Re: Bugs introduced that affected jde-import-all (patch for last import has comment bug)

Eric, thanks for the info.

Given that it fails in/after the jde-import-all; which is basically: (insert 
(concat "import " new-import ";\n"))
And that it seems to work on the old grammar and fails with the new grammar
(but works invoked in a separate command)

I'm guessing that the condition-case has disabled the after-hooks... ?
Or it is simply that the tags are not being marked?

semantic-highlight-edits-mode does *not* indicate any changes when 
jde-import-all does the insert.
[but again: semantic does not appear to have any tag/region/overlays in that 
area...
 can the apparent loss of the import-tag be a result of some after-hook not 
running?]

And that is the question:
  Why does it not see the tag/region when jde-import does its thing,
  But it apparently *does* see a region and reparse when invoked as a 
separate command.


----- Original Message ----- 
From: "Eric M. Ludlam" <[email protected]>
To: "Jeff Peck" <[email protected]>
Cc: "JDEE Development" <[email protected]>
Sent: Friday, January 08, 2010 3:06 PM
Subject: Re: [jdee-devel] Bugs introduced that affected jde-import-all 
(patch for last import has comment bug)


> Hi
>
> Jeff Peck wrote:
>> But! If I try to run (semantic-parse-changes) using advice around/after 
>> jde-import-all,
>> (or even directly putting it into the end of 
>> jde-import-insert-imports-into-buffer)
>> it has no beneficial effect; it apparently sees no changes and makes no 
>> updates.
>>  Maybe someone who understands how semantic reparse really works can 
>> explain...?
>
> Semantic tracks changes as a buffer is edited.  If you enable 
> semantic-highlight-edits-mode, and then edit code, you can watch how it 
> tracks what you edit, and how it periodically will re-parse everything 
> highlighted.  This is the incremental parser in action.
>
> It works by coalescing edits into groups defined by tag boundaries. Those 
> tags are then reparsed.  If you ever looked inside tags, you would see a 
> 'reparse-symbol' attribute.  It uses this to jump back into the grammar at 
> the right place for that tag.
>
> Based on your loose description, I would guess that the edits are not 
> being tracked, or the edits are not coalescing to the right tags because 
> the tag boundaries are incorrect.  Either of those might be due to a 
> non-local exit which was hidden by a condition-case or some-such.
>
> Since semantic uses after-change-functions, and that is documented as 
> getting cleared if an error is thrown, this is another possibility.
>
> Lastly, the change hooks can be inhibited by inhibit-modification-hooks, 
> which is something that SRecode uses when making big changes, and which 
> also provides a big speed improvement, since there are a lot of minor 
> modes that use change hooks.  See srecode-insert-fcn for an example on how 
> to use it safely with Semantic.
>
> If there is a test case I can examine that doesn't include a successful 
> JDEE install I can probably help, though I have never delved into that 
> specific code before.  Since the JDEE has a lot of code generators, it 
> would be good for a JDEE developer to be familiar with some of the tricks 
> too, similar to what SRecode uses.
>
> If the JDEE ever implements some unit tests, I'd like to know, as I would 
> set it up to run when I test CEDET releases.
>
> Thanks
> Eric 


------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev