| 1 |
On Fri, Nov 11, 2005 at 11:47:17PM -0500, Alec Warner wrote: |
| 2 |
> There is a problem with DepSet in 3.0, it doesn't parse dependencies |
| 3 |
> properly in the case of dev-php/php-4.4.0-r1. |
| 4 |
> |
| 5 |
> This is using pkg.get.data("RDEPEND", "") |
| 6 |
pkg.data["RDEPEND"] I assume (no get method). |
| 7 |
|
| 8 |
> Package deps: |
| 9 |
> berkdb? ( =sys-libs/db-1* || ( >=sys-libs/db-4.0.14-r2 |
| 10 |
> >=sys-libs/db-3.2.9-r9 ) ) |
| 11 |
|
| 12 |
<leave only what's relevant please, scanning for berkdb is easy, but |
| 13 |
people are lazy> |
| 14 |
|
| 15 |
> Notice the correct || ( bla foo ) dependency with berkdb? |
| 16 |
> |
| 17 |
> Now I switch to pkg.rdepends(): |
| 18 |
> Package rdeps: |
| 19 |
> berkdb? ( =sys-libs/db-1* ( >=sys-libs/db-4.0.14-r2 || |
| 20 |
> >=sys-libs/db-3.2.9-r9 ) ) |
| 21 |
|
| 22 |
> Notice now that the berkdb? depend is ( bla || foo ) which is now |
| 23 |
> invalid, causing StateGraph to go bonkers. |
| 24 |
*cough* StateGraph losing it's marbles isn't due to this. |
| 25 |
|
| 26 |
>>> d=DepSet("|| ( a b c )") |
| 27 |
>>> print str(d) # yes that's redundant, making it clear __str__ is involved |
| 28 |
( a || b || c ) |
| 29 |
|
| 30 |
An OrRestriction is an ordered list of restrictions, short circuit or |
| 31 |
searching for a True... in python the equivalent code would be |
| 32 |
|
| 33 |
>>> if a or b or c: |
| 34 |
|
| 35 |
the || placement is just a difference in how __str__ outputs... it's |
| 36 |
not a bug, it's a difference in how it's presented for str output. |
| 37 |
Internally, DepSet parses it correctly. |
| 38 |
|
| 39 |
~harring |
Archives