tehn
commented
7 years ago thank you for spotting this, and the fix sounds correct.
On Mon, Mar 6, 2017 at 9:24 AM, Sam Doshi notifications@github.com wrote:
As far as I can tell this bug has been there since the beginning. Or at least for quite a long time.
Consider the following
TR.PULSE 1 : TR.PULSE 2
That command doesn't make any sense as the PRE (i.e. before the :) has no MOD statement (e.g. IF). But it passes validation and runs!
What happens is the parser splits the entire command into a PRE and a POST, the validation step doesn't check to see if the first WORD of the PRE is a MOD. And finally the processor only executes the PRE (assuming that execution of the POST is delegated to the non-existent MOD).
The fix is fairly straightforward, we just need to insert a check at the end of validate to assert that if there is a PRE it must start with a MOD.
Unless I hear any objections I'll get it fixed in my parser branch.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/monome/teletype/issues/72, or mute the thread https://github.com/notifications/unsubscribe-auth/AAPEcBG8KN6fsHaQEkvIV094IFqg4lxNks5rjBc4gaJpZM4MUNiO .
samdoshi
As far as I can tell this bug has been there since the beginning. Or at least for quite a long time.
Consider the following
That command doesn't make any sense as the
PRE(i.e. before the:) has noMODstatement (e.g.IF). But it passes validation and runs!What happens is the parser splits the entire command into a
PREand aPOST, the validation step doesn't check to see if the firstWORDof thePREis aMOD. And finally the processor only executes thePRE(assuming that execution of thePOSTis delegated to the non-existentMOD).The fix is fairly straightforward, we just need to insert a check at the end of
validateto assert that if there is aPREit must start with aMOD.Unless I hear any objections I'll get it fixed in my parser branch.