| Submitter | Andrei Gherzan |
|---|---|
| Date | Jan. 11, 2012, 12:45 p.m. |
| Message ID | <1326285902-31525-1-git-send-email-andrei@gherzan.ro> |
| Download | mbox | patch |
| Permalink | /patch/19067/ |
| State | New |
| Headers | show |
Comments
On Wednesday 11 January 2012 14:45:02 Andrei Gherzan wrote: > Some upstream layers (ex: meta-oe) use '&&' or '||' syntax in LICENSE > variable. Because of my last commit, these bbfiles are failing as LICENSE > is tested with oe.license where '&&' and '||' are the bug source. > The fix is to replace multiple characters of '&' or '|' with a single one. Sorry to be a pain but could we not just fix those layers to use the accepted syntax? Doing so should be just a one-line find/sed command. Cheers, Paul
On 01/11/2012 03:56 PM, Paul Eggleton wrote: > On Wednesday 11 January 2012 14:45:02 Andrei Gherzan wrote: >> Some upstream layers (ex: meta-oe) use '&&' or '||' syntax in LICENSE >> variable. Because of my last commit, these bbfiles are failing as LICENSE >> is tested with oe.license where '&&' and '||' are the bug source. >> The fix is to replace multiple characters of '&' or '|' with a single one. > Sorry to be a pain but could we not just fix those layers to use the accepted > syntax? Doing so should be just a one-line find/sed command. > > Cheers, > Paul > That'd be a solution as well but because those files were accepted this way until now i thought that this is an accepted syntax. More, this would scale and give a little freedom to the developers especially for those that are coming from "C" part or the world. (smile) ag
On 01/11/2012 06:11 AM, Andrei Gherzan wrote: > On 01/11/2012 03:56 PM, Paul Eggleton wrote: >> On Wednesday 11 January 2012 14:45:02 Andrei Gherzan wrote: >>> Some upstream layers (ex: meta-oe) use '&&' or '||' syntax in LICENSE >>> variable. Because of my last commit, these bbfiles are failing as >>> LICENSE >>> is tested with oe.license where '&&' and '||' are the bug source. >>> The fix is to replace multiple characters of '&' or '|' with a single >>> one. >> Sorry to be a pain but could we not just fix those layers to use the >> accepted >> syntax? Doing so should be just a one-line find/sed command. >> >> Cheers, >> Paul >> > That'd be a solution as well but because those files were accepted this > way until now i thought that this is an accepted syntax. More, this > would scale and give a little freedom to the developers especially for > those that are coming from "C" part or the world. (smile) > I think I agree with Paul here, Beth has sent out a proposal for that and by making a change here may allow the bad behavior to continue. Sau! > ag > > _______________________________________________ > Openembedded-core mailing list > Openembedded-core@lists.openembedded.org > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core >
On 01/12/2012 09:19 AM, Saul Wold wrote: > I think I agree with Paul here, Beth has sent out a proposal for that > and by making a change here may allow the bad behavior to continue. > > Sau! In this case we can drop this and the problem will remain (will have to be fixed) in meta-oe layers. @g
Patch
diff --git a/meta/lib/oe/license.py b/meta/lib/oe/license.py index 7ab66e7..5b8d47c 100644 --- a/meta/lib/oe/license.py +++ b/meta/lib/oe/license.py @@ -18,6 +18,7 @@ license_pattern = re.compile('[a-zA-Z0-9.+_\-]+$') class LicenseVisitor(ast.NodeVisitor): """Syntax tree visitor which can accept OpenEmbedded license strings""" def visit_string(self, licensestr): + licensestr = re.sub(r'([|&])\1+', r'\1', licensestr) new_elements = [] elements = filter(lambda x: x.strip(), license_operator.split(licensestr)) for pos, element in enumerate(elements):