| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA1 |
| 3 |
|
| 4 |
On 06/02/2012 10:08 PM, Mike Frysinger wrote: |
| 5 |
> # @FUNCTION: _multijob_fork # @INTERNAL # @DESCRIPTION: # Do the |
| 6 |
> actual book keeping. _multijob_fork() { [[ $# -eq 1 ]] || die |
| 7 |
> "incorrect number of arguments" |
| 8 |
> |
| 9 |
> local ret=0 [[ $1 == "pre" ]] && : $(( ++mj_num_jobs )) if [[ |
| 10 |
> ${mj_num_jobs} -ge ${mj_max_jobs} ]] ; then multijob_finish_one |
| 11 |
> ret=$? fi [[ $1 == "post" ]] && : $(( ++mj_num_jobs )) return |
| 12 |
> ${ret} } |
| 13 |
|
| 14 |
The "pre" logic seems wrong. Consider an initial state of |
| 15 |
mj_num_jobs=0 and mj_max_jobs=1. It will increment mj_num_jobs to 1, |
| 16 |
so [[ 1 -ge 1 ]] is true, and then call multijob_finish_one even |
| 17 |
though no jobs have started yet? Wouldn't that deadlock |
| 18 |
multijob_finish_one, as it waits for a reply from a job that doesn't |
| 19 |
exist yet? |
| 20 |
- -- |
| 21 |
Thanks, |
| 22 |
Zac |
| 23 |
-----BEGIN PGP SIGNATURE----- |
| 24 |
Version: GnuPG v2.0.19 (GNU/Linux) |
| 25 |
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ |
| 26 |
|
| 27 |
iEYEARECAAYFAk/L4j4ACgkQ/ejvha5XGaPyuQCfSHRUHA1KoVc97yRZa8FlF+TS |
| 28 |
n04An1/c7IQaH4mqUtm8P305WKKDOgvE |
| 29 |
=EgJz |
| 30 |
-----END PGP SIGNATURE----- |
Archives