305998 - Improve the new netError.dtd for Firefox 1.5 (package as override)

Status: VERIFIED FIXED

(Firefox :: General, defect)

Description

[Mike Beltzner [:beltzner, not reading bugmail]]

Now that bug 302729 has landed on trunk and branch, Firefox is ready to accept
the friendlier netError.dtd messages and appstrings. However, there's still the
problem that these strings use "Firefox" not &productShortName; meaning that
embedded apps would get broken (bug 302309 addresses this, but doesn't look
hopeful to have it fixed in time for Firefox 1.5)

In order to get the new messages in for Firefox 1.5, this bug suggests that we
pacakage the netError.dtd and appstrings.properties in /browser as an override
of the stuff in /dom.

Comment 1

[Mike Beltzner [:beltzner, not reading bugmail]]

Requesting 1.8b4+ to be carried over from bug 301208

Comment 2

[Mike Beltzner [:beltzner, not reading bugmail]]

Attachment: New network error messages from bug 302108

Bringing the patch over ... see bug 302108 comment 24 to see where the various
pieces of any given error message come from in these two files. I'm personal
computerless for the next few days, so if someone would be so kind as to modify
these so that they get packaged as described in comment 0, I'd be grateful.

Comment 3

[Mike Beltzner [:beltzner, not reading bugmail]]

Gah, all above references should be to bug 301208 (damned fingers)

Comment 4

[Jeff Walden [:Waldo]]

Attachment: Patch which implements a browser-specific override of netError.dtd/appstrings.properties

This patch implements Firefox-specific overrides of netError.dtd (which is
actually chrome://branding/locale/brand.dtd-dependent) and
appstrings.properties (which has hardcoded "Firefox" references - see below). 
The Firefox-specific versions are stored in
<browser/locales/en-US/chrome/overrides/*> and are overridden in
<browser/locales/jar.mn>.  Note that preprocessed files are always overridden,
so a simple "*" in jar.mn will do the trick as well as a "+" (and we'll get the
license preprocessed out of appstrings.properties).

This could be good enough if necessary, but it would be best if we extended the
parsing of appstrings.properties to allow for another item representing
brandShortName to be parsed out of it.	This would primarily allow easier
redistribution with less of a need for unlicensed redistributors to hack at
some part of source to remove "Firefox" references.  That would require hacking
in docshell and uriloader, tho, so it's not nearly as much a slam-dunk,
no-effort patch that can be quickly reviewed and committed easily.

(A location for overridden files had been suggested to me on IRC, but as I'd
already closed Chatzilla when I started on the patch I couldn't check to see
what that location was.  I took a reasonable guess and left it at that,
figuring something was better than nothing.)

Comment 5

[Mike Connor [:mconnor]]

Comment on attachment 193907 [details] [diff] [review]
Patch which implements a browser-specific override of netError.dtd/appstrings.properties

This is so evil.  I'd ask you to rebrand this to Deer Park, but we're
rebranding back really soon now.

That said, r=me, please file a bug on making the appstrings stuff
branding-friendly.

Comment 6

[Mike Beltzner [:beltzner, not reading bugmail]]

If we don't care so much about making the files branding friendly for FFx 1.5
then I think we can continue to track the task of making things branding
friendly for trunk back in bug 302309 and bug 301208, both of which are targeted
to trunk.

Comment 7

[Jeff Walden [:Waldo]]

The patch is checked in on branch and trunk.  Bug 306369 has been filed to
address the incredible evils which I just committed to trunk code, but it's
unlikely the fix there will be accepted for branch too.

Comment 8

[Alexander L. Slovesnik]

<!ENTITY netTimeout.longDesc "&sharedLongDesc;">
<!ENTITY netReset.longDesc "&sharedLongDesc;">
<!ENTITY netInterrupt.longDesc "&sharedLongDesc;">
<!ENTITY connectionFailure.longDesc "&sharedLongDesc;">

Can you please explain how localizers should translate these strings?
Or this is mistake?

Comment 9

[Alexander L. Slovesnik]

Oops, sorry.
I found it in the end.

Comment 10

[Marek Stępień [:marcoos, inactive]]

Comment on attachment 193907 [details] [diff] [review]
Patch which implements a browser-specific override of netError.dtd/appstrings.properties

+  <li>Check to make sure your system has the Personal Security Manager
+    installed.</li>

I just wonder how Joe Sixpack is going to figure out if his Firefox has the
Personal Security Manager or not. ;-)

Comment 11

[Mike Beltzner [:beltzner, not reading bugmail]]

(In reply to comment #10)
> I just wonder how Joe Sixpack is going to figure out if his Firefox has the
> Personal Security Manager or not. ;-)
> 

I checked into that, actually. This error  message won't ever appear in Firefox.
We're including it for embedders or Seamonkry users (mostly on Linux iirc) who
might not have installed that package. ;)

Comment 12

[Axel Hecht]

Jeff, this needs a follow-up announcement on .l10n, too.

Whatever you come up with for the trunk, it should get a l12y review, could you
make sure that I'm CCed on those bugs? I can't really keep track on this by 
tracking burning l10n boxes and bonsai.

Comment 13

[Jeff Walden [:Waldo]]

(In reply to comment #12)
> Jeff, this needs a follow-up announcement on .l10n, too.

Yes, I meant to do that but got sidetracked.  Post coming after I finally get
some sleep tonight (it's a long story)...

> Whatever you come up with for the trunk, it should get a l12y review, could you
> make sure that I'm CCed on those bugs? I can't really keep track on this by 
> tracking burning l10n boxes and bonsai.

Will do.