Sounds good, Henry. I certainly have a use for that. But I can't see it in pacman for either JGTK or j602.
On Sun, Aug 12, 2012 at 9:10 PM, Henry Rich <[email protected]> wrote: > I have an alpha-test version of a syntax checker for J scripts. It is in > the addons, as > > general/misc/lint.ijs > > The goal is to catch errors before execution. It can find only gross > errors, but those are the kind my programming students make, so this is > progress IMO. What it looks for is: > > explicit definitions lacking trailing ), and ) with no definition > undefined names, including names not defined in all paths > verbs used with invalid valences > non-noun results at the end of condition blocks and verbs > syntax errors > sentences with no effect on execution (eg verb verb) > > The idea is that lint can be used in place of load. It loads the script, > then checks it for errors by executing each line of each explicit definition > in 'safe mode' where side effects are prevented. > > example: > > lint '~addons/general/misc/lint.ijs' > > The result of the verb is a table of (line number);(error message). > If there are errors, it brings up a grid showing the program and error > messages. > > I would test this a lot more before releasing it, but school starts tomorrow > & I'm not going to have much more time to put on it. I think it's better > than nothing - not freshman-proof, but I'll fix that as the freshmen tear > into it. Meanwhile I would appreciate suggestions and bug reports. > > (My thanks to Dan King who suggested this project at the J Conference in > response to my bellyaching about J's error-detection deficiencies) > > Henry Rich > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
