2022-01-31 14:20:49 <f-a         > https://planetasinclair.blogspot.com/2022/01/fig-forth-ts-2068-mia.html those were the days
2022-01-31 20:48:50 <a88         > I have words that can fail and I am returning (... -- <value> failed? ), failed? is 0 if it did not fail. I find it useful because in the calling word I have to test for failure, so I simply do WORD IF handle error THEN ...  any comments?
2022-01-31 20:49:21 <a88         > It seems a good arrangement to me, but I am open to any improvements
2022-01-31 20:57:47 <MrMobius    > one thing ive done is set an error number and do RDROP before returning so the called word returns not to what call it but to the word that called that
2022-01-31 20:58:09 <MrMobius    > that way you can assume anything that returns succeeded and only have to check for errors once at the top most level
2022-01-31 20:58:55 <MrMobius    > this introduces so inflexibilities though because the called word needs to know how many levels to RDROP and return to, so hard to make words general purpose that way
2022-01-31 21:04:21 <a88         > MrMobius, ah yes, I saw that trick of R> DROP in the Thinking forth book; I thought it was one of the specialties that you can do in forth. Like you I my concern is that callee needs to know something about caller levels, so I didn't use it; but worth considering.
2022-01-31 21:08:51 <MrMobius    > ya hard to do something like that in C
2022-01-31 21:11:35 <a88         > It gave me a great smile, thinking: "ha, look at the power of forth, you can even alter the return sequence after the fact"
2022-01-31 21:12:27 <MrMobius    > yep. its even worse than spaghetti code :P
2022-01-31 21:12:32 <MrMobius    > just kidding
2022-01-31 21:13:09 <a88         > :) that was one of my other concerns, I was thinking about maintainability
2022-01-31 21:13:43 <a88         > nonetheless, it is quite useful if you are sure your word will be used only in that calling context