| 1 |
On Mon, Sep 24, 2012 at 01:05:35PM -0400, Mike Frysinger wrote: |
| 2 |
> On Monday 24 September 2012 06:25:35 Gregory M. Turner wrote: |
| 3 |
> > On cygwin, there is a problem with bi-directional pipe support in bash. |
| 4 |
> > |
| 5 |
> > I used to solve this with an ugly reversion in portage and an |
| 6 |
> > ultra-simple stubbification patch for multiprocessing.eclass (both |
| 7 |
> > serialized everything). |
| 8 |
> > |
| 9 |
> > However, this really sucked for numerous reasons, including the obvious |
| 10 |
> > one: it makes stuff slow as hell. |
| 11 |
> |
| 12 |
> if cygwin sucks, it doesn't get parallel jobs. add stubs to the end of the |
| 13 |
> file to disable parallel support. |
| 14 |
> |
| 15 |
> if is_cygwin ; then |
| 16 |
> numjobs() { return 1; } |
| 17 |
> multijob_init() { return 0; } |
| 18 |
> ...etc... |
| 19 |
> fi |
| 20 |
> |
| 21 |
> interleaving the cygwin logic otherwise makes it unmaintainable |
| 22 |
|
| 23 |
Seconded; if cygwin environment sucks for this, just suppress the |
| 24 |
parallelization for it. Better to be working then not, plus I doubt |
| 25 |
cygwin users are going to notice a huge diff in performance. :) |
| 26 |
|
| 27 |
~harring |
Archives