search

tabemann / zeptoforth

A not-so-small Forth for Cortex-M
MIT License
192 stars 18 forks source link

Zeptofort 1.7.0: Error while writing file with Zeptoed (external memory card) #123

Open mcctuxic opened 4 months ago

mcctuxic commented 4 months ago

Setup: Linux (GENTOO, Linux kernel 6.10.2) kitty 0.35.2 zsh 5.9 picocom 3.1 Raspberry Pie Pico W Zeptoforth 1.7.0

Setting up the Pico W: Using Zeptoforth 1.7.0 big full usb uf2 to flash the forth system. entering the forth system via picocom entering compile-to-flash mode leaving picocom without resetting the tty (CTRL-a CTRL-q) from the linux console (kitty/zsh) utils/codeload3.sh -B 115200 -p /dev/ttyACM0 serial extra/common/fixed32.fs utils/codeload3.sh -B 115200 -p /dev/ttyACM0 serial extra/common/zeptoed_all.fs entering the forth system via picocom compile-to-ram reboot entering the forth system via picocom

(after preparing the sd<>-card/the fs) s" file.fs" zed

CTRL-W message appears: Wrote file in the midth of the screen of zed it prints in red "**heap internal error**" This happens not all the time but it happens twice this morning after installing Zeptoforth 1.7.0 The file seems to be ok, though.
mcctuxic commented 4 months ago

I think, I found what triggers the error.

Setup as before.

s" newfile.fs" zed Added some random lines. Important seems to add lines only. No removal! CTRL-W CTRL-V no problems so far

s" newfile.fs" zed now delete some lines CTRL-W and when pressing CTRL-V either one gets an "heap internal errror" and then the system hangs hard (need to use the reset button) or it hangs instantly.

mcctuxic commented 4 months ago

Got

HARDWARE EXCEPTION, RETURNING TO PROMPT

...what happens here?

mcctuxic commented 4 months ago

I nuked my board and setup everything again. Cannot say, whether it fixes or changes something yet. But fscking the memory card, which I freshly partioned and formatted before starting using it with Zeptoforth gives me this message:

sudo fsck.vfat /dev/sdb1 fsck.fat 4.2 (2021-01-31) /TEST Invalid '..' entry in the second slot. Fixing. /TEST Has a large number of bad entries. (57/64)

/TEST is a directory in the root of that card.

mcctuxic commented 4 months ago

So... Summa summarum...

After encountering the above issue and getting the HARDWARE EXCEPTION, RETURNING TO PROMPT exception, I nuked the flash of the board and reformatted the memory card, which had fs-error according to fschk.vfat.

I reinstalled zeptoforth 1.7.0 big full usb uf2. I installed fixed32 and zeptoed (via the _all.fs-file) into flash as described above. I rewrote block 10...13 to flash from a backup via copy and paste.

The blocks have this content: B10:

11 13 load-range

B11:

oo import pin import simple-fat32 import fat32-tools import

class-size buffer: my-fs 10 11 12 15 1 my-fs init-object my-fs current-fs! : ls [inlined] list-dir ; : md [inlined] create-dir ; : cp [inlined] copy-file ; : cat [inlined] list-file ; : rm [inlined] remove-file ; : clr [inlined] page ; B12: : dec [inlined] decimal ; : oct [inlined] octal ; : bin [inlined] binary ; : cls depth 0 ?do drop loop ; xon-xoff-enabled enable-line page B13: 0 UART::DISABLE-UART 0 output-pin low 0 pin! I wrote the files I backuped from the memory card back onto the memory card. I use picocom -f x -b 115200 --logfile zeptolog.txt /dev/ttyACM0 to connect to the board. But when using zeptoeds CTRL-W to write a changed file I get "heap internal error" still when hitting CTRL-V to leave zed.
mcctuxic commented 4 months ago

Sorry for that many comments Travis, but... :)

I did the same with the previous version (1.6.4.3) and the above error with CTRL-W/CTRL-V no longer appeards. So...I think, my hardware is ok.....

tabemann commented 4 months ago

On Sun, Jul 28, 2024 at 4:06 AM mcctuxic @.***> wrote:

Sorry for that many comments Travis, but... :)

Oh, no need to be sorry.

I did the same with the previous version (1.6.4.3) and have no problems with zed. So I think, my hardware is ok.....

I was going to say that I was having trouble reproducing the issue... then I got it to crash. I will be looking into it further.

Travis

Message ID: @.***>

tabemann commented 4 months ago

On Sun, Jul 28, 2024 at 12:42 PM Travis Bemann @.***> wrote:

On Sun, Jul 28, 2024 at 4:06 AM mcctuxic @.***> wrote:

Sorry for that many comments Travis, but... :)

Oh, no need to be sorry.

I did the same with the previous version (1.6.4.3) and have no problems with zed. So I think, my hardware is ok.....

I was going to say that I was having trouble reproducing the issue... then I got it to crash. I will be looking into it further.

I was only able to get it to crash once, and since then have had no problems, and I don't remember what I exactly did when it did crash. Yet you seem to have been able to get it to crash reliably. Could you write down your exact steps that are triggering a crash, so I can try to reproduce it reliably?

Thanks,

Travis

Message ID: @.***>

tabemann commented 4 months ago

On Sun, Jul 28, 2024 at 12:53 PM Travis Bemann @.***> wrote:

On Sun, Jul 28, 2024 at 12:42 PM Travis Bemann @.***> wrote:

On Sun, Jul 28, 2024 at 4:06 AM mcctuxic @.***> wrote:

Sorry for that many comments Travis, but... :)

Oh, no need to be sorry.

I did the same with the previous version (1.6.4.3) and have no problems with zed. So I think, my hardware is ok.....

I was going to say that I was having trouble reproducing the issue... then I got it to crash. I will be looking into it further.

I was only able to get it to crash once, and since then have had no problems, and I don't remember what I exactly did when it did crash. Yet you seem to have been able to get it to crash reliably. Could you write down your exact steps that are triggering a crash, so I can try to reproduce it reliably?

Okay, I have made some progress. I have to make files of sufficient size and then either attempt to close them or close zeptoed. Clearly the heap is getting corrupted at some point, and the reason why you weren't getting these errors with 1.6.4.3 is that older versions of zeptoed made no effort to clean up after themselves as there was no need to close open files.

Travis

Message ID: @.***>

tabemann commented 4 months ago

On Sun, Jul 28, 2024 at 1:34 PM Travis Bemann @.***> wrote:

On Sun, Jul 28, 2024 at 12:53 PM Travis Bemann @.***> wrote:

On Sun, Jul 28, 2024 at 12:42 PM Travis Bemann @.***> wrote:

On Sun, Jul 28, 2024 at 4:06 AM mcctuxic @.***> wrote:

Sorry for that many comments Travis, but... :)

Oh, no need to be sorry.

I did the same with the previous version (1.6.4.3) and have no problems with zed. So I think, my hardware is ok.....

I was going to say that I was having trouble reproducing the issue... then I got it to crash. I will be looking into it further.

I was only able to get it to crash once, and since then have had no problems, and I don't remember what I exactly did when it did crash. Yet you seem to have been able to get it to crash reliably. Could you write down your exact steps that are triggering a crash, so I can try to reproduce it reliably?

Okay, I have made some progress. I have to make files of sufficient size and then either attempt to close them or close zeptoed. Clearly the heap is getting corrupted at some point, and the reason why you weren't getting these errors with 1.6.4.3 is that older versions of zeptoed made no effort to clean up after themselves as there was no need to close open files.

I am pretty sure I have this sorted now ─ I have made a new release ( https://github.com/tabemann/zeptoforth/releases/tag/v1.7.1) which appears to fix the zeptoed issues. This release also includes new words for simplifying accessing files and directories in FAT32 filesystems.

Travis

Message ID: @.***>

mcctuxic commented 4 months ago

On 07/28 09:08, tabemann wrote:

On Sun, Jul 28, 2024 at 1:34 PM Travis Bemann @.***> wrote:

On Sun, Jul 28, 2024 at 12:53 PM Travis Bemann @.***> wrote:

On Sun, Jul 28, 2024 at 12:42 PM Travis Bemann @.***> wrote:

On Sun, Jul 28, 2024 at 4:06 AM mcctuxic @.***> wrote:

Sorry for that many comments Travis, but... :)

Oh, no need to be sorry.

I did the same with the previous version (1.6.4.3) and have no problems with zed. So I think, my hardware is ok.....

I was going to say that I was having trouble reproducing the issue... then I got it to crash. I will be looking into it further.

I was only able to get it to crash once, and since then have had no problems, and I don't remember what I exactly did when it did crash. Yet you seem to have been able to get it to crash reliably. Could you write down your exact steps that are triggering a crash, so I can try to reproduce it reliably?

Okay, I have made some progress. I have to make files of sufficient size and then either attempt to close them or close zeptoed. Clearly the heap is getting corrupted at some point, and the reason why you weren't getting these errors with 1.6.4.3 is that older versions of zeptoed made no effort to clean up after themselves as there was no need to close open files.

I am pretty sure I have this sorted now ─ I have made a new release ( https://github.com/tabemann/zeptoforth/releases/tag/v1.7.1) which appears to fix the zeptoed issues. This release also includes new words for simplifying accessing files and directories in FAT32 filesystems.

Travis

Hi Travis,

...how many hours of sleep do you get between DO and LOOP or even +LOOP ? :) ;)

An extra big THANK YOU for this super fast update, Travis!

I have already updated my two RP2040 boards.

Cheers! Tuxic

Message ID: @.***>

-- Reply to this email directly or view it on GitHub: https://github.com/tabemann/zeptoforth/issues/123#issuecomment-2254902946 You are receiving this because you authored the thread.

Message ID: @.***>

tabemann commented 3 months ago

On Mon, Jul 29, 2024, 11:19 mcctuxic @.***> wrote:

Hi Travis,

...how many hours of sleep do you get between DO and LOOP or even +LOOP ? :) ;)

More than enough!

An extra big THANK YOU for this super fast update, Travis!

You're welcome! I try to respond to bug reports in a timely manner.

I have already updated my two RP2040 boards.

I may make another release soon even without bug reports because I have noticed some weirdness with IMPORT-FROM but I did not want to wait on fixing zeptoed.

Travis

mcctuxic commented 3 months ago

On 07/29 01:00, tabemann wrote:

On Mon, Jul 29, 2024, 11:19 mcctuxic @.***> wrote:

Hi Travis,

...how many hours of sleep do you get between DO and LOOP or even +LOOP ? :) ;)

More than enough!

An extra big THANK YOU for this super fast update, Travis!

You're welcome! I try to respond to bug reports in a timely manner.

I have already updated my two RP2040 boards.

I may make another release soon even without bug reports because I have noticed some weirdness with IMPORT-FROM but I did not want to wait on fixing zeptoed.

Travis

I will preheat my Flash-O-mat, engage the extra fusion power adapter and switch on the Zero-point module now to be prepared!

Let's flash! :)

Cheers! Tuxic

-- Reply to this email directly or view it on GitHub: https://github.com/tabemann/zeptoforth/issues/123#issuecomment-2256790196 You are receiving this because you authored the thread.

Message ID: @.***>