gnodet-bot commented on code in PR #28198:
URL: https://github.com/apache/camel/pull/28198#discussion_r4333480670
##########
components/camel-jira/src/main/java/org/apache/camel/component/jira/consumer/NewIssuesConsumer.java:
##########
@@ -129,16 +129,19 @@ protected int doPoll() throws Exception {
* @return queue of new issues
*/
private Queue<Issue> getNewIssues() {
- String jqlFilter;
+ Queue<Issue> issues;
// If we have processed an issue before, use its timestamp to get only
the newer ones
if (latestIssue != null) {
- jqlFilter = String.format(NEW_ISSUES_JQL_FORMAT,
getServerTimestamp(), getEndpoint().getJql());
+ String jqlFilter = String.format(NEW_ISSUES_JQL_FORMAT,
getServerTimestamp(), getEndpoint().getJql());
+ // all of them: the newest one becomes the latest issue, so a new
issue left out by maxResults would never
+ // be returned again; maxResults is the page size here
+ int maxResults = getEndpoint().getMaxResults();
+ int pageSize = maxResults > 0 ? Math.min(maxResults,
SEARCH_MAX_PER_QUERY) : SEARCH_MAX_PER_QUERY;
+ issues = getIssues(jqlFilter, SEARCH_START_AT, pageSize,
Integer.MAX_VALUE);
Review Comment:
💡 **Behavioural change worth documenting**: the user-configured `maxResults`
is used here only as `pageSize`; the actual result cap is `Integer.MAX_VALUE`.
The previous code capped the total at `maxResults`. The comment explains the
rationale, but this change is non-obvious from the parameter name — a Javadoc
note on `JiraConfiguration.getMaxResults()` clarifying its dual role (page size
for `newIssues`, total cap for other consumers) would help future readers.
##########
components/camel-jira/src/main/java/org/apache/camel/component/jira/consumer/AbstractJiraConsumer.java:
##########
@@ -155,6 +151,27 @@ protected Queue<Issue> getIssues(String jql, int start,
int maxPerQuery, int max
return new LinkedList<>(issues);
}
+ /**
+ * Jira Cloud cannot page by startAt: its search client throws
UnsupportedOperationException for a startAt other
+ * than 0. The pages are read with the next page token of the previous
page instead, from the first issue, until the
+ * last page (no next page token) or maxResults (ignored if it's <= 0).
+ */
+ private Queue<Issue> getCloudIssues(
+ AsynchronousCloudSearchRestClient searchRestClient, String jql,
int maxPerQuery, int maxResults) {
+ Set<Issue> issues = new LinkedHashSet<>();
+ String nextPageToken = null;
+ do {
+ SearchResult searchResult
+ = searchRestClient.enhancedSearchJql(jql, maxPerQuery,
nextPageToken, Set.of("*navigable"), null).claim();
+ for (Issue issue : searchResult.getIssues()) {
+ issues.add(issue);
+ }
+ nextPageToken = searchResult.getNextPageToken();
+ } while (nextPageToken != null && !(maxResults > 0 && issues.size() >=
maxResults));
Review Comment:
⚠️ **Unbounded memory accumulation**: when `maxResults = Integer.MAX_VALUE`
(as passed from `NewIssuesConsumer`), the `issues.size() >= maxResults` guard
never fires. After a long outage or a large import, all new issues are
accumulated in-heap before any exchange is routed. Consider logging a WARN at a
threshold (e.g. every 10 000 issues) so runaway polls are visible in the logs.
##########
components/camel-jira/src/main/java/org/apache/camel/component/jira/consumer/AbstractJiraConsumer.java:
##########
@@ -122,6 +122,10 @@ protected Queue<Issue> getIssues(String jql, int start,
int maxPerQuery, int max
maxPerQuery = maxResults;
}
+ if (endpoint.getClient().getSearchClient() instanceof
AsynchronousCloudSearchRestClient cloudSearchRestClient) {
+ return getCloudIssues(cloudSearchRestClient, jql, maxPerQuery,
maxResults);
Review Comment:
ℹ️ **Broader scope than the PR title suggests**: this `instanceof` dispatch
sits in `AbstractJiraConsumer.getIssues()`, which is shared by
`NewCommentsConsumer` and `WatcherConsumer` as well. Those consumers now
silently switch to token-based Cloud paging too. Worth a note in the PR
description and ideally in the method Javadoc so maintainers know this is an
intentional cross-consumer change.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
