It looks like this toMap method is only used (transitively) by SolrConfigHandler.handleGet to print the config and show it to a client. Probably safe to add the other elements as well, good catch.
On Tue, Jun 30, 2015 at 3:49 PM, Christine Poerschke (BLOOMBERG/ LONDON) < [email protected]> wrote: > Hi, > > Follow-on question, also re: SolrIndexConfig's toMap method. The > constructor reads "/mergedSegmentWarmer" and "/infoStream" and > "/infoStream/@file" elements but the toMap method does not write them. > > Would that be intended or is it maybe an unintended omission? If intended > would be great to have a comment in toMap to clarify why/which elements are > skipped, so that if/when a new element is added its easier to determine if > toMap should write the new element also. > > Happy to open a JIRA with a patch, or to add to Mike's JIRA/patch. > > Christine > > From: [email protected] At: Jun 30 2015 21:03:19 > To: [email protected] > Subject: Re:Solr Config MergePolicy/MergeScheduler Naming Bug? > > Concur, that looks like a copy/paste bug. The JIRA(s) associated with the > commit are no longer open, so a new JIRA with the one-line patch would seem > the next logical step. This appears to affect trunk but not branch_5x or > branch_4x. Well spotted! > > From: [email protected] At: Jun 30 2015 20:34:48 > To: [email protected] <#14e463b75da28f90_> > Subject: Re:Solr Config MergePolicy/MergeScheduler Naming Bug? > > I was looking through code for unrelated reasons and this line stuck out > to me: > > > https://github.com/apache/lucene-solr/blob/trunk/solr/core/src/java/org/apache/solr/update/SolrIndexConfig.java#L180 > > if(mergeSchedulerInfo != null) > m.put("mergeScheduler",mergeSchedulerInfo.toMap()); > if(mergePolicyInfo != null) m.put("mergeScheduler",mergePolicyInfo. > toMap()); > > Are they both supposed to be using the "mergeScheduler" key? If not, happy > to open a JIRA and provide the one-line patch. > > Mike > >
