This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
Summary: | [74cat] Throwable at org.netbeans.modules.editor.fold.FoldChildren.insertImpl | ||
---|---|---|---|
Product: | editor | Reporter: | muellermi <muellermi> |
Component: | Code folding | Assignee: | Svata Dedic <sdedic> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | abjerkho, brettryan, cwt137, ender01, gtzabari, ionuion, janario, messaged, mfukala, mhaungs, michbarsinai, mkhramov, mkroll, mmetelka, mmirilovic, mps77, mrpc, mslama, nadavvin, nickdick74, ouardisoft, padraigdoran, palman, paolosca, peppertech, ravshansbox, tmysik, tnleeuw, tomjnsn, victork, vriha, whedro |
Priority: | P1 | Keywords: | NO74 |
Version: | 7.4 | ||
Hardware: | All | ||
OS: | All | ||
Issue Type: | DEFECT | Exception Reporter: | 204105 |
Attachments: |
stacktrace
stacktrace stacktrace stacktrace stacktrace stacktrace stacktrace stacktrace stacktrace stacktrace stacktrace stacktrace stacktrace |
Description
muellermi
2013-09-24 11:12:43 UTC
Created attachment 140412 [details]
stacktrace
Created attachment 140447 [details]
stacktrace
editing a json file, nothing extraordinarly complex though:
{
"elements": {
"foo": {
"parent": null,
"attributes": {
"cool": "boolean",
"id": "id"
}
},
"infoo": {
"parent" : "foo"
}
}
,
"attributes": {
"one": {
"parent" : null
}
}
}
I think I've added a brace before and "one" element
100% reproducible on following sample. Just type "{" at the pipe position. { "elements": { "foo": { "parents": null, "attributes": { "cool": "boolean", "id": "id" }, "elements": { "infoo":{ } } }, "infoo2": { "parent" : "foo" }, "infoo3": { "parent" : | } } , "attributes": { "one": { "parent" : null } } } I got this while editing a Java class in dev 13-09-26. Have not been able to reproduce reliably, still trying. Created attachment 140587 [details]
stacktrace
.
Created attachment 140588 [details]
stacktrace
.
Created attachment 140589 [details]
stacktrace
.
still same annoying problem with latest buil (20130930...) This problem occurs if I edit a JSF page (maybe HTML in general) There is no such problem, whilst editing Java source. Created attachment 140677 [details]
stacktrace
I was editing a php/html file. I am not sure, but I think this happened when I entered <p> and netbeans autocompleted the </p> tag.
Created attachment 140686 [details]
stacktrace
Edit jsp
Created attachment 140687 [details]
stacktrace
Edit jsp
Created attachment 140688 [details]
stacktrace
Edit jsp
This bug already has 50 duplicates see http://statistics.netbeans.org/exceptions/detail.do?id=204105 Svata, could you please look at this as soon as possible ? With 50+ dups in last 13 days, it became pretty serious issue ... Thanks in advance Although the bug is marked as [74cat] and there's indeed a defect valid for 7.4 the issue is not so highly visible in 7.4. I've added additional sanity checks into the dev version. All entries in this exception report actually come from dev version. (In reply to Svata Dedic from comment #16) > Although the bug is marked as [74cat] and there's indeed a defect valid for > 7.4 the issue is not so highly visible in 7.4. > I've added additional sanity checks into the dev version. All entries in > this exception report actually come from dev version. Ok, thanks for evaluation. Created attachment 140776 [details]
stacktrace
typing text for pattern value in jsf xhtml code:
<f:convertDateTime pattern="DD MM"
(In reply to Svata Dedic from comment #16) > Although the bug is marked as [74cat] and there's indeed a defect valid for > 7.4 the issue is not so highly visible in 7.4. > I've added additional sanity checks into the dev version. All entries in > this exception report actually come from dev version. BTW have you noticed the 100% reproducible case in comment#3? Created attachment 140872 [details]
stacktrace
was taking out a try/catch block in my PHP code and got this error
Created attachment 140981 [details]
stacktrace
change:
if (file.match(//) == null) {
to:
if (file.match(/ /) == null) {
Created attachment 140986 [details]
stacktrace
Changed a <button> tag into a <input> tag. Decided to go back to the original tag and did Ctrl-Z a few times and got this error once I got to the original code.
The defect can be also reproduced by running automatic Fold random manager test. It seems the UNDO operation is a culprit. the Undo operation does not fire pre-change events. So updates of folds affected by 'remove' operation do not work well. Furthermore the undo seems to restore some Positions to their previous state, but the hierarchy might have changed since then including reparenting a fold. The data become damaged and the affected fold cannot be found using binary search. invalid state during remove/replaceByChildren reported by exception reports and automatic test seems fixed. Marek's testcase in comment #3 is fixed by a change in FoldOperationImpl.Refresher that prevents bounds update if the fold contains some children - rev. http://hg.netbeans.org/jet-main/reva1b86e5a5a03 While most of the issue is really resolved now with Svata's latest patch there seems to be some leftout part of this which is still not fixed. It occurs relatively rare(Unlike the already fixed parts which occured a lot). Stacktraces are the same(Only offsets slightly changed due to code updates) and located here: http://statistics.netbeans.org/analytics/detail.do?id=204105 While i'm not sure about builds 131010-131015, builds 131020+ have your last fix applied for sure(I build it myself and know the changes which got in - almost no fold exceptions also indicate it is in) - As you see there is some small case of this folding bug which is not catched yet and reported by only 2 users meantime(Much smaller error-rate than before). See issue #238175; I almost completely changed the matching code; available from build 2013-11-24. |