Originator said...
I see whats up, when you save as, there is no spin at the end on my PC, you must add it manually each time, but that works fine so I wouldn't bother.
When you say "PC". Linux/Windows - Version?
I'll make sure its fixed in the next release anyway.
Brad, This code gives a compile error, but the cursor ends up on the wrong line. Mac Version, BST V0.12, Compiler V0.11
con
QPOS = $08 'Query Position
QSPD = $10 'Query Speed
CHFA = $18 'Check for Arrival
TRVL = $20 'Travel Number of Positions
CLRP = $28 'Clear Position
SREV = $30 'Set Orientation as Reversed
STXD = $38 'Set TX Delay
SMAX = $40 'Set Speed Maximum
SSRR = $48 'Set Speed Ramp Rate
'--- User Constants & Variables ---
AllWheels = 0
RightWheel = 1 'ID of the right side Position Controller
LeftWheel = 2 'ID of the left side Position Controller
CommPin = 12 'communication bus pin
BaudValue = 32 'for 19.2kbps
rxPin = 1
txPin = 2
var
long Wheel 'Specifies which wheel to command for subroutines
long Distance 'Used to set the travel distance
long RxData 'Used to receive data
obj
G : "cr_Globals"
encoder : "FullDuplexSerial"
PUB start : okay
'' Starts uart object (at baud specified) in a cog
'' -- returns false if no cog available
okay := encoder.start(rxpin, txpin, %0000, 19200) : okay
Can you elaborate Chuck? I had to comment out
G : "cr_Globals" but it gives me an error at
(40,55) which is correct in my listing as the ":" between the ")" and the "okay"
If I don't comment out the G : " cr_Globals" then it just gives me an error about being unable to locate that object with no line locations (which is a bug in itself really)
The error is at the location listed in the message (I forgot to delete the ": ok" after I copied the start syntax from the object) . The problem was that I did not look at the error location numbers at first, I looked at the spot where it left the insertion point. This was 5 or 6 lines above the true error location. As a guess, maybe when it calculates the location to insert the cursor, it does not take the mac's two byte line endings?
Another problem. Annoyance really. I have the folding turned on, which is nice, but I often use a double click to select a whole word. When I do this on any word in the PUB line, it collapses in addition to selecting the word. It would be better if it only collapsed the line if I clicked on the [noparse][[/noparse]+].
It would also be nice, if a double click on a word followed by a drag would select the second and third, etc. words in addition to the first word. Currently it selects the first word, but after that, it only ads characters to the selections.
A Triple Click on a line should select the entire line, including leading tab/spaces and the CR/LF. Currently a triple click only selects the words, so to duplicate 5 lines, I have to triple-click, then paste and fix, paste and fix, etc. I should only have to Paste, Paste, Paste, Paste, Paste.
Chuck Rice said...
The error is at the location listed in the message (I forgot to delete the ": ok" after I copied the start syntax from the object) . The problem was that I did not look at the error location numbers at first, I looked at the spot where it left the insertion point. This was 5 or 6 lines above the true error location. As a guess, maybe when it calculates the location to insert the cursor, it does not take the mac's two byte line endings?
I can snap a screen shot if that will help.
Actually, yeah a screen shot would help heaps here.
As for the other niggles, I'll certainly look into them tomorrow for you and try to get them fixed before the next release.
Just an opinion, when double clicking on the first line of the block to toggle fold or unfold, there are many times I double click on a word to highlight the word, and having the fold/unfold turns into a battle of wills sometimes. I will get used to it, but I think the fold icon is sufficient.
TChapman said...
Just an opinion, when double clicking on the first line of the block to toggle fold or unfold, there are many times I double click on a word to highlight the word, and having the fold/unfold turns into a battle of wills sometimes. I will get used to it, but I think the fold icon is sufficient.
I've removed that feature and fixed the triple click line select that was bothering Chuck.
The first was there by design. I don't use double/triple clicking to select stuff so I figured it was a good way to do fold/unfold. I'm obviously not correct [noparse];)[/noparse]
I have to reproduce / fix the cursor being in the wrong place when highlighting an error yet, but I think I know where the problem might be.
The double-click select word not selecting by word when you drag is something I'm going to have to get into the bowels of the editor code to fix, so it's not likely to happen very soon I'm afraid.
There are enough bugs and annoyances there now to justify another release.
Chuck Rice said...
Here is is. It may be that it is just not setting the cursor, so it is left wherever I had it last.
Edit: The cursor does not show, but it is under RxData in the var section where the pink artifact was left.
Quick question. Are you clicking on the error in the error box to highlight the error? It does nothing to the editor or cursor until the error is clicked on.
BradC said...
Quick question. Are you clicking on the error in the error box to highlight the error? It does nothing to the editor or cursor until the error is clicked on.
Oops. No. There is only 1 error, so I assumed that it would be selected.
BradC said...
Quick question. Are you clicking on the error in the error box to highlight the error? It does nothing to the editor or cursor until the error is clicked on.
Oops. No. There is only 1 error, so I assumed that it would be selected.
No, it wasn't. I've set it up now so it will automatically select the first error though. Thought that sounded like a good idea.
I've got a clipboard bug with Ubuntu 8.10 to finish and I'll have a new release out. It's proving to be a bit of a problem though so I'm not sure when I'll get it buried.
On XP using bst 12, I am experiencing that if I highlight several lines to delete, lines get deleted some number of rows below where I have highlighted. If I go to one line only, I can delete that line ok. This is with line number in on, and most code folded up except the few consecutive blocks that I am tweaking.
Another observation:
If I accidentally erase or leave out a ] or ) on a line, the compile error says 1,1 always, not the line that the error appears on. I have since learned how to work around these errors but thought I'd mention it in case you wanted all bugs listed.
On XP using bst 12, I am experiencing that if I highlight several lines to delete, lines get deleted some number of rows below where I have highlighted. If I go to one line only, I can delete that line ok. This is with line number in on, and most code folded up except the few consecutive blocks that I am tweaking.
I can't reproduce this one yet, but I'll keep trying.
TChapman said...
Another observation:
If I accidentally erase or leave out a ] or ) on a line, the compile error says 1,1 always, not the line that the error appears on. I have since learned how to work around these errors but thought I'd mention it in case you wanted all bugs listed.
Wow, can't believe I missed that one. It's an ugly artifact of the way I parse brackets. I'll make sure its taken care of. Thanks!
On XP using bst 12, I am experiencing that if I highlight several lines to delete, lines get deleted some number of rows below where I have highlighted. If I go to one line only, I can delete that line ok. This is with line number in on, and most code folded up except the few consecutive blocks that I am tweaking.
I can't reproduce this one yet, but I'll keep trying.
I have seen something like it. If i quadruple-click on a line, it will select the whole line including leading whitespace and the line ending. if you then press delete, it deletes that line and the following paragraph. This was in the Mac version, and I was deleting a comment line. Undo undeleted all the deleted lines.
Despite the minor issues you are already working on, just wanted to say "THANK YOU" for a fine
editor/compiler for Linux. I'm using it exclusively now for all my spin projects.
On XP using bst 12, I am experiencing that if I highlight several lines to delete, lines get deleted some number of rows below where I have highlighted. If I go to one line only, I can delete that line ok. This is with line number in on, and most code folded up except the few consecutive blocks that I am tweaking.
I can't reproduce this one yet, but I'll keep trying.
I have seen something like it. If i quadruple-click on a line, it will select the whole line including leading whitespace and the line ending. if you then press delete, it deletes that line and the following paragraph. This was in the Mac version, and I was deleting a comment line. Undo undeleted all the deleted lines.
Nice catch! Quad click is "Select entire paragraph" and it was only highlighting the first line as selected.
I've rectified that now, let's see if that fixes it for TChapman also.
Despite the minor issues you are already working on, just wanted to say "THANK YOU" for a fine
editor/compiler for Linux. I'm using it exclusively now for all my spin projects.
Thank you for all your hard work on this...
OBC
Cheers OBC, I really appreciate that.
I don't mind the bugs so much, it gives me a reason to procrastinate doing block intent markers [noparse];)[/noparse]
Mostly bug fixes. Nothing visibly major but a bit of re-architecting.
Think I have squashed all the bugs to date. Quadruple click to highlight paragraph does not seem to work on Mac (I don't see that as a problem as I'd never really intended it to work).
Still occasional serial problems on windows and the serial terminal still has the reset-on-disconnect on Mac.
I've tried to improve the error reporting in the compiler and I've managed to speed it up a bit again. I think I've pretty much wrung all the performance out of it I can now without a fairly major redesign.
Chuck Rice said...
As always, Every release is better and better! Thanks Brad!
Cheers Chuck! I'd be worried if they were getting worse, although I really have an uneasy feeling about this release for some reason.
I'm not sure it's been as heavily tested as I usually get time to do and there was some pretty major voodo changes internally in the highlighter.
Time will tell I guess.
As always, bugs and feedback warmly welcomed.
I think I've reached tipping point now. I was writing bst to help me write spin code, but I now seem to spend more time on bst than I do on the propeller itself. Hopefully things will start to settle down soon and I can get back to bashing out bytes [noparse]:)[/noparse]
I'm thinking about adding a native Linux-x86_64 compile based on Ubuntu and GTK2 (same as now just native 64 bit) to the builds. Would there be any interested parties?
Brad: I noticed that there are a couple of things (No show stoppers of course):
If you exit with the mouse pointer to the right of the end of the BST window it stays in "enter text mode".
The buttons in the config dialogs are... hesslich! I mean, on the Mac version they are not standard apple issued buttons.
In the list of filenames to the left, sometimes the names are painted in white with withe background when you use the scroll bar on the text editor, difficult to reproduce.
If there is just whitespace in the top line, it will be painted with different colors depending on the section you select (independent of the color it has).
Mac: If you close it from the close button on the top left of the window it closes the application... it shoudn't, well most apps just hide themselves, except the c1sco vpn client...
btw. It is just great, and gets better and better !
Edit: I just noticed that I used the same expression as Chuck, well it gets wtf, great!
Ale said...
Brad: I noticed that there are a couple of things (No show stoppers of course):
If you exit with the mouse pointer to the right of the end of the BST window it stays in "enter text mode".
Can I ask you to elaborate on this a little more? I'm not quite sure I understand the nature of the problem.
If I exit what?
Ale said...
The buttons in the config dialogs are... hesslich! I mean, on the Mac version they are not standard apple issued buttons.
Actually, believe it or not they are apple supplied buttons (Ugly, I agree completely - but Apple supplied none the less). Carbon has a bug in 10.5 where a button encapsulated in a panel component does not look like a nice normal OSX button. It looks ok in 10.4 though.
I could get around it by re-designing the UI to avoid the use of panels, but that just makes my life a lot harder for some nice looking buttons.
I figure the configure dialogs are not day to day kinda things, so for now they can stay ugly. My development machine is still 10.4, so they look ok here [noparse];)[/noparse]
I know Apple is deprecating carbon, so I'll have to do something about it eventually.
Ale said...
In the list of filenames to the left, sometimes the names are painted in white with withe background when you use the scroll bar on the text editor, difficult to reproduce.
I'll look into this one. Which particular version of OSX are you using (for bugs like this even the point release matters).
I've never seen anything like it here. Are you talking about the file tree up the top or the file list down the bottom?
If you ever manage to catch it, I'd positively _love_ a screenshot of it happening.
Ale said...
If there is just whitespace in the top line, it will be painted with different colors depending on the section you select (independent of the color it has).
Oh neat, I'd never noticed that one. Now I've seen it I know why it happens and it should be a pretty easy fix. (Famous last words - the GTK2 cut and paste bug took me 16 hours)
Ale said...
Mac: If you close it from the close button on the top left of the window it closes the application... it shoudn't, well most apps just hide themselves, except the c1sco vpn client...
This is the sort of feedback I really need from Mac users to make sure I can "Macify" things properly. Thanks! I'll look into how I might fix that.
Ale said...
btw. It is just great, and gets better and better !
Edit: I just noticed that I used the same expression as Chuck, well it gets wtf, great!
Thanks, I _really_ appreciate the feedback. It's only because you guys are so good with bug reports and feedback that it gets better. I miss or learn to ignore lots of this stuff in my day to day usage. I need to be slapped around the bonce with it from time to time to motivate me to find and fix it.
What started out as a quick and dirty ~3K line hack on top of a 10K line compiler is now in excess of 26K LOC. Bugs happen! [noparse];)[/noparse]
This solved all my known XP issues, other than a rare unexplainable crash that I can't pin down the action that prompts it. It seems to follow a condition where the window stops scrolling properly with the mouse wheel, where the screen sort of jumps a bit but doesn't scroll, although the slider will still scroll it. Could be a computer based issue, I usually solve it with quit/start on the bst. Not a big deal, except cases like last night I didn't save for a while. I love the response on relaunch, "it seems we are recovering from a crash", to which I want to say "yeah, true, and how bout putting my stuff back the way it was prior please" [noparse]:)[/noparse]
TChapman said...
This solved all my known XP issues, other than a rare unexplainable crash that I can't pin down the action that prompts it. It seems to follow a condition where the window stops scrolling properly with the mouse wheel, where the screen sort of jumps a bit but doesn't scroll, although the slider will still scroll it. Could be a computer based issue, I usually solve it with quit/start on the bst. Not a big deal, except cases like last night I didn't save for a while. I love the response on relaunch, "it seems we are recovering from a crash", to which I want to say "yeah, true, and how bout putting my stuff back the way it was prior please" [noparse]:)[/noparse]
Is this a new bug or an old bug? I ask as I've made some not insignificant changes to the internals of the highlighter in this release, particularly around memory allocation.
If it's a new bug that will narrow it down for me to try and fix.
On XP .12 was crashing, often it says a popup error that says "press continue if you want to risk......". If I press continue, things seem ok though.
I also got the same crash tonight too on .13, it usually is having to do with deleting stuff (or backspace maybe I forget what does it).
I noticed some things tonight new, clicking on the "+" fold down icon was folding blocks several blocks lower. It seems that deleting lines, or parts of lines of a block is weirding out the fold feature as described, but this is a guess. I did notice that once I hit Fold All Blocks it seems to rid the strange problem of blocks getting un/folded lower than the block I am clicking on.
The crash seems to follow certain conditions, I am getting closer to identifying those conditions, or at least trying to track the preceding motions that led to the crash.
Comments
When you say "PC". Linux/Windows - Version?
I'll make sure its fixed in the next release anyway.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
con QPOS = $08 'Query Position QSPD = $10 'Query Speed CHFA = $18 'Check for Arrival TRVL = $20 'Travel Number of Positions CLRP = $28 'Clear Position SREV = $30 'Set Orientation as Reversed STXD = $38 'Set TX Delay SMAX = $40 'Set Speed Maximum SSRR = $48 'Set Speed Ramp Rate '--- User Constants & Variables --- AllWheels = 0 RightWheel = 1 'ID of the right side Position Controller LeftWheel = 2 'ID of the left side Position Controller CommPin = 12 'communication bus pin BaudValue = 32 'for 19.2kbps rxPin = 1 txPin = 2 var long Wheel 'Specifies which wheel to command for subroutines long Distance 'Used to set the travel distance long RxData 'Used to receive data obj G : "cr_Globals" encoder : "FullDuplexSerial" PUB start : okay '' Starts uart object (at baud specified) in a cog '' -- returns false if no cog available okay := encoder.start(rxpin, txpin, %0000, 19200) : okayG : "cr_Globals" but it gives me an error at
(40,55) which is correct in my listing as the ":" between the ")" and the "okay"
If I don't comment out the G : " cr_Globals" then it just gives me an error about being unable to locate that object with no line locations (which is a bug in itself really)
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
I can snap a screen shot if that will help.
It would also be nice, if a double click on a word followed by a drag would select the second and third, etc. words in addition to the first word. Currently it selects the first word, but after that, it only ads characters to the selections.
A Triple Click on a line should select the entire line, including leading tab/spaces and the CR/LF. Currently a triple click only selects the words, so to duplicate 5 lines, I have to triple-click, then paste and fix, paste and fix, etc. I should only have to Paste, Paste, Paste, Paste, Paste.
Actually, yeah a screen shot would help heaps here.
As for the other niggles, I'll certainly look into them tomorrow for you and try to get them fixed before the next release.
Ta for the feedback [noparse]:)[/noparse]
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
Edit: The cursor does not show, but it is under RxData in the var section where the pink artifact was left.
Post Edited (Chuck Rice) : 11/16/2008 7:23:50 PM GMT
I've removed that feature and fixed the triple click line select that was bothering Chuck.
The first was there by design. I don't use double/triple clicking to select stuff so I figured it was a good way to do fold/unfold. I'm obviously not correct [noparse];)[/noparse]
I have to reproduce / fix the cursor being in the wrong place when highlighting an error yet, but I think I know where the problem might be.
The double-click select word not selecting by word when you drag is something I'm going to have to get into the bowels of the editor code to fix, so it's not likely to happen very soon I'm afraid.
There are enough bugs and annoyances there now to justify another release.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
Quick question. Are you clicking on the error in the error box to highlight the error? It does nothing to the editor or cursor until the error is clicked on.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
Oops. No. There is only 1 error, so I assumed that it would be selected.
No, it wasn't. I've set it up now so it will automatically select the first error though. Thought that sounded like a good idea.
I've got a clipboard bug with Ubuntu 8.10 to finish and I'll have a new release out. It's proving to be a bit of a problem though so I'm not sure when I'll get it buried.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
On XP using bst 12, I am experiencing that if I highlight several lines to delete, lines get deleted some number of rows below where I have highlighted. If I go to one line only, I can delete that line ok. This is with line number in on, and most code folded up except the few consecutive blocks that I am tweaking.
Another observation:
If I accidentally erase or leave out a ] or ) on a line, the compile error says 1,1 always, not the line that the error appears on. I have since learned how to work around these errors but thought I'd mention it in case you wanted all bugs listed.
I can't reproduce this one yet, but I'll keep trying.
Wow, can't believe I missed that one. It's an ugly artifact of the way I parse brackets. I'll make sure its taken care of. Thanks!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
I have seen something like it. If i quadruple-click on a line, it will select the whole line including leading whitespace and the line ending. if you then press delete, it deletes that line and the following paragraph. This was in the Mac version, and I was deleting a comment line. Undo undeleted all the deleted lines.
Despite the minor issues you are already working on, just wanted to say "THANK YOU" for a fine
editor/compiler for Linux. I'm using it exclusively now for all my spin projects.
Thank you for all your hard work on this...
OBC
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
New to the Propeller?
Getting started with a Propeller Protoboard?
Check out: Introduction to the Proboard & Propeller Cookbook 1.4
Updates to the Cookbook are now posted to: Propeller.warrantyvoid.us
Got an SD card connected? - PropDOS
Nice catch! Quad click is "Select entire paragraph" and it was only highlighting the first line as selected.
I've rectified that now, let's see if that fixes it for TChapman also.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
Cheers OBC, I really appreciate that.
I don't mind the bugs so much, it gives me a reason to procrastinate doing block intent markers [noparse];)[/noparse]
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
Mostly bug fixes. Nothing visibly major but a bit of re-architecting.
Think I have squashed all the bugs to date. Quadruple click to highlight paragraph does not seem to work on Mac (I don't see that as a problem as I'd never really intended it to work).
Still occasional serial problems on windows and the serial terminal still has the reset-on-disconnect on Mac.
I've tried to improve the error reporting in the compiler and I've managed to speed it up a bit again. I think I've pretty much wrung all the performance out of it I can now without a fairly major redesign.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
Cheers Chuck! I'd be worried if they were getting worse, although I really have an uneasy feeling about this release for some reason.
I'm not sure it's been as heavily tested as I usually get time to do and there was some pretty major voodo changes internally in the highlighter.
Time will tell I guess.
As always, bugs and feedback warmly welcomed.
I think I've reached tipping point now. I was writing bst to help me write spin code, but I now seem to spend more time on bst than I do on the propeller itself. Hopefully things will start to settle down soon and I can get back to bashing out bytes [noparse]:)[/noparse]
I'm thinking about adding a native Linux-x86_64 compile based on Ubuntu and GTK2 (same as now just native 64 bit) to the builds. Would there be any interested parties?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
If you exit with the mouse pointer to the right of the end of the BST window it stays in "enter text mode".
The buttons in the config dialogs are... hesslich! I mean, on the Mac version they are not standard apple issued buttons.
In the list of filenames to the left, sometimes the names are painted in white with withe background when you use the scroll bar on the text editor, difficult to reproduce.
If there is just whitespace in the top line, it will be painted with different colors depending on the section you select (independent of the color it has).
Mac: If you close it from the close button on the top left of the window it closes the application... it shoudn't, well most apps just hide themselves, except the c1sco vpn client...
btw. It is just great, and gets better and better !
Edit: I just noticed that I used the same expression as Chuck, well it gets wtf, great!
Can I ask you to elaborate on this a little more? I'm not quite sure I understand the nature of the problem.
If I exit what?
Actually, believe it or not they are apple supplied buttons (Ugly, I agree completely - but Apple supplied none the less). Carbon has a bug in 10.5 where a button encapsulated in a panel component does not look like a nice normal OSX button. It looks ok in 10.4 though.
I could get around it by re-designing the UI to avoid the use of panels, but that just makes my life a lot harder for some nice looking buttons.
I figure the configure dialogs are not day to day kinda things, so for now they can stay ugly. My development machine is still 10.4, so they look ok here [noparse];)[/noparse]
I know Apple is deprecating carbon, so I'll have to do something about it eventually.
I'll look into this one. Which particular version of OSX are you using (for bugs like this even the point release matters).
I've never seen anything like it here. Are you talking about the file tree up the top or the file list down the bottom?
If you ever manage to catch it, I'd positively _love_ a screenshot of it happening.
Oh neat, I'd never noticed that one. Now I've seen it I know why it happens and it should be a pretty easy fix. (Famous last words - the GTK2 cut and paste bug took me 16 hours)
This is the sort of feedback I really need from Mac users to make sure I can "Macify" things properly. Thanks! I'll look into how I might fix that.
Thanks, I _really_ appreciate the feedback. It's only because you guys are so good with bug reports and feedback that it gets better. I miss or learn to ignore lots of this stuff in my day to day usage. I need to be slapped around the bonce with it from time to time to motivate me to find and fix it.
What started out as a quick and dirty ~3K line hack on top of a 10K line compiler is now in excess of 26K LOC. Bugs happen! [noparse];)[/noparse]
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
The issue is the mouse pointer not changing from edit mode (That sort of upper case 'I') to the arrow.
Are you using Carbon ? well, they are not gonna fix it! :-(
I tried to reproduce that painting problem... so far no luck... as soon as I can I'll post it.
You got in a couple of months much more than you know who, even after we begged and promised to pay for it. Well, let's enjoy !, thanks.
Ale
Post Edited (Ale) : 11/18/2008 7:38:59 PM GMT
Love the new red parenthesis.
OBC
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
New to the Propeller?
Getting started with a Propeller Protoboard?
Check out: Introduction to the Proboard & Propeller Cookbook 1.4
Updates to the Cookbook are now posted to: Propeller.warrantyvoid.us
Got an SD card connected? - PropDOS
Is this a new bug or an old bug? I ask as I've made some not insignificant changes to the internals of the highlighter in this release, particularly around memory allocation.
If it's a new bug that will narrow it down for me to try and fix.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!
I also got the same crash tonight too on .13, it usually is having to do with deleting stuff (or backspace maybe I forget what does it).
I noticed some things tonight new, clicking on the "+" fold down icon was folding blocks several blocks lower. It seems that deleting lines, or parts of lines of a block is weirding out the fold feature as described, but this is a guess. I did notice that once I hit Fold All Blocks it seems to rid the strange problem of blocks getting un/folded lower than the block I am clicking on.
The crash seems to follow certain conditions, I am getting closer to identifying those conditions, or at least trying to track the preceding motions that led to the crash.