1 |
On Tue, 12 May 2009, Frank Hellmuth wrote: |
2 |
> On Monday 11 May 2009 13:43:44 François Bissey wrote: |
3 |
> > Hi all, |
4 |
> > |
5 |
> > For anyone interested in using opendx I just pushed a new revision |
6 |
> > of it in the overlay along with a new revision of opendx-samples. |
7 |
> > [...] |
8 |
> > There is still a minor issue to solve properly with szip but I would like |
9 |
> > some feedback from the current change by interested users before |
10 |
> > going further, I also spent a week perusing opendx code and need a small |
11 |
> > break from it. |
12 |
> |
13 |
> Hi, |
14 |
> |
15 |
> thanks for your hard work! |
16 |
> |
17 |
> I think there is a problem with parallel make processes. If I try to emerge |
18 |
> opendx-4.4.4-r2 with MAKEOPTS="-j3" it fails with |
19 |
> |
20 |
> [...] |
21 |
> Making all in dxmods |
22 |
> make[3]: Entering directory |
23 |
> `/var/tmp/portage/sci-visualization/opendx-4.4.4- |
24 |
> r2/work/dx-4.4.4/src/exec/dxmods' |
25 |
> echo MYINC: |
26 |
> rm -f dxcm.mdf |
27 |
> MYINC: |
28 |
> rm -f dx.mdf |
29 |
> cp ./dxmdf.src tmp.c |
30 |
> cp ./dxmdf.src tmp.c |
31 |
> x86_64-pc-linux-gnu-gcc -E -I../../../include -I./../dpexec - |
32 |
> I/var/tmp/portage/sci-visualization/opendx-4.4.4-r2/work/dx-4.4.4/include - |
33 |
> Dlinux -DHAVE_CONFIG_H tmp.c > dxcm.mdf |
34 |
> x86_64-pc-linux-gnu-gcc -E -I../../../include -I./../dpexec - |
35 |
> I/var/tmp/portage/sci-visualization/opendx-4.4.4-r2/work/dx-4.4.4/include - |
36 |
> Dlinux -DHAVE_CONFIG_H -DCPLUSPLUS_UI -DEXECUTIVE_MODS tmp.c > dx.mdf |
37 |
> rm -f tmp.c |
38 |
> cc1: error: tmp.c: No such file or directory |
39 |
> make[3]: *** [dx.mdf] Error 1 |
40 |
> make[3]: *** Waiting for unfinished jobs.... |
41 |
> [...] |
42 |
> |
43 |
> With MAKEOPTS="-j1" it compiles and installs fine. |
44 |
> |
45 |
Hi Frank, |
46 |
we already have a patch for that but for some reason it doesn't always work. |
47 |
And I think I know why and I thought I had it solved but obviously it wasn't |
48 |
enough. I will see if the solution I think for it works in the next 12 hours. |
49 |
For info the patch is concurrent-make-fix and it use to apply to local.mk |
50 |
unfortunately this file is meant to be generated from local.mk.in so it can |
51 |
be overwritten I changed the patch to local.mk.in but that's obviousvly not |
52 |
enough to ensure it is always generated correctly. I will remove the extra |
53 |
local.mk that shouldn't be in the tarball in the first place and see if it |
54 |
really works without problem. |
55 |
|
56 |
Cheers, |
57 |
Francois |