Re: Update : rm -rf *Decorator ;-)

Hi guys! Back from a week off... This last sunday, I had a bit of time fixing server-integ issues related to the changes in the LDAP API I've made those last week. This morning, I cleaned most of the failing tests, and I have only 10 of them failing : [ERROR] Tests run: 521, Failures: 8, Errors

Re: Update : rm -rf *Decorator ;-)

Ok, the first problem has been fixed. I'm now down from : [ERROR] Tests run: 521, Failures: 43, Errors: 22, Skipped: 18 to [ERROR] Tests run: 521, Failures: 35, Errors: 12, Skipped: 18 in server-integ. That is 20 less errors, still 47 to get fixed :-) Note that many of them are caused by a ba

Re: Update : rm -rf *Decorator ;-)

Last hours before a week off :-) I was not able to complete the refactoring and having all the server-integ tests passing. There are two main issues that need to be addressed at this point : - the buffer I use to store the encoding result can grow during the encoding, but I have made a mistake: i

Re: Update : rm -rf *Decorator ;-)

Hi ! finally fixed the OSGi failing test, which was about fixing the Control Factories registration. We have ow two lists, one for requests controls and another for response control. API tests are passing green. I started testing the server with the updated API, and hit a first wall, due to the

Re: Update : rm -rf *Decorator ;-)

Ok, more about the failing OSGi test. The PasswordPolicy draft (https://tools.ietf.org/html/draft-behera-ldap-password-policy-10#page-24) defines two different controls with the same OID. They are logical distinguished by the fact that those controls are sent with a request or a response. It's a

Re: Update : rm -rf *Decorator ;-)

Done ! All decorators have been removed, and I have the build working (almost) ! Ok, I wrote 'almost' because there is a couple of glitches I need to address : o first, there is a failing OSGi test. This test, ApiLdapExtrasCodecOsgiTest, tries to get the PasswordPolicyRequest control from the L

Re: Update : rm -rf *Decorator ;-)

Hi ! Another insmnia, another extended operation cleaned up :-) So far, the Cancel request and response have been processed, 2 unit test classes have been added for them, and the StartTLS extended request and response have also been completed. There are 7 remaining extended requests and 8 extend

Re: Update : rm -rf *Decorator ;-)

Hi guys ! After controls, I have worked on extended operations. It's a bit of more refactoring, as an extended operation (request or response) is an LDAP operation, not a sub-element of a ldap operation (the difference being in the way we process it). At this point, I started with the Cancel Req

Update 2 : rm -rf *Decorator ;-)

Hi guys, as of today, Controls have all been converted to not use Decorators. Regarding operations, I still have to deal with extended operations and Intermediate. This is really similar to controls, because we have many extended operations to deal with. Anyway, Controls took a while because we h

Update : rm -rf *Decorator ;-)

A quick update on this sunny sunday afternoon, while travelling to Tübingen to join the OpenLDAP conference : - all the LDAP messages decorator have been removed except the IntermediateResponse and the Extended Request/Response - atm, I'm dealing with controls that need to be encoded too. The idea

Re: rm -rf *Decorator ;-)

Le 03/10/2018 à 16:36, Radovan Semancik a écrit : > On 10/03/2018 04:12 PM, Emmanuel Lécharny wrote: >> I remember Alex reaction when he discovered those Decorator additions, >> which was kind of "what the HECK is THAT ??? Ok, your choice, I'm not >> going to touch it with a 10 feet pole..." (ki

Re: rm -rf *Decorator ;-)

On 10/03/2018 04:12 PM, Emmanuel Lécharny wrote: I remember Alex reaction when he discovered those Decorator additions, which was kind of "what the HECK is THAT ??? Ok, your choice, I'm not going to touch it with a 10 feet pole..." (kind of. He may have been less polite...) 6 or 7 years later (I

Re: rm -rf *Decorator ;-)

> On Oct 3, 2018, at 9:12 AM, Emmanuel Lécharny wrote: > > The problem is that the implementation was really a nightmare (and still > is). Anyone who wants to add a new extendedOperation or a nex Control > has to go through many classes and is likely to get lost (I experienced > it last month wh

Re: rm -rf *Decorator ;-)

On Wed, Oct 3, 2018 at 7:42 PM Emmanuel Lécharny wrote: > Hi ! > > a quick mail as a follow up of my last night insomnias... > > > Last week-end I wa scompleting my rewrite of the Message encoding part. > The gain is clear : > - simpler code (way simpler !!!) > - faster code (way faster, too, 2.5

rm -rf *Decorator ;-)

Hi ! a quick mail as a follow up of my last night insomnias... Last week-end I wa scompleting my rewrite of the Message encoding part. The gain is clear : - simpler code (way simpler !!!) - faster code (way faster, too, 2.5x average) At the end of this refactoring, I faced some issues with the