Useful NeXTSTEP Research Notes: Difference between revisions
Mineman3000 (talk | contribs) fixed formatting again |
Mineman3000 (talk | contribs) fixed formatting again again |
||
| Line 49: | Line 49: | ||
Probably - it's a 3GB filesystem, though, so well under the limit, for this version of NT | Probably - it's a 3GB filesystem, though, so well under the limit, for this version of NT | ||
</div> | </div> | ||
'''suzuran (Today at 7:50 PM)''' | |||
<div style="margin-left: 2em; line-height: calc(100% * 1.25);"> | <div style="margin-left: 2em; line-height: calc(100% * 1.25);"> | ||
Yeah, 2GB is the 2^31 line. | Yeah, 2GB is the 2^31 line. | ||
Revision as of 07:14, 29 December 2022
vmlemon (Today at 7:33 PM)
I never thought I'd see that materialised
Would be cooler, if I could actually breathe life into this stuff, and see it work, though
Feels kinda weird, seeing a GCC-based compiler, for Windows, engineered by NeXT, too
And, for whatever reason, they've nerfed the ability to get help for it, too
Interesting tuple, there
Feels really weird, seeing Windows-style paths, in a UNIX shell, and the tooling is pretty austere
CygWin's version of GDB? That's unexpected
That version of "df" seems kinda fucky, and cursed, too
Dunno why it reports negative values
Seems to ship its own, private copies of PostScript fonts, too
suzuran (Today at 7:47 PM)
Because you rolled over 2^31?
vmlemon (Today at 7:49 PM)
Probably - it's a 3GB filesystem, though, so well under the limit, for this version of NT
suzuran (Today at 7:50 PM)
Yeah, 2GB is the 2^31 line.
but knowing gnu stuff, they just used "int" and not "unsigned int"
vmlemon (Today at 7:50 PM)
That makes sense
The whole thing feels kinda kitbashed together, from disparate parts, and not everything seems to have source available for it
Managed to pull the "NeXT" directory out, from the Windows disk image, though, so I can have a better look at it
Seems to have more GNU stuff in it, than BSD stuff, somehow
And, the WindowServer seems to paint PostScript output onto GDI canvases/contexts
The Mach emulation stuff superficially seems less complex, than I expected, too - mostly just threading, and NT MailSlot communication
They don't seem to use that much actual Mach code, in the headers - there's only NeXT copyright headers, on that stuff, but they mention snippets being taken from CMU Mach