acmfiu@serss0.fiu.edu (ACMFIU) (11/16/90)
I have the format of OMF v2.1 headers but NO ONE is outputting them. therefore, could someone relay the differences in the headers between v2.0 and 2.1. also, if there are any differences in the expressions (i.e. EXPR, etc), post those also. albert
gwyn@smoke.brl.mil (Doug Gwyn) (11/17/90)
In article <1720@kluge.fiu.edu> acmfiu@serss0.fiu.edu (ACMFIU) writes: >I have the format of OMF v2.1 headers but NO ONE is outputting them. >therefore, could someone relay the differences in the headers between >v2.0 and 2.1. also, if there are any differences in the expressions >(i.e. EXPR, etc), post those also. As I recall, apart from bugs in Apple's documentation, the only real difference from 2.0 to 2.1 is that 2.1 provides support for another form of ORG, used for IIGS cross-development under MPW. Mike Westerfield of ByteWorks tells me he argued with Apple DTS about this, claiming that it wasn't necessary, but DTS went ahead and added it.
acmfiu@serss0.fiu.edu (ACMFIU) (11/17/90)
In article <14508@smoke.brl.mil> gwyn@smoke.brl.mil (Doug Gwyn) writes: >In article <1720@kluge.fiu.edu> acmfiu@serss0.fiu.edu (ACMFIU) writes: >>I have the format of OMF v2.1 headers but NO ONE is outputting them. >>therefore, could someone relay the differences in the headers between >>v2.0 and 2.1. also, if there are any differences in the expressions >>(i.e. EXPR, etc), post those also. > >As I recall, apart from bugs in Apple's documentation, the only real >difference from 2.0 to 2.1 is that 2.1 provides support for another >form of ORG, used for IIGS cross-development under MPW. Mike >Westerfield of ByteWorks tells me he argued with Apple DTS about >this, claiming that it wasn't necessary, but DTS went ahead and >added it. I thought so too but it seems the KIND field has been changed. at least when i look at an OMF v2.0 file and see the position of the KIND field it's NOT in the position the OMF v2.1 docs say it suppose to be. Also, the format of the KIND field, if it's in the right place in OMF v2.0, is different. To see for yourself, just 'dumpobj -n ~GLOBALS -o <prog>' and you will see a KIND field of $40, 'private static code segment'. now look at OMF v2.1 specs of the KIND field and $40 IS NOT the representation of 'private static code segment'. mike did argue about the tempORG field. it does what 'obj' does in orca/m so he saw no reason for it. albert
gwyn@smoke.brl.mil (Doug Gwyn) (11/17/90)
In article <1730@kluge.fiu.edu> acmfiu@serss0.fiu.edu (ACMFIU) writes: >I thought so too but it seems the KIND field has been changed. Hmm, I thought that changed between 1.0 and 2.0, not 2.0 and 2.1. I guess I'll have to go look this up. By the way, quite some time ago I started writing my own implementation of DumpObj, but due to the large number of bugs in ORCA/C 1.0 I gave up on it. Perhaps with ORCA/C 1.1 it would be worth resuming.