12 replies to this topic

[steve22]

Does anyone know if we'll see any new Snow Leopard builds/seeds any time soon or should I download the current Developer Preview to try it out.  

I remember we got the first Tiger/Leopard seeds around 2 months after WWDC 04/06.

~Matt

[Colonel]

Nope, there haven't been any new Snow Leopard builds yet. Go get the WWDC build.

[Adrian Fogge]

I guess it all depends on your definition of soon...
Are we talking about cosmic, geologic, or 3Drealms?

We will obviously be seeing an update sometime "soon".
For approximately a month after WWDC, there were no new launchd submissions, however development has resumed and appears to be coming along with quite a few are directly related to 64-bit including incorrect memory allocations being assigned, improper rounding of 64-bit registers, and improper deallocation of memory.

There are also additional changes that reference additional debugging being added.
In the past, debugging is added approximately 2 weeks prior to a seed.

The majority of the debugging that is being added appears directly related to exit failures, and the basis of additional error checking that *could* be used to try to recover from a recursive loop error that would normally result in an application crash depending on where it is implemented in the system. Could definitely cause some data loss, but we are talking about crash recovery so data loss is already on the table.

As always, no real guarantees on when we will see much of anything... just my interpretation of the publicly available launchd trac changelogs.

~Adrian

[steve22]

Thanks. It's been over 3 months since WWDC and still no new builds

http://launchd.macos...g/trac/timeline

Looking at the launchd timeline there hasn't been any changes in over a week. What could this indicate?

I also noticed that there haven't been any mentions of Snow Leopard builds for 4 months now? http://launchd.macos...?...&ticket=off

Adrian Fogge, on Aug 16 2008, 08:44 PM, said:

I guess it all depends on your definition of soon...
Are we talking about cosmic, geologic, or 3Drealms?

We will obviously be seeing an update sometime "soon".
For approximately a month after WWDC, there were no new launchd submissions, however development has resumed and appears to be coming along with quite a few are directly related to 64-bit including incorrect memory allocations being assigned, improper rounding of 64-bit registers, and improper deallocation of memory.

There are also additional changes that reference additional debugging being added.
In the past, debugging is added approximately 2 weeks prior to a seed.

The majority of the debugging that is being added appears directly related to exit failures, and the basis of additional error checking that *could* be used to try to recover from a recursive loop error that would normally result in an application crash depending on where it is implemented in the system. Could definitely cause some data loss, but we are talking about crash recovery so data loss is already on the table.

As always, no real guarantees on when we will see much of anything... just my interpretation of the publicly available launchd trac changelogs.

~Adrian

[Embio]

coming up on month 4 and no builds - surely it has to be soon?

if they get released tomorrow I'm an Apple insider

[mikecwest]

Has any heard of build 10B87?   I saw it mentioned on a Wikipedia page, only briefly being released on 10B87.  Does anyone have any details?

[Adrian Fogge]

Normal naming convention is that 'B' refers to the 10.x.1 release.

So, 10B87 would mean 10.6.1 - Build 87.

~Adrian

[mikecwest]

Adrian Fogge, on Oct 11 2008, 11:41 PM, said:

Normal naming convention is that 'B' refers to the 10.x.1 release.

So, 10B87 would mean 10.6.1 - Build 87.

~Adrian

Hmm, I though B referred to beta, and A referred to Alpha? (A quick yahoo search indicates that you are correct, as 10.x.c would then be a 10.x.2 release?)  

I guess that might have been inaccurate information on the wikipedia?  Thanks for the tip on build numbers, I never noticed that before.

With that being the case....I notice that the numbers that follow the letter seem higher on later versions....does that number continue to increase and never start over at 0 until a new "cat" arrives?

[Adrian Fogge]

Builds like 9a528a mean 10.5.0, Day 528 (Revision A).

Right now, we should be on 10a219 for Snow Leopard.

~Adrian

[steve22]

That's interesting. Then why was build 10a127 mentioned on the launchd Trac last week?

Adrian Fogge, on Oct 11 2008, 07:06 PM, said:

Builds like 9a528a mean 10.5.0, Day 528 (Revision A).

Right now, we should be on 10a219 for Snow Leopard.

~Adrian

[Adrian Fogge]

Two things...

10a127 was mentioned 2008-09-17, not last week.
At that point, we are talking about a build that was created July 8, 2008 given what we know about 10a96.

Issues such as "20s stall at installer shutdown" are generally not the highest priority fixes around, and are generally fixed on the builds where they were first identified and then merged into the more recent code base after the issues are resolved and tested working where they were first identified.

Lower priority issues such as delays are more an inconvenience than ones such as "calling vproc_standby_end core dumps" and are thus put on hold for a while. As a matter of fact, both issues were resolved with the same changeset.

At least that is standard programming practices within my organization.

~Adrian

[steve22]

Thanks for the explanation.

~Matt

Adrian Fogge, on Oct 11 2008, 07:23 PM, said:

Two things...

10a127 was mentioned 2008-09-17, not last week.
At that point, we are talking about a build that was created July 8, 2008 given what we know about 10a96.

Issues such as "20s stall at installer shutdown" are generally not the highest priority fixes around, and are generally fixed on the builds where they were first identified and then merged into the more recent code base after the issues are resolved and tested working where they were first identified.

Lower priority issues such as delays are more an inconvenience than ones such as "calling vproc_standby_end core dumps" and are thus put on hold for a while. As a matter of fact, both issues were resolved with the same changeset.

At least that is standard programming practices within my organization.

~Adrian

[Capeman]

Things are starting to move along...

http://www.appleinsi..._imageboot.html