shoot, no dice on that approach

[ You must provide a valid url ]

password has a * in it though so that might be borking it as well..

this is an attempt at a patch for this issue on the plexis UrlSourceValidator

I couldn't locate the mechanism that plexus would use for a cleaner way of obtaining url, username and password from the https:// url string so I added a couple of scrapeX methods to the bottom.

it builds in plexus-formica just fine...I haven't tried it yet with my continuum build but I built it from the commandline testing version I was playing with to get the things right should it should be good to go.

I looked into adding to the test class for it but I don't know the policy for locating a https:// box for general testing of it.

just grab me on irc or email if you have questions or want me to revisit parts of it. it did dirty up a pretty simple class to get it worked around.

/home/jesse/osrc/continuum/secure-url-continuum-pre.patch

this is just a initial dump of what I had to do to get things working for review so I can clean things up nicely for a real patch

I am in the process of creating a plexus-url-http component that wraps this functionality into MungedHttpsURL...leaving us the chance to do a better solution in DefaultHttpsURL down the road that prompts the user for cert acceptance and username/password..It is pretty close to done I think, will need to play around getting it plugged in bit should be a much cleaner patch coming once I get a few hours...hopefully this week

/home/jesse/osrc/plexus/secure-url.patch

this patchs in the usage of the file I'll be attaching next for resolution of funky secure urls

/home/jesse/osrc/plexus/secure-url-plexus.patch

used in combination with the MungedHttpsURL file attached

/home/jesse/osrc/continuum/secure-url-continuum-api.patch

includes pom.xml addition to add the plexus-formica package as a dependency for the new class

/home/jesse/osrc/continuum/secure-url-continuum-core.patch

includes pom.xml addition to add the plexus-formica package as a dependency for the new class

in all cases the change to the interface has forces changes in a number of unit tests which are as included and all still validates locally (given the patch into plexus)

There was talk about new components, or different ways to solve the problem, but this at least gets the change in a fairly nonintrusive manner..if someone wants me to look into building out a different type of component or mechanism then let me know.

cheers

attachments from the 25th are the ones that ought to be good to go, the others were for review and whatnot