Fix runtime error inside m3quake QScanner#1024
Fix runtime error inside m3quake QScanner#1024VictorMiasnikov wants to merge 1 commit intomodula3:masterfrom
Conversation
Fix runtime error inside m3quake QScanner
...
***
*** runtime error:
*** An enumeration or subrange value was out of range.
*** file "../src/QScanner.m3", line 49
***
*** execution of [<function _BuildGlobalFunction at 0x0000000002F2DD68>, <function _ShipFunction at 0x0000000002F2DDD8>] failed ***
...
...
t.buflen := M3File.Read (f, t.buffer^, size); (* VVM: it is line 49 *)
...
|
Are you sure you have everything up to date? |
|
} I don't think this is the right fix.
It do not "bad things". But I am not sure that it ( path) do "good things"
… 19.04.2022, 13:51, "Jay Krell" ***@***.******@***.*** commented on this pull request.In m3-sys/m3quake/src/QScanner.m3:> @@ -46,7 +46,7 @@ PROCEDURE Init (t: T; f: File.T; map: Quake.IDMap): T =
DO
size := VAL(stat.size, INTEGER);
t.buffer := NEW (REF ARRAY OF CHAR, MAX (0, size) + 1);
- t.buflen := M3File.Read (f, t.buffer^, size);
What is size when this fails? Is it negative? I don't think this is the right fix.
|
|
19.04.2022, 13:49, "Jay Krell" ***@***.***>:
Are you sure you have everything up to date?It seems like you are hitting things I already fixed, like the slashes (long ago, not today).In this case, the problem I hit months ago when first running DJGPP is that stat.size does not account for text conversion (\r\n to \n) and reading does, so the sizes would mismatch. Open should always be binary. I thought I made it so. Maybe I missed some.I don't think this is the right fix.The right fix is to always open binary.I agree it is wonky.
} I don't think this is the right fix.
Yes. In best case it is "temporary fix".
} fixed, like the slashes (long ago, not today).
I remember this fix. ( May be 'this', I am not sure)
} DJGPP
This patch also for other MinGw targets
} text conversion (\r\n to \n)
It is news for me. In this case "sizes would mismatch" of course . . .
|
|
@VictorMiasnikov this can be closed now right? |
This patch ( in any key) is part of: Nothing terrible happened with: But this patch is not be "a way to real solving of problem" ;-( |
|
It is actual |
|
How is the error produced? |
|
We need to stop with manual releases at this point, and stick to automation. |
|
How is the error produced?
I don't remember. ( See 2022 year)
If need I try reproduce again
In any key "my branch" is total good on AMD64_LINUX with GCC backend
And 90/100 tests passed in anothers backend
…
----------------
Кому: modula3/cm3 ***@***.***);
Копия: Mention ***@***.***);
Тема: [modula3/cm3] Fix runtime error inside m3quake QScanner (PR #1024);
07.05.2026, 12:45, "Jay Krell" ***@***.***>:
jaykrell left a comment (modula3/cm3#1024)
How is the error produced?
|
|
We need to stop with manual releases at this point, and stick to automation.
There's been too many supply chain attacks in the larger world. Sorry.
I don't understand that You want say...
(
In any key "my branch" is total good on AMD64_LINUX with GCC backend
And 90/100 tests passed in anothers backend
)
…---------------
Кому: modula3/cm3 ***@***.***);
Копия: Mention ***@***.***);
Тема: [modula3/cm3] Fix runtime error inside m3quake QScanner (PR #1024);
07.05.2026, 12:46, "Jay Krell" ***@***.***>:
jaykrell left a comment (modula3/cm3#1024)
We need to stop with manual releases at this point, and stick to automation.
There's been too many supply chain attacks in the larger world. Sorry.
|
|
} to stop with manual releases at this point, and stick to automation
Today / now it looks as fantastic:
Boot1.py AMD64_LINUX c keep
^^^^^^^ broken in cm3-d5.12.0It generate .tar.gz with outdated makefile
} to stop with manual releases at this point, and stick to automation
…----------------
Кому: modula3/cm3 ***@***.***);
Копия: Mention ***@***.***);
Тема: [modula3/cm3] Fix runtime error inside m3quake QScanner (PR #1024);
07.05.2026, 12:46, "Jay Krell" ***@***.***>:
jaykrell left a comment (modula3/cm3#1024)
We need to stop with manual releases at this point, and stick to automation.
There's been too many supply chain attacks in the larger world. Sorry.
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: ***@***.***>
|
|
AI / LLM understand You right? ( What exactly named files I upload to PR 1024 ? Translation of the Quote:
Explanation of the MeaningThis message was written by a developer or project manager explaining a change in the software release process. Here are the key points: 1. Stopping Manual Work ("manual releases") 2. Why Automation is Needed ("stick to automation") 3. Supply Chain Attacks ("supply chain attacks")
4. "Sorry" Short Summary: |
2 participants
Fix runtime error inside m3quake QScanner
...
*** runtime error:
*** An enumeration or subrange value was out of range.
*** file "../src/QScanner.m3", line 49
*** execution of [<function _BuildGlobalFunction at 0x0000000002F2DD68>, <function _ShipFunction at 0x0000000002F2DDD8>] failed ***
...
...
t.buflen := M3File.Read (f, t.buffer^, size); (* VVM: it is line 49 *)
...