The work was never done to carbonate the installer. it still links against non- carbon libraries as well as a bunch of metrowerks libraries that i don't know if they're carbonated or not... bugger.
this is so a beta blocker for osx.
Whiteboard: OSX++, PDT+
sfraser is jumping on the grenade.
Assignee: ssu → sfraser
Carbonation is what you do to high fructose corn syrup.
Status: NEW → ASSIGNED
Summary: Installer needs to be carbonated → Installer needs to be carbonized
*** Bug 84037 has been marked as a duplicate of this bug. ***
I attached some work-in-progress diffs that Carbonize the installer code. There are a lot of changes, because the installer code is very poorly factored. Not attached are the new prefix files for each of the 4 new targets (mozilla/netscape carbon/classic). We really need to rejig the targets in the project, since it seems that only resources differ between Netscape and Mozilla targets. Netscape targets should not be in this tree anyway. To build with these changes, you also need to add CarbonLib and CarbonAccessors.o to the MIW.mcp project, and remove all the classic stub libs from the carbon taregets. In addition, libjar.Lib incorrectly links with an MSL lib, which needs to be fixed. I did find carbonized versions of GUSI on the net, but the main GUSI has not been carbonized yet. I got as close as having 1 link error in the carbon target with the carbonized GUSI. Having seen the installer code up-close and personal, it is my strong opinion that this code should be thrown away, and the installer written from scratch using a well-establish application framework (like PowerPlant). Since we've decided to not ship an installer for the X build, throwing back to sdagley.
Assignee: sfraser → sdagley
Status: ASSIGNED → NEW
Let me clarify simon's comment "we've decided to not ship an installer for the X build" - that applies to the beta release for X. We're going to need something working by the time we RTM the X version
Clearing PDT+ and Blocker severity. We can live without the installer for the beta stage of the X version
Severity: blocker → major
Whiteboard: OSX++, PDT+ → OSX++
Added kw: nsenterprise+. Priority: P1, and Target milestone: Mozilla0.9.4
Priority: -- → P1
Target Milestone: --- → mozilla0.9.4
Adding syd to cc: list since he 'owns' installer these days
Grega agrees that we can move installer bugs out from the M0.9.5 milestone. So that Steve D can focus on some more urgent UI bugs for the M0.9.5. Reason: Steve said there is a walkaround for Mac OSX user even if we do not provide a installer.
Target Milestone: mozilla0.9.5 → mozilla0.9.6
It occurs to me that an installer is largely unnecessary for Mozilla, proper, and only supports the Netscape cruft (lovingly). A .dmg with an .app would be great for Mozilla.
nominate nsbranch. I feel this is needed for a Netscape release. What is the workaround as mentioned in the 9-11 comments which is acceptable to the general public?
The workaround is to go with a 'blob' install but one that's formatted as a bundled application (it packages the app and requisite files as a single icon the user can drag anywhere on their hard drive) inside of a compressed disk image. This is the format adopted by most OS X apps that don't have/need installers.
I forgot to add that jj has a seperate bug for doing the build work for the packaging and disk image generation
That bug is 100168 - I'm working on it
Nominating for PDT+ because this is a P1 and nsenterprise+. Why is this targeted for 0.9.6???
Keywords: nsbranch → nsbranch+
Because we're not going to have it for the Mac OS X release and AFAIK this isn't an enterprise requirement (I should have cleared those keywords earlier). As was commented on for the N6.1 preview for Mac OS X people liked the fact we had a simple drag installer (not realizing we had no installed at all for the OS X version)
Keywords: nsbranch+, nsenterprise+
Priority: P1 → P3
Whiteboard: OSX++ → OSX
Target Milestone: mozilla0.9.6 → Future
Thanks for the explanation. Where would the end user license file/read me be placed since we don't have an installer? Would the user license file be accessible to the user from the About page? Would the readme file be available in the Netscape folder? Thanks.
We'll still have a folder with support files, such as the Read Me and the Profile Manager icon where the application bundle will 'live'. It's just the app and critical files/folders will be encapsulated as an application bundle so the user won't see the parts unless they ask to (there's a Finder option to reveal the contents of a bundle).
answer is "yes" to all of the above. both license and read me files will be located next to the app (bundle) in the Nescape folder, and the EULA is also accessible from the about page as it's always been.
This bug might be in need of closing as WONTFIX but I'm tossing it to Syd to make that determination.
Assignee: sdagley → syd
You don't need an installer to display the EULA. You can make Disk Copy images which won't mount unless the user agrees to a EULA (although the legal status of such agreements is suspect regardless of which method is used). I agree with Bill McGonigle, an installer isn't necessary and Apple's reccomended method for distribution is a Disk Copy image with a stand-alone application which can be dragged wherever the user wants. Note: 0.9.5 doesn't appear to work if the app bundle is located on a volume without write access. That's definitely a bug
Strobe: what you describe about the EULA is exactly what we're doing for Mozilla 0.9.5 and beyond. (pops up when the disk image is mounted) Regarding your comment about the crash on startup when launched from the disk image, this has been logged as bug 104954. Unless we absolutely need a "modular" installer like we have on other platforms, I think the current "drag-to-install" solution is satisfactory on MacOS X, and we can close this bug as "wontfix".
Closing as WONTFIX since we have settled the issue of the OS X installer
Status: NEW → RESOLVED
Last Resolved: 17 years ago
Resolution: --- → WONTFIX
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.