Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Status
The first post of this thread is a WikiPost and can be edited by anyone with the appropiate permissions. Your edits will be public.
Is the GCC PowerPC compatible? Perhaps the parts of the 2.6.1 Xcode that are failing can be patched with the working components from 10A096 version, or recompiled?

As of now I am unsure gcc is functional. I tried to build macports on that system with Xcode from 10A261, and it fails:

Screenshot 2021-11-10 19-25-12.png


How can I test gcc specifically?

I can also move some Xcode components under root from one system to another and find out when things start working. But doing that may take time, there are many components scattered around.
 
  • Like
Reactions: ChrisCharman
As of now I am unsure gcc is functional. I tried to build macports on that system with Xcode from 10A261, and it fails:

View attachment 1907315

How can I test gcc specifically?

I can also move some Xcode components under root from one system to another and find out when things start working. But doing that may take time, there are many components scattered around.
It says the C compiler isn’t working but that’s not very useful - perhaps use ‘file’ on the binaries inside the Developer Tools and see what architectures they’re compiled for? It does look as though ‘make’ is intel on your system now. If GCC has been compiled for PowerPC then we can use it as a replacement for the earlier version, even without the rest of the IDE. I may have a look through the 10A432 version for later builds too actually, it’s something I hadn’t considered doing until now.
 
It says the C compiler isn’t working but that’s not very useful - perhaps use ‘file’ on the binaries inside the Developer Tools and see what architectures they’re compiled for? It does look as though ‘make’ is intel on your system now. If GCC has been compiled for PowerPC then we can use it as a replacement for the earlier version, even without the rest of the IDE. I may have a look through the 10A432 version for later builds too actually, it’s something I hadn’t considered doing until now.

Will do that once back to office. By the way, what can I use to batch compare contents of folders with architectures and versions listed? Like, to find out what components differ between two Xcode installations at once?
 
  • Like
Reactions: ChrisCharman
I’m unsure. Both @B S Magnet and myself would like to find a script that would easily achieve this as well.

I see. And how to check architectures of executables, unless they are explicitly stated in a file name?

P. S. By the way, this post mentions 10.6 builds that our Wikipage lacks: 10A246 and 10A250.
 
I see. And how to check architectures of executables, unless they are explicitly stated in a file name?

P. S. By the way, this post mentions 10.6 builds that our Wikipage lacks: 10A246 and 10A250.
Yeah there are a number of builds i’ve seen mentioned across the net over the last few days that we haven’t seen before - i’ve not seen 10A246 and 10A250 before though.

To check and individual binary you can simply type ‘file’ in the Terminal followed by the path to the executable, or ‘otool -l’ instead of ‘file’ to list the library dependencies for the binary.
 
  • Like
Reactions: barracuda156
It says the C compiler isn’t working but that’s not very useful - perhaps use ‘file’ on the binaries inside the Developer Tools and see what architectures they’re compiled for? It does look as though ‘make’ is intel on your system now. If GCC has been compiled for PowerPC then we can use it as a replacement for the earlier version, even without the rest of the IDE. I may have a look through the 10A432 version for later builds too actually, it’s something I hadn’t considered doing until now.

Perhaps it is worth trying to install Xcode from 10A222 on a clean system too. I can do that, since I have now a subpartition for experiments that can be reformatted.

Does anyone know what components Xcode installs outside its own directory Developer (exhaustive list)?
 
  • Like
Reactions: ChrisCharman
Thanks. The command counted 104 hits for Build 10A96.

find . -type f -perm +111 -exec lipo -info {} \; | grep ppc64 | wc -l

reveals 24 hits on 10A190. On the other hand, on my 10.5.8 install, the same command reveals 1140 hits, although it's not really a clean install anymore. Indeed doesn't sound like ppc64 was sitting on top of their to-do list for SL.

Indeed it looks as though, as things were recompiled and replaced, the ppc64 flag wasn’t used. In many cases the ppc or ppc7400 flag was still used. This is either due to the elusive internal builds, as the G5 will run 32 bit ppc binaries, or possibly for Rosetta compatibility. It can’t have been accidental for over 430 separate builds.

Reviewing the Apple reference library for guides on building for ppc64 leads me to believe that it would have taken a lot more effort for the engineers to continue to work on 64bit PowerPC builds as well and, even as a contingency, it would be very ‘Apple’ to keep the 64bit USP of its flagship operating system for only its shiny new products.
 
  • Like
Reactions: B S Magnet
I see. And how to check architectures of executables, unless they are explicitly stated in a file name?

P. S. By the way, this post mentions 10.6 builds that our Wikipage lacks: 10A246 and 10A250.

I found the source post. I’ll be updating the wikipost in a moment. Thanks for this heads-up!
 
This is my current understanding of the timeline for the seeds we’re aware of so far:

—————— 10.6 Development ——————

09 Jun 2008 10A96 WWDC DP 1

10A096 (Server)
———————————————————-

25 Oct 2008 10A190 ADC Seed

04 Dec 2008 10a222 ADC Seed

## Jan 2009 10A246 Internal Build

## Jan 2009 10A250 Internal Build

04 Feb 2009 10A261 ADC Seed

06 Mar 2009 10A286 ADC Seed

02 Apr 2009 10A314 ADC Seed

## Apr 2009 10A335 [Seed Update]

23 Apr 2009 10A354 [Seed Update]

09 May 2009 10A355 [Seed Update]

———————————————————-

11 Jun 2009 10A380 WWDC DP 2

10A380 (Server)

———————————————————

28 Jun 2009 10A394 [Seed Update]

09 Jul 2009 10A402a [Seed Update]

10A403 (Server)

17 Jul 2009 10A411a [Seed Update]

24 Jul 2009 10A421a [Seed Update]

…………………………………………………………

## ___ 2009 10A428 Internal Build?

## ___ 2009 10A430 Internal Build?

…………………………………………………………

12 Aug 2009 10A432 GM ADC Seed

10A433 GM (Server)
 
This is my current understanding of the timeline for the seeds we’re aware of so far:

—————— 10.6 Development ——————

09 Jun 2008 10A96 WWDC DP 1

10A096 (Server)
———————————————————-

25 Oct 2008 10A190 ADC Seed

04 Dec 2008 10a222 ADC Seed

## Jan 2009 10A246 Internal Build

## Jan 2009 10A250 Internal Build

04 Feb 2009 10A261 ADC Seed

06 Mar 2009 10A286 ADC Seed

02 Apr 2009 10A314 ADC Seed

## Apr 2009 10A335 [Seed Update]

23 Apr 2009 10A354 [Seed Update]

09 May 2009 10A355 [Seed Update]

———————————————————-

11 Jun 2009 10A380 WWDC DP 2

10A380 (Server)

———————————————————

28 Jun 2009 10A394 [Seed Update]

09 Jul 2009 10A402a [Seed Update]

10A403 (Server)

17 Jul 2009 10A411a [Seed Update]

24 Jul 2009 10A421a [Seed Update]

…………………………………………………………

## ___ 2009 10A428 Internal Build?

## ___ 2009 10A430 Internal Build?

…………………………………………………………

12 Aug 2009 10A432 GM ADC Seed

10A433 GM (Server)

Here someone claims to have 10A314 and 10A314e: https://www.insanelymac.com/forum/t...betasclassics/?do=findComment&comment=1637759
 
  • Like
Reactions: ChrisCharman
## Apr 2009 10A335 [Seed Update]

23 Apr 2009 10A354 [Seed Update]

09 May 2009 10A355 [Seed Update]

I’ve spent some time over the past couple of days to review threads from 2009. Mentions of a “10A355” appeared to be conflated with Build 10A335, as mentions of it preceded mention of a build 10A354 by several days and appeared in mentions very shortly after seed release of 10A335.

———————————————————-

09 Jul 2009 10A402a [Seed Update]

The builds with an “a”-suffix appear to be outnumbered by the same without the suffix. Here’s one such case. For now, in Table 1 I’ve updated those builds without a suffix, as they appear to be the same build released at the same time.

10A403 (Server)

17 Jul 2009 10A411a [Seed Update]

24 Jul 2009 10A421a [Seed Update]

…………………………………………………………

## ___ 2009 10A428 Internal Build?

## ___ 2009 10A430 Internal Build?

…………………………………………………………

12 Aug 2009 10A432 GM ADC Seed

10A433 GM (Server)

I wasn’t aware that the GM of 10.6.0 Server was 10A433. I’ll make an update reflecting that.

Also, because the only mention anywhere that I can find of a “Build 10B87” is on this thread’s wikipost, I’m tentatively removing it until or if we find a reference we can cite.


The link from there mentions Demonoid had those images. Can anyone check with an active account at Demonoid? I had one ages ago, but apparently it got removed when they moved to a different domain.

The current iteration of Demonoid features nothing prior to 2019, meaning database entries between 2003 and 2019 are lost.
 
Last edited:
I’ve spent some time over the past couple of days to review threads from 2009. Mentions of a “10A355” appeared to be conflated with Build 10A335, as mentions of it preceded mention of a build 10A354 by several days and appeared in mentions very shortly after seed release of 10A335.
I would agree however i have seen video footage of the build on YouTube (screenshot below) though descriptions of experiences between the separate builds is more than likely error prone. It could be that 354 and 355 are updates to client and server?

D72D475D-E934-42B4-AC5C-B1224128EF1F.jpeg


The builds with an “a”-suffix appear to be outnumbered by the same without the suffix…
I wasn’t aware that the GM of 10.6.0 Server was 10A433…

Also, because the only mention anywhere that I can find of a “Build 10B87” is on this thread’s wikipost, I’m tentatively removing it until or if we find a reference we can cite…
Yes i’m unsure as to the origins and meanings of the suffix as well.

Indeed, the retail 10A433 is 10.6.0 Server.

I can’t find any reference to Build 10B87 anywhere other than the wiki either, i agree that it should be removed.
 
Last edited:
  • Like
Reactions: B S Magnet
I would agree however i have seen video footage of the build on YouTube (screenshot below) though descriptions of experiences between the separate builds is more than likely error prone. It could be that 354 and 355 are updates to client and server?

View attachment 1907795

Very possibly, especially at that later stage of product development, though the screen cap here shows 10A335, not 10A355, and 10A335 is thoroughly documented by several independent sources and whose date of release comports well with the dates associated with other releases.

Sidebar: it’s noteworthy to mention how the internal builds were probably happening daily on most weekdays though probably not every single day (with times where a couple of days might have been needed). This puts the start of development (when a “Build 10A1” got underway) sometime between 23 November 2007 and 22 January 2008.

Yes i’m unsure as to the origins and meanings of the suffix as well.

In time, maybe someone with knowledge on that might emerge to clear it up.

Indeed, the retail 10A433 is 10.6.0 Server.

I can’t find any reference to Build 10B87 anywhere other than the wiki either, i agree that it should be removed.

The duty is done!
 
  • Like
Reactions: ChrisCharman
Very possibly, especially at that later stage of product development, though the screen cap here shows 10A335, not 10A355, and 10A335 is thoroughly documented by several independent sources and whose date of release comports well with the dates associated with other releases.

Sidebar: it’s noteworthy to mention how the internal builds were probably happening daily on most weekdays though probably not every single day (with times where a couple of days might have been needed). This puts the start of development (when a “Build 10A1” got underway) sometime between 23 November 2007 and 22 January 2008.



In time, maybe someone with knowledge on that might emerge to clear it up.



The duty is done!

This is all i’ve been able to find on 10A355 though 10A354 seems to be well reported by many sources.

From what i can gather the internal build production accelerated, so there would have been days where multiple builds were generated. Given lengthy compile times based on the technology available at the time it’s probably safe to assume that the earlier and larger builds would have taken longer, possibly days, so it’s difficult to pinpoint when builds were generated without looking at the binaries themselves. You could look at the confirmed dates that we have for the developer seeds and then calculate the number of builds generated between them, then average that across time to a predicted ‘Build 0’ i suppose, as a curiosity.

It’s interesting to note that all developer seeds released after WWDC 2009 were only available via software update, with the exception of the GM, unless i’m mistaken. Prior to that they were made available on ADC premiere downloads page.
 
Last edited:
  • Like
Reactions: B S Magnet
It’s interesting to note that all developer seeds released after WWDC 2009 were only available via software update, with the exception of the GM, unless i’m mistaken. Prior to that they were made available on ADC premiere downloads page.

From a review of surviving torrent links (there really aren’t many), it does appear post-WWDC 2009 builds were available both as Software Updates (like a delta update) and also as standalone installations ostensibly posted to the ADC site. It’s also possible pre-WWDC 2009 seeds were like this, but it’s just as likely Software Update on the earlier builds might have had some issues preventing “delta”-styled updates from being feasible. It’s hard to know without having been there in the moment or relying on someone’s memory who had.
 
  • Like
Reactions: ChrisCharman
From a review of surviving torrent links (there really aren’t many), it does appear post-WWDC 2009 builds were available both as Software Updates (like a delta update) and also as standalone installations ostensibly posted to the ADC site. It’s also possible pre-WWDC 2009 seeds were like this, but it’s just as likely Software Update on the earlier builds might have had some issues preventing “delta”-styled updates from being feasible. It’s hard to know without having been there in the moment or relying on someone’s memory who had.
It’s my understanding from reading reports and seed notes from the time that as of 10A394 onwards, updates were provided as Deltas instead of complete builds that needed to be imaged onto a partition. If that’s correct, the only complete builds (Other than the GM) following DP2 10A380 that we’ll be able to find will have been either created and uploaded to torrent sites by testers at the time or leaked internal builds. The update packages are much smaller in size, obviously, so will be easy to identify if they show up anywhere - 10A394 for example should be around 700mb. It does explain why we’ve not yet found any builds, with the exception of the ‘accidental’ posting of 10A403 to ADC, between 10A380 and 10A432 GM.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.