Assembler bugs
Wilton
Posts: 11
I'm not sure if there's a more appropriate place to raise this, but I've encountered three minor bugs in the assembler:
1.· When a bad label (duplicate in most recent case) is encountered, it stops adding symbols to the table or something, causing any labels after that point to not be recognized and references to them to be undefined.· This creates spurious errors that will go away once the bad label is fixed.
2.· Somehow the editor managed to get some text at the beginning of a line that it didn't display and the assembler didn't like.· It kept flagging the two lines as errors with extra characters, but I couldn't get rid of the problem.· I solved it by opening the file in Notepad where the first character in each bad line was shown as a little square box.· Deleting it and saving the file solved the problem.
3.· This is way to vague, but I encountered a situation where an error caused the wrong line to be referenced hand highlighted.· It seemed to have something to do with not counting blank lines near the error or something.· I found the error (as I recall) several lines above where it claimed it was by commenting out lines until it went away.· Sorry, that was a couple of days ago and I don't have an example.
Wilton
·
1.· When a bad label (duplicate in most recent case) is encountered, it stops adding symbols to the table or something, causing any labels after that point to not be recognized and references to them to be undefined.· This creates spurious errors that will go away once the bad label is fixed.
2.· Somehow the editor managed to get some text at the beginning of a line that it didn't display and the assembler didn't like.· It kept flagging the two lines as errors with extra characters, but I couldn't get rid of the problem.· I solved it by opening the file in Notepad where the first character in each bad line was shown as a little square box.· Deleting it and saving the file solved the problem.
3.· This is way to vague, but I encountered a situation where an error caused the wrong line to be referenced hand highlighted.· It seemed to have something to do with not counting blank lines near the error or something.· I found the error (as I recall) several lines above where it claimed it was by commenting out lines until it went away.· Sorry, that was a couple of days ago and I don't have an example.
Wilton
·
Comments
As always with these sort of things, providing short source files that demonstrate the problem goes a long way toward getting a bug fixed. In many cases, it's the only way to get a bug fixed.
The first bug you describe (duplicate labels causing problems) is something where it would help a lot to have a sample file that shows the problem. Without this, I'm just guessing at how to re-create your problem. As a result, it won't get fixed without a sample.
The second bug is something I've only recently heard about. Apparently you got a control character into the source file. The IDE doesn't show non-printable characters, so it appears invisible. JohnnyMac recently sent me a reproducible sample of this error, so it's something I'm looking into.
As you can imagine, the third problem is too vaguely described to try and fix. It's not clear what the problem is, so if it happens again, save the file and post it here.
Thanks,
PeterM
Sorry for not getting back to this earlier. I upgraded my DSL and they had to turn it off for a week to do it. Why? I don't know. Ask the phone company.
Anyway, I'm back online. I really need you to create a small source file that demonstrates the problem. Stepping through thousands of lines of assembly code in SASM under the Visual C++ debugger is not something that will happen. Create the smallest program that demonstrates the problem and I can work on it. Hand me a behemoth, and it will unfortunately go to the bottom of the work stack and probably never see the light of day.
You will likely find that in creating the small source file, you learn more about the problem and can provide me with more clues.
Thanks,
PeterM
I tried to create a program from scratch that had the problem, thinking it would be easy--I failed.· So then I stripped out a bunch of stuff from the previous code keeping an eye on what was going on.· If you change the first instance of SerStart to something else, say SerStart2, it will compile error free.· However as is, it generates three errors, the last two of which are spurious.
Based on my attempt to build an example from scratch, I would guess that the org statements have a role in the failure, since that is the only thing I can think of that I didn't try in the "from scratch" attempt.· It had one at 0, but nowhere else.· The other (related) possibility is that the @ operator may be involved, since the "from scratch" attempt didn't need that.
Wilton
P.S: Sorry if I attached it twice.· The attachment manager is a litle brain dead:
··· 1.· It doesn't give any indication that it has done its job
··· 2.· It doesn't go away when its done or have a mechanism for removing it other than the close box
··· 3.· The Reply Post areas does not give any indication of what (if anything) is attached.
Flying blind, I wasn't sure the first one worked, although it probably did.
36 lines? Excellent! I'll download this and see what I can find. It will take a little while as I have a number of other things going on right now, but I will investigate.
Thanks,
PeterM