Index contents and format [message #199755] |
Wed, 30 March 2005 05:04  |
Eclipse User |
|
|
|
Hello,
We want to implement search capabilities (references) similar to the ones
that JDT provides for Cobol and PL1 resources. To do so, we would like to
duplicate the architecture that JDT uses and the indexing mechanism in
particular. In our case, we will need to trigger a dedicated parser on the
source files in order to populate the indexes. I am looking for detailed
information about the indexes:
- What is the precise contents of the information stored by the indexes?
- I understand that the indexer is tightly linked to the compiler. What is
the relationship between the two? Which subset of the compiler is used by
the indexer?
Thank you for your help!
Nicolas
|
|
|
|
Re: Index contents and format [message #199785 is a reply to message #199769] |
Wed, 30 March 2005 09:49  |
Eclipse User |
|
|
|
Hello Jerome,
Thanks for your answer.
Would the searchParticipant work for non Java language ? I've seen that it
is used by JSP, which is a language close to Java.
I've seen that you can define your own parsing but that it would record
"Java index entries". Does that just mean that my Cobol index would be
stored in the jdt plugin metadata or that it would be mixed with Java
indexes?
Is it possible to run the indexing in background during checkout, ressource
change, etc. as you do with Java files? I've seen that
IndexManager>>addSource manipulates a defaultSearchParticipant, which is the
Java one. I think we can't afford to index live during the search.
Thanks again,
Nicolas
"Jerome Lanneluc" <jerome_lanneluc@fr.ibm.com> wrote in message
news:d2e2g0$l9g$1@news.eclipse.org...
> The format of the indexes is not specified and is subject to change.
> With that said, an index contains essentially index keys to documents
> maps. An index key encodes a reference or a declaration as it appears in
> the source. So no resolved information is stored in the index. You can
> find a more detailed description of index keys by looking at the
> createIndexKey(...) methods in the JavaSearchPattern hierarchy.
>
> The part of the compiler that is used is a specialized parser
> (SourceElementParser). This parser is a subclass of the compiler parser.
> That is the only relationship it has with the compiler.
>
> Also did you look at participating in search (and indexing) using a
> SearchParticipant ?
>
> Jerome
>
>
> Nicolas Dangeville wrote:
> > Hello,
> >
> > We want to implement search capabilities (references) similar to the
ones
> > that JDT provides for Cobol and PL1 resources. To do so, we would like
to
> > duplicate the architecture that JDT uses and the indexing mechanism in
> > particular. In our case, we will need to trigger a dedicated parser on
the
> > source files in order to populate the indexes. I am looking for detailed
> > information about the indexes:
> > - What is the precise contents of the information stored by the indexes?
> > - I understand that the indexer is tightly linked to the compiler. What
is
> > the relationship between the two? Which subset of the compiler is used
by
> > the indexer?
> >
> > Thank you for your help!
> >
> > Nicolas
> >
> >
> >
> >
|
|
|
Powered by
FUDForum. Page generated in 0.05993 seconds