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.
070328. Not sure how to reproduce. Syntax coloring in Javadoc has been messed up today, with most of the Javadoc grey but seemingly random characters blue.
Created attachment 40117 [details] Stack trace
Honzo are you aware of any changes that could affect this? Javadoc module has not been modified within last 2 weeks.
BTW I noticed this just above the exception in the log: space/src/versionability_89629/ant/project/src/org/netbeans/modules/project/ant/ProjectLibraryProvider.java:0: warning - @param tag has no arguments. To clarify: the bad syntax coloring was in and around the methods that I was editing at the time, all in new code.
The warning is generated by javadoc tool logger that is out of control of the javadoc module. It does not relate to ISE. ISE was thrown due to impossibility to find the tag @param returned by javadoc tool in lexer token sequence. Bad syntax coloring could indicate some lexer issue. But without steps to reproduce it, it is just my guess unfortunately.
Sorry Honzo, no idea offhand. According to cvsps, the last change in java/lexer was elimination of part token IDs on 5 March. Jesse, if it happens again, could you please attach the source code? Thanks.
Perhaps you can put a hook in the code to detect this case and trigger a source dump to the log directory. Even if this happened to me again, it is unlikely I would stop typing the exact instant this exception was thrown and remember to copy and paste the source text.
improved diagnostics /cvs/javadoc/src/org/netbeans/modules/javadoc/hints/JavadocUtilities.java,v <-- JavadocUtilities.java new revision: 1.4; previous revision: 1.3
I just had a similar problem in today's build. Working on RequestProcessor.java, I accepted the hint to add a return value to RP.Task.getPriority. I started typing: @return the priority level (see e.g. {@link Thread# Now the "ad" in "Thread" are in blue, like "@return". Same once I complete the sentence - just those two letters blue: @return the priority level (see e.g. {@link Thread#NORM_PRIORITY} But when I edit it to say @return the priority level (see e.g. {@link Thr ead#NORM_PRIORITY} then it is the "ea" which is blue. Log says space/src/nb_all/openide/util/src/org/openide/util/RequestProcessor.java:0: warning - @return tag has no arguments. When I later accept the hint to add a @param tag to setPriority: /** Changes the priority the task will be performed with. * @param priority the priority level (see e.g. {@link Thread#NORM_PRIORITY} */ public void setPriority(int priority) { then "@link" is correctly blue - but the hint remains on the setPriority line. Both issues are fixed after I close and reopen the file.
IMO the token hierarchy of the java source is out of sync after document changes. Bad syntax highlighting and not recomputed hints are consequences of it. Reassigning to java/source.
Created attachment 40643 [details] Sample of corrupted syntax coloring
Happens quite routinely for newly added code with Javadoc. Rather disconcerting. BTW is there no way to suppress the messages such as ..../Clazz.java:0: warning - @return tag has no arguments. from being printed in the log file? They clutter the log with meaningless lines. This should be fixed before release, IMHO.
suppressing javadoc warnings will require to patch JSR199, there is no api
I filed issue #100488 for the warnings in the log file, so as to keep this issue clearly about the corrupted syntax coloring (and perhaps the one ISE, which I have not seen again).
The broken coloring is covered by issue #99229. Not sure about the ISE - Honza P. AFAIK added some logging because of it. May be related to the fact that we do not use snapshots for the token hieararchy (we have issue(s) for this, but I do not have them handy). Sorry for slow response.
IMHO this is a dup of issue 99229 which could lead to an incorrect offset (and thus a wrong text) of an embedded javadoc token. *** This issue has been marked as a duplicate of 99229 ***