Possible bug in SequencePoint - StartLine value off the charts #80

New issue

Closed

opened 2011-10-28 00:48:34 -05:00 by provegard · 2 comments

provegard commented

2011-10-28 00:48:34 -05:00

(Migrated from github.com)

I've been using Cecil in a hobby project, and overall it works great! I found something odd when studying the sequence points of one of my test methods. The start lines of the sequence points are:

nop : 13
newobj : 14
newobj : 16
ldloc.1 : 17
ldloc.2 : 16707566
ldstr : 20
br.s : 21
ldstr : 23
br.s : 24
ret : 26

The ldloc.2 instruction stands out a bit there... It seems to me that it's a bug, but I know too little about Cecil and CIL to say for sure what's going on.

I have created a tiny VS2010 solution that demonstrates the problem. Simply run from within VS and you should get the printout as above:

https://github.com/provegard/CecilSeqPointBug

I've been using Cecil in a hobby project, and overall it works great! I found something odd when studying the sequence points of one of my test methods. The start lines of the sequence points are: nop : 13 newobj : 14 newobj : 16 ldloc.1 : 17 ldloc.2 : 16707566 ldstr : 20 br.s : 21 ldstr : 23 br.s : 24 ret : 26 The ldloc.2 instruction stands out a bit there... It seems to me that it's a bug, but I know too little about Cecil and CIL to say for sure what's going on. I have created a tiny VS2010 solution that demonstrates the problem. Simply run from within VS and you should get the printout as above: https://github.com/provegard/CecilSeqPointBug

jbevain commented

2011-10-28 00:59:45 -05:00

(Migrated from github.com)

Hey,

Thanks for the bug and the repro, but this time I'm lucky, it's no bug.
See http://blogs.msdn.com/b/jmstall/archive/2005/06/19/feefee-sequencepoints.aspx

Thanks!

Hey, Thanks for the bug and the repro, but this time I'm lucky, it's no bug. See http://blogs.msdn.com/b/jmstall/archive/2005/06/19/feefee-sequencepoints.aspx Thanks!

provegard commented

2011-10-28 01:24:16 -05:00

(Migrated from github.com)