At that point all the gaps would be filled. Then a new EXA would get sent to Drive 2, and do the same thing as the first Then one for Drive 3. If there are no files left, it would get passed a purposely invalid file number, crashing the EXA. ![]() then drop the file and return to the first step. read all the data from the file and pass it to the "writer" over global M receive one of the file numbers over global M The "reader" will send a 1 once it crosses the link, telling the "sender" that it made it if not, another is spawned to try again. Something to do with the way the EXAs were communicating.Īs I recall, my original solution was basically: Send out a "reader" EXA to Drive 1. Originally posted by thewifiwhisperer:Why would it not just crash while the first EXA is holding the file and trying to communicate?I don't remember exactly why the solution softlocked, but it softlocked. ![]() The comms itself is ended with an unambiguous end signal.Ĩ5 lines / 5027 cycles, room for improvement in these loops. I make sure to a nice END TRANS value (all 'real'-values are 4-digit).Ī complicated way of saying: I just did the TEST MRD route (I wait 2 ticks: one to cross the link, one to grab the file.), but only to check the first arrival of the EXA. First time, all is copied, second and third, only write whatever value is not 'FAIL'. The filemaster is an ugly beast, that has 3 repeated bits hardcoding for the 3 drives, sending out a 'copier', that messages the entire contents of the file. It waits for the local ok from the filemaster. 'central dispatch' holds the file list and per file spawns a 'filemaster' (in the controller host) I see that I have a bit of a 3-'tiered' system. , and opened immediately after, causing an unneeded EXA to be spawned and block up the solution.Why would it not just crash while the first EXA is holding the file and trying to communicate?
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |