Very nice... thanks for the post. Your chapters have taken me through areas of the code I haven't imagined visiting before, which means at least for me they proved their worth.
Couple of notes: 1. I particularly liked the write barrier discussion in Chapter 7. It pointed to an area of code that looked too time consuming to understand on my own, but seems Ok now. I would have liked it a little expanded to describe how the GC uses the card tables (since, if I understand correctly, the cleanup is not "deterministic" - there could be marked card table bits that represent heap areas that have been collected already.) 2. Having an explanation of how things are is great - having an explanation of why things are the way they are is awesome. For instance, in the discussion on preemptive vs. cooperative GC, it would help to understand some of the complexities of the preemptive model in order to appreciate the decission MS made to go with the cooperative GC in SSCLI. (This discussion could also introduce some of the tradeoffs). In other words, your description was clear and accurate (as far as I could tell anyway) but it described a state of facts not the process that led to this state of facts. I find that describing the process that led you to your decision is very enlightening. 4. I was left slightly confused by the discussion of generations / heap_segments. Specifically, the text seems to suggest there is a 1:1 relationship between the two, but certain factors could dictate 1:n (maybe the case in the commercial CLR?) What is the rationale behind this design choice? (Again the 'why' versus the 'how'). 5. Maybe the book contains this elsewhere, but given how central MethodTables are, it would really help to have a diagram of the object layout. This is especially true of all objects that aggregate an "out-of- band", negative-offset object, like CGCDesc for MethodTable or ObjHeader for Object. These are difficult to track from code alone. 6. References to the generated assembler and well-chosen references to SSCLI code rock. It's great to be able to follow through code this way. Very nice job. Now, can you push the release date back a couple of months? :-) Cristian On Tue, 30 Jul 2002 16:55:31 -0700, Ted Neward <[EMAIL PROTECTED]> wrote: >David Stutz, Geoff Shilling and I, along with O'Reilly, are happy to >announce that two sample chapters of our forthcoming book, "Shared Source >CLI Essentials", are available on the O'Reilly website at >http://www.oreilly.com/catalog/sscliess/chapter/index.html. > >We're making these chapters available not only to offer a preview of what's >to come, but also as an opportunity for interested parties to offer up >feedback, criticisms, and reactions. And yes, we're still on track (so far, >at least) for the November 2002 release. > >Comments, questions, flames, criticisms--bring 'em on. :-) > >Ted Neward >{.NET || Java} Course Author & Instructor, DevelopMentor >(http://www.develop.com) >http://www.javageeks.com/tneward >http://www.clrgeeks.com/tneward
