Bug #223

reconsider Java indexing

Added by Michael Krkoska 929 days ago. Updated 924 days ago.

Status:New Start:07/20/2009
Priority:Normal Due date:
Assigned to:- % Done:

0%

Category:Document Scanner
Target version:-

Description

Hi,

A naive user would expect, that the search syntax examples from the user guide apply to all file types:

http://www.supose.org/wiki/supose/SearchingContents

But there is no field contents for a Java file. Instead there is only data in the undocumented(?) methods and comments fields. This seems to be very unexpected.

Admitted, all you have to do "to make it work as expected" is comment the line starting with java= in etc/fileextension.properties. But to find this, the new user has to read either source code, or find the appropriate bug report (did I miss something obvious?). The user guide does not mention this.

While adding fields like methods and comments is certainly useful for some cases, contents should be filled as well in order to stay compatible with the other file types.

Please reconsider the Java indexing strategy.

Thanks,
Michael

History

Updated by Karl Heinz Marbaise 924 days ago

Hi Michael,

Hi,

A naive user would expect, that the search syntax examples from the user guide apply to all file types:

http://www.supose.org/wiki/supose/SearchingContents

But there is no field contents for a Java file. Instead there is only data in the undocumented(?) methods and comments fields.

Not really undocumented (see the fields list), but you're right it can be improved...

Admitted, all you have to do "to make it work as expected" is comment the line starting with java= in etc/fileextension.properties.

This depends on what you expect...but it could be improved...(Users Guide)

But to find this, the new user has to read either source code, or find the appropriate bug report (did I miss something obvious?). The user guide does not mention this.

Correct. As mentioned above it could be improved... May be you can issue an enhancement for the Users guide...

While adding fields like methods and comments is certainly useful for some cases, contents should be filled as well in order to stay compatible with the other file types.

There are three options for this:

  1. Index Java files all the time as text based on fileextenstion.properties configuration which results in contents field.
  2. Index Java files as text and as Java files which results in contents, methods and comments fields but with reducing in scanning performance
  3. Index Java only as Java with methods and comments fields.

Please reconsider the Java indexing strategy.

I'm thinking about it over and over, cause in the future i will support different files types like Pyhton, Perl, C, C++, Ruby etc.

Kind regards
Karl Heinz Marbaise

Also available in: Atom PDF