| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA256 |
| 3 |
|
| 4 |
On 01/09/14 08:02 AM, Micha³ Górny wrote: |
| 5 |
> Dnia 2014-08-31, o godz. 14:14:20 Pacho Ramos <pacho@g.o> |
| 6 |
> napisa³(a): |
| 7 |
> |
| 8 |
>> This tries to solve: |
| 9 |
>> https://bugs.gentoo.org/show_bug.cgi?id=519682 |
| 10 |
>> |
| 11 |
>> Looks to do the job but maybe there are shorter or saner ways of |
| 12 |
>> doing it |
| 13 |
> |
| 14 |
> There are no sane ways of doing this. For a sane alternative, take |
| 15 |
> a look at python_gen_any_dep(). |
| 16 |
> |
| 17 |
> Long story short, define a list of allowed variables and do: |
| 18 |
> |
| 19 |
> x=${x//\${PF}/${PF}} x=${x//\${PN}/${PN}} |
| 20 |
> |
| 21 |
> etc. |
| 22 |
> |
| 23 |
|
| 24 |
...we could do that, couldn't we? If the list of variables expanded |
| 25 |
is limited to a standard expected list (ie all one sees in 'man 5 |
| 26 |
ebuild' plus say, MY_{P,PN,PV} as those are also fairly common. |
| 27 |
|
| 28 |
I've never compared the efficiency of bash substitution compared to |
| 29 |
say, pipe'ing through awk or sed, though, so i don't know if this |
| 30 |
method would be better or worse.. |
| 31 |
|
| 32 |
|
| 33 |
-----BEGIN PGP SIGNATURE----- |
| 34 |
Version: GnuPG v2 |
| 35 |
|
| 36 |
iF4EAREIAAYFAlQF9q4ACgkQ2ugaI38ACPDSegEArBPDRoWSjbHEp0Khu746+cFO |
| 37 |
6JMYviZX2bchXrTNprgBALaY42zlej0hwEjGcRyUs96oDINJxu3ZKArWbPDLIWYh |
| 38 |
=6vGm |
| 39 |
-----END PGP SIGNATURE----- |
Archives