Partition Manager
Pages: 1 ... 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
Jon Abbott (1421) 2641 posts |
What line was the error on?
Could you create a debug log please and post or email me the result. Set DEBUG%=TRUE in !RunImage |
Bryan (8467) 468 posts |
160 Drive_Info%(D%, D_Partitions%)=0 Interesting error reporting from BASIC |
Colin Ferris (399) 1811 posts |
If the Iyonix’s main disk was filecore formatted to say 128GB and then Fat32 formatted for the rest – would the Filecore 128GB run slow? |
Steve Pampling (1551) 8164 posts |
Hmm, if the disc was “partitioned”1 would the Iyonix know that there were two “discs” to look at on the one physical spinning rust? 1 My understanding is that RO knows absolutely diddly about partitions and assumes the first bit of disc space is it. Could be wrong, but the filesystem bounty has making the system partition aware as part of the work. |
Raik (463) 2059 posts |
Sorry, I forget … 136.
You find here |
David J. Ruck (33) 1631 posts |
No, the first 128GB is fast, the rest is slow. As long as the filecore partition is all within the first 128GB, it’s fast. |
Jon Abbott (1421) 2641 posts |
Unless I’m missing something, the debug logs don’t show any issues. ADFS::4 is the only drive showing a FAT partition and is 500GB, is that the drive you’re referring to? 160 Drive_Info%(D%, D_Partitions%)=0 Are you saying that line errors? What’s the error?
Line 136 in GPT.Partitions errors? |
Raik (463) 2059 posts |
I think … The GPT drive is connected to SCSI.
No. ADFS::4 is detected correct. I ignore Linux in RISC OS and mount the second partition with fat32fs:mount -fp2 :12 { > null: }. I changed the order of the devices. Plug GPT as ADFS::7 and “ADFSBackup” to SCSI. GPT is unknown now. Is the same output if I use the GPT device on ARMX6 and try PartMgr. |
Colin Ferris (399) 1811 posts |
Nice if BASIC gave the name of the variable missing. Is the Disc formatter a RO Prog – and available for download? |
Raik (463) 2059 posts |
I think you mean the FAT partition behind FileCore… |
Steve Drain (222) 1620 posts |
There is no variable where one would be expected, so no name available. ;-) |
Colin Ferris (399) 1811 posts |
FAT 42 the answer to everything – Hitch hiker guide to the Galaxy :-) Doesn’t BASIC know what Variable it’s looking for? |
Rick Murray (539) 13818 posts |
I’m still waiting for BASIC to report error line numbers when in the desktop (you’ll notice subtly different behaviour between that and command line use).
Yes, there is. |
Alan Adams (2486) 1147 posts |
Seconded. Here my approach to deal with the deficiency, is to add a *report on the line before and list all the variables in the erroring line. Reporter has the advantage of NOT breaking when the variable doesn’t exist, so it makes it possible to see which one (or more) is undefined. Reporter does break on array(y) when y doesn’t exist (or I may misremember, it might be when array() doesn’t exist) Fixing BASIC would be much better. |
Steve Drain (222) 1620 posts |
Oh, no there isn’t. ;-) Try something simple such as I agree that the four-decade old error reporting is showing its age. Perhaps printing the line of the error and highlighting the program pointer would be good. When programming with Reporter and throwback I do not have any problem with errors as they are. |
Martin Avison (27) 1491 posts |
I suspect the ‘normal’ error is where a variable name is mistyped, rather than omitted with WIDTH (or similar). Where BASIC has looked for a variable name but not found it, I think that when the error is detected registers hold the first character and the address of the rest of the string used. In those (common) cases it would seem possible to include the variable name in the message. Another similar message I find infuriating is |
Chris Hall (132) 3554 posts |
Perhaps we should add ‘but can’t be arsed to tell you what the filename was and I have erased my memory of its name now. Nerr!’ I can remember when the BASIC error messages were limited to two characters such as ‘FC error’ so we have it easy. |
Steve Pampling (1551) 8164 posts |
FAT 42 the answer to everything – Hitch hiker guide to the Galaxy :-) These are not the variables you are looking for :) (Star Wars the abused version) |
Steve Drain (222) 1620 posts |
I suspect there may be more omitted errors than you think, certainly for me.
I am sure it would, but first you would need two new errors for “Variable not found” and “Expression omitted”, which is the dual purpose of the current error. I suspect that the “Variable not found” part only arises from the |
Jon Abbott (1421) 2641 posts |
0.66 now available. I’ve resolved the “missing variable in GPT.Partitions line 146” issue and also fixed a bug that could cause partitions created in gaps to overlap the next partition. @Raik see if this resolves either the ADFS 7 partitions not being reported correctly, or the missing FAT after the ADFSBackup partition. |
John WILLIAMS (8368) 493 posts |
Thank you for adding the version number to the title bar – that will be a great help to us NetSurf downloaders!. And my RPi 50MB boot “partition” now shows-up as well! |
Raik (463) 2059 posts |
Sorry for this…
GPT nearly ok (partitions names are SSDp1 and SSDp2), ADFS 7 ok, but ADFS 5 + 6 are gone… |
Jon Abbott (1421) 2641 posts |
I wondered if fixing ADFS 7 might break something. It’s broken full disc FileCore drives formatted with recent versions of HForm. The problem is HForm writes a malformed MBR which has no partition entry for the FileCore area, so it’s indistinguishable from an unpartitioned MBR drive…hence ADFS 5 and 6 are now showing as unpartitioned. I’ll modify it to detect an MBR with “Castle” in the OS field that has no partitions, and just assume it’s full disc FileCore. I also need to completely rewrite the FileCore detection to not rely on the OS at all, which is a slightly bigger piece of work. I’ll try to get the former fixed later today. What did you use to partition SCSI 0:0? The GUID’s are the not the official FileCore value listed in PartMan’s source. |
Raik (463) 2059 posts |
Is the “backup mirror” of my ARMX6 SSD. |
Jon Abbott (1421) 2641 posts |
0.67 now available. It checks for an HForm generated MBR with no partitions and falls back to FileCore full disc.
The gpttest source uses the correct GUID. Did you self-compile it, or was it supplied with the MX6? I’m trying to understand if we have two “official” GUID’s for FileCore that I need to handle, or if yours is a one-off that I can ignore. |
Pages: 1 ... 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29