp, then that's probably your problem...
Good idea. I checked and found, by putting explicit paths to cp in
the makefile that /usr/bin/cp, /
%% Darren Hiebert <[EMAIL PROTECTED]> writes:
dh> I encountered a strange bug when make-3.79 is used on IRIX 6.5,
dh> # Begin Makefile
dh> #DATE=-d "19 Dec 1994"
dh> b: a
dh> cp -p $< $@
dh> a:
dh> touch $(DATE) a
dh> # End Makefile
dh> Note that every time you run make
OK, I added those. Thanks for the report.
--
---
Paul D. Smith <[EMAIL PROTECTED]> Find some GNU make tips at:
http://www.gnu.org http://www.ultranet.com/~pauld/gmake/
"Please remain cal
%% "=?iso-8859-1?Q?H=E4gglund=2C_Ulf?=" <[EMAIL PROTECTED]> writes:
There's no need for HTML on the mailing lists. Please use plain text.
Thanks!
uh> GNU Make version 3.79, by Richard Stallman and Roland
uh> McGrath.Built for i686-pc-linux-gnu.
uh> I am trying to port a recursive Makefil
This is PR/1696. There's a patch attached as a resolution to the bug
report.
http://www-gnats.gnu.org:8080/cgi-bin/wwwgnats.pl/full/1696
--
---
Paul D. Smith <[EMAIL PROTECTED]> Find some GNU make tips at:
Paul D. Smith writes:
> Does BeOS have enough of a shell that you can run configure and get
> useful results?
Yes. It uses GNU bash 2.03.
> bh> 2) Error during "make distclean":
>
> bh> Apparently $(MAKE) = /bework/gnubuild/make-3.79/./make has been
> bh> removed during the "make distcle
%% Bruno Haible <[EMAIL PROTECTED]> writes:
bh> 1) Compilation error in arscan.c. BeOS doesn't have the
bh> header file. But it has archive libraries, and uses the GNU
bh> binutils to create them. I therefore copied the contents of
bh> glibc's into arscan.c, and verified using "od -x"
27 matches
Mail list logo
