354449 - nsMicrosummaryService.js not regenerated when enabling/disabling Places

Status: RESOLVED FIXED

(Firefox Graveyard :: Microsummaries, defect)

Description

[Daniel Brooks [:db48x]]

the dialog that asks for the title and location comes up, but clicking on the add button doesn't do anything. Here are the js errors I see:

Error: ServiceManager::GetService returned failure code: = NS_ERROR_XPC_GS_RETURNED_FAILURE
Source file: chrome://browser/content/browser.js
Line: 1427

Error: Component returned failure code: 0x80004002 (NS_NOINTERFACE) [nsISupports.QueryInterface]
Source file: file:///home/db48x/moz/mozilla/firefox-opt-i686-pc-linux-gnu/dist/bin/components/nsMicrosummaryService.js
Line: 637

Based on something philor mentioned on IRC, this may be a dup. Better to record it while I'm thinking of it though.

Comment 1

[Steve England [:stevee]]

User Agent String?

Comment 2

[Phil Ringnalda (:philor)]

Not a dupe of mine, I was waiting for someone with more build-fu to tell me whether I just didn't know what I was doing. After Vlad turned Places off, I did a "make -f client.mk clobber" and "make -f client.mk build" and near as I could tell, everything that needed to know about #ifdef MOZ_PLACES was rebuilt, except for the copy of nsMicrosummaryService.js sitting in my objdir, refusing to go away until I rm'ed it. (Mac and Linux, so -> All)

Comment 3

[Daniel Brooks [:db48x]]

in my case the build id is 20060927.

Phil: that description helps, yes. Looking at the Makefile, there's no way for make to know it ought to have rebuilt it. I think there's a better way of preprocessing a js component that would handle this automatically, but I'll have to look. Alternately, bsmedberg can probably tell us, so I'll cc him.

Barring that, making it a phony target would also work.

Comment 4

[Daniel Brooks [:db48x]]

EXTRA_PP_COMPONENTS maybe? EXTRA_something_COMPONENTS I'm sure…

Comment 5

[Benjamin Smedberg]

EXTRA_PP_COMPONENTS should work

Comment 6

[Daniel Brooks [:db48x]]

Attachment: 354449-1.diff

actually, EXTRA_PP_COMPONENTS won't work because the file was checked in with the .in extension. I've added a comment to that effect.

Comment 7

[Daniel Brooks [:db48x]]

and the patch does indeed fix the bug.

Comment 8

[Myk Melez [:myk] [@mykmelez]]

If EXTRA_PP_COMPONENTS is the right fix, then we should just get a CVS admin like Dave Miller to move the file for us.

Comment 9

[Benjamin Smedberg]

Comment on attachment 240326 [details] [diff] [review]
354449-1.diff

What myk said. You can you my client-side script to do the move if that's easier.

Comment 10

[Daniel Brooks [:db48x]]

Attachment: 354449-2.diff

Presumes nsMicrosummaryService.js gets moved in some fashion.

Comment 11

[Dave Miller [:justdave]]

The cvs copy for this bug made the tree red because of a CVS conflict.  The choices are back out the move, or check this in and trigger a clobber build.

Question: does nsMicrosummaryService.js.in need to get removed also? (The cvs copy was just a copy, the original is still there until deleted)

Comment 12

[Robert Sayre]

Checking in Makefile.in;
/cvsroot/mozilla/browser/components/microsummaries/src/Makefile.in,v  <--  Makefile.in
new revision: 1.2; previous revision: 1.1
done

Comment 13

[Dave Miller [:justdave]]

(In reply to comment #11)
> Question: does nsMicrosummaryService.js.in need to get removed also? (The cvs
> copy was just a copy, the original is still there until deleted)

As far as I know, this ^^ still hasn't been done.  So if it needs to be, don't forget. :)

Comment 14

[Myk Melez [:myk] [@mykmelez]]

> As far as I know, this ^^ still hasn't been done.  So if it needs to be, don't
> forget. :)

Yup.  I have CVS removed the file after getting r=gavin on IRC.