Yoshiki,
I guess this tool only works on image of the same version and with a
change set only.
No, no, there is nothing that depends on specifically 3.2 or even 3.x, except a line in #openSourceFilesForImage: which specifies the .sources file name. As long as corresponding .changes exists, it should be possible to load V2 image onto V3 and vice versa.
I have quite a few of 1.x and 2.x and 3.x images on my machines (with full source and changes files) only the 3.x work as expected.
I don't quite get what you mean by "with a change set," but that the way it works.
Sorry, I meant .changes file. :-)
Is it doable ? Would you consider that as a feature request.
If you look at the class comment of ImageBrowser, you'll find the answer:-)
Just read it and got all that I want to know ;-)
Thanks for the tips, Yoshiki.
Cheers,
PhiHo
PS:
P.S. PhiHo, if you are going to reply my email, please include the entire part of this message. I still have a suspision that this one won't go through the normal channel. Thank you,
I received only 1 copy of your mail, so assumingly it did not get to the list.
----- Original Message ----- From: Yoshiki.Ohshima@acm.org To: squeak-dev@lists.squeakfoundation.org Cc: "PhiHo Hoang" phiho.hoang@rogers.com; sgelkins@bellsouth.net; cg@cdegroot.com; evan@telemessage.co.za Sent: Friday, November 01, 2002 4:06 AM Subject: Re: [ENH] Image Browser: for browsing code in the other .image file
PhiHo,
I guess this tool only works on image of the same version and with a
change set only.
No, no, there is nothing that depends on specifically 3.2 or even 3.x, except a line in #openSourceFilesForImage: which specifies the .sources file name. As long as corresponding .changes exists, it should be possible to load V2 image onto V3 and vice versa.
I don't quite get what you mean by "with a change set," but that the way it works.
It would be nice if I can use it to browse a headless image with no
source and change.
Yes.
Is it doable ? Would you consider that as a feature request.
If you look at the class comment of ImageBrowser, you'll find the answer:-)
I think it is not actually that hard. All you have to do is to write a modified version of InstructionStream and Decompiler which access the bytes and bits through the interpreter simulator's #longAt: and #fetchPointer:ofObject: and alike. The bytecodes are just same byte value. The literals have to be located in the simulated image, but it should be too hard.
-- Yoshiki
P.S. PhiHo, if you are going to reply my email, please include the entire part of this message. I still have a suspision that this one won't go through the normal channel. Thank you,