The problem will GO away. The TDs need to upload DB format that is lining up with current parser. 
The incident with uploads from a Mac not working and the same files from a Windows PC working is somewhat disturbing, Leago has been made aware of that and has that set of files to test with.
Game stats has been unwell for over 24 hours now.
Not urgent but just as a design choice, this could be confusing. It looks like it is mixing actual games played (7) with supplement ratings (not in a supplement yet). For consistency, I think it should either be Current rating & Current games played or supplement rating and supplement games played.
I know you are doing all you can, Mike, but at the same time it is frustrating that there are so many bugs that are highly testable before going live (like stress testing performance and parsing the DBF files).
Stress testing is actually rather difficult to do well. My son sets up stress tests for major companies computer networks and it takes him months to design a full-blown stress test.
Leago was given over a thousand sets of upload files to test with and they’re in their automated testing routine so they can get checked with each feature rollout. (Some are files we know will fail because they’re corrupted.)
We’ve been asking people to send us DBF files that won’t upload and pass them on to Leago for testing and to have them added to the test suite. It looks like a common thread the last few days has been SwissSys, but the recent report of it working from a PC and not a Mac is a bit disturbing, because the files should be identical regardless of the browser used, which suggests some kind of timing issue.
So this is a SwissSys thing?
Even if it is, that isn’t an edge case. It is a pattern that shows up in every file that would have been tested and with the largest pairing program by usage.
They were - the files were literally off of a shared Google drive (although the Mac also tried locally). The Mac also tried three times, twice roughly at the same time and then a third time an hour later and failed each time, with the Windows computer succeeding at roughly the same time as the Mac failed the second time last night (around 10pm Central).
No, or at least that’s not the only issue.
In the Mac / PC report, the underlying software was WinTD.
I don’t think it’s really a SwissSys thing, but SwissSys account for about 50% of the upload files, and the errors have indicated they came from several different versions of the program.
Some years ago I wrote a program to diagnose DBF file issues, and a common thread in failed files seems to be that the file ends improperly with a Hex 00 byte at the end. This seems to happen mostly in the TH file, which is also interesting, as it’s the only one of the 3 DBF files that always has just one data record in it. There may be some DBF code library routines that have a bug in them.
Some bugs persist for years despite heavy usage. Some years ago I found and reported to Microsoft a bug in the Solitaire program, which is one of the most heavily used programs in PC history. The guy I sent it to took it to his weekly staff meeting and got a standing ovation for it. It turns out the bug was inadvertently fixed in the Windows 5 rewrite of that program but it was still astonishing that nobody had discovered that bug over several years.
The parser I wrote in 2005 is apparently more tolerant of minor issues than the one Leago wrote (I know they wrote a custom one as I did), but I don’t know where they are on testing the failed files I’ve sent them over the last 3 days, that’ll be one of the issues discussed in today’s regularly scheduled tag-up call.
Again, many of these are not edge cases. At a certain point, things cross over from understandable bugs to just plain sloppiness or worse.
I appreciate that you are working hard to help Leago fix the issues; however, we should not hand wave away what is going on here as normal software rollout pains
In computing terms, I’m a dinosaur. I”m 76, I started writing code in 1967 and have been doing it professionally since 1972 with a few years off to get my MBA (and even in grad school I kept on writing code on the side to help pay the bills.)
I’ve probably made about every kind of coding error you can make and I’ve probably offered about every kind of excuse you can offer for why something didn’t work, so I think my BS Detector is set quite high.
I’m going to close this topic because it’s getting kind of long and a bit off the subject.
New topics for specific issues people are having are fine.