freenode/#mezzano - IRC Chatlog

21:54:29 ebrasca froggey: PR done! I just left license wichout changes.

22:11:25 ebrasca I going to start working on Loopback Device for Mezzano.

23:07:44 ebrasca fittestbits__: Hi

23:08:51 fittestbits__ Hi

23:10:31 ebrasca fittestbits__: It is strange for me check-gpt-crc is changed.

23:11:36 ebrasca fittestbits__: I thinked extended partition can only be last one.

23:12:43 fittestbits__ Not according to https://thestarman.pcministry.comasm/mbrPartTables2.htm

23:13:31 ebrasca fittestbits__: Can't open this link. "Hmm. We’re having trouble finding that site"

23:14:08 fittestbits__ Yeah, looks bad, just a minute ...

23:15:06 fittestbits__ The "/"s got lost https://thestarman.pcministry.com/asm/mbr/PartTables2.htm

23:15:44 ebrasca fittestbits__: What do you think about crc32c ?

23:16:13 fittestbits__ The original check-gpt-crc does (let ((header-size ...) (expected-crc ...))

23:16:32 fittestbits__ then does (setf (get-header-crc buf) 0)

23:16:49 fittestbits__ then does (let ((actual-crc ...)) ...)

23:17:10 fittestbits__ The new code just does (let ((header-size ...) (expected-crc ...) (actual-crc ...)) ...)

23:17:25 fittestbits__ But doesn't zero the crc in the buffer ...

23:18:05 fittestbits__ Have you gotten the crc32c to work? I have a different version, but it's failing on my mezzano ext4 partition.

23:18:24 ebrasca ACTION is thinking , maybe some day he is going to be error free.

23:19:01 ebrasca fittestbits__: I only tested 1 input with crc32c , used this link https://crccalc.com/

23:19:30 fittestbits__ And, strangely, my linux ext4 partitions don't actually have crcs! the CRC type field are 0 (instead of 1) and the CRC value fields are 0.

23:20:05 fittestbits__ Yeah, I version of crc32 is getting the right value for a simple test case ... but I haven't figured out what's going on with my mezzano ext4 partition.

23:20:10 fittestbits__ So, I haven't released it.

23:20:27 fittestbits__ I version -> my version

23:20:43 ebrasca fittestbits__: I plan to add some write to ext4.lisp , for now metadata.

23:20:58 ebrasca If you like I can delete my PR.

23:22:39 ebrasca fittestbits__: How is your crc32c version?

23:23:20 fittestbits__ I suggest testing the crc32c code on an ext4 partition - either on mezzano, or by using the data from some other ext4 superblock.

23:24:04 ebrasca Good idea.

23:24:04 fittestbits__ The only ext4 I've found with a crc is the one I'm using on mezzano. So, I'm not sure the crc is valid.

23:24:39 fittestbits__ Also, I know how to fix the defconstant problem ... just a second while I look it up ...

23:25:38 fittestbits__ Try (eval-when (:compile-toplevel :load-toplevel :execute) (defconstant ...) (defconstant ...)) to fix the cold boot issue.

23:28:33 ebrasca fittestbits__: It keep giving "The function MEZZANO.DISK::MAKE-CRC-TABLE is undefined." in cold boot.

23:29:37 fittestbits__ Oh, include that function in the eval-when form.

23:34:02 ebrasca fittestbits__: Now I get http://ix.io/2tWb

23:35:32 fittestbits__ That's a problem with SBCL. I guess go back to defvar.

23:36:49 ebrasca with "(if (boundp '+crc32-table+) (symbol-value '+crc32-table+) (make-crc-table +crc32-polynomial+))" does work!

23:37:05 ebrasca At leat in make cold-image

23:37:45 fittestbits__ So progress ...

23:40:24 ebrasca fittestbits__: How are you doing?

23:40:53 fittestbits__ Good. Haven't been writing much code lately ... doing other stuff.

23:40:59 fittestbits__ How are you?

23:42:14 ebrasca fittestbits__: I am file. I have make cache.lisp block based ( Now it don't need to load full file to ram )

23:42:43 fittestbits__ I saw that. A good step forward.

23:43:46 ebrasca Do you like to make optimized read/write-sequence for it?

23:46:44 fittestbits__ Any idea how often read/write-sequence is used?

23:47:37 fittestbits__ In tag-fs, I left them as a loop just being lazy.

23:48:37 ebrasca Can tag-fs work with cache.lisp ?

23:50:25 fittestbits__ The streams code in tag-fs is already block oriented. So, I haven't looked at how it would work with cache.lisp.

23:51:31 ebrasca I think the more FSs use cache.lisp the more care it is going to get.

23:53:03 fittestbits__ I wrote a pretty good test for the block code. Take a look at file-fuzzing-test in http://ix.io/2tWh

23:53:46 fittestbits__ It does seeks in both directions reading and writing random number of bytes.

23:54:02 fittestbits__ At the end it checks the file against a memory buffer that should contain the same info.

23:54:30 fittestbits__ It assumes that when you seek past the end of the file, the bytes are written as zero.

23:55:17 fittestbits__ You'll want the file size to be much larger than the block size and the number of iterations to be large enough that a lot of file is written.

23:55:21 ebrasca I am not going to read it today.

23:55:27 ebrasca It is very interesting.

23:55:28 fittestbits__ That's OK.

23:55:38 fittestbits__ I have to go now. TTYL

23:55:54 ebrasca fittestbits__: Bye!

23:58:48 MetaYan ebrasca: Just a tiny thing: Happened to notice (sup:debug-print-line "Detected MBR partition table" disk) in partition.lisp not being formatted like its GPT counterpart.

0:01:10 ebrasca MetaYan: I think it does (sup:debug-print-line "Detected GPT on disk " disk)

0:05:46 MetaYan ebrasca: I just meant the space at the end. ;)

0:12:54 ebrasca MetaYan: Thx , now it is this (sup:debug-print-line "Detected GPT partition table " disk)

0:22:51 catonano_ ** NICK catonano

0:23:39 MetaYan ebrasca: Oh, I meant that the MBR should be like the GPT.

0:23:49 ebrasca ok

0:25:38 MetaYan Got to go. Bye.