Git 1.2

[iain.m...@gmail.com]

I've finally updated Git to understand the latest Glulx spec (3.1),
for compatibility with gamefiles created with the latest version of
Inform.

I haven't yet implemented the "malloc" and "mfree" opcodes -- instead,
I've just indicated via the "gestalt" opcode that they're not
available. Apologies if this causes any problems (does Inform need
them?). As soon as I get around to implementing these opcodes, I'll
release another Git update.

For now, get version 1.2 on the archive, or at http://diden.net/if/git

Iain

[Andrew Plotkin]

Here, iain.m...@gmail.com wrote:
> I've finally updated Git to understand the latest Glulx spec (3.1),
> for compatibility with gamefiles created with the latest version of
> Inform.
>
> I haven't yet implemented the "malloc" and "mfree" opcodes -- instead,
> I've just indicated via the "gestalt" opcode that they're not
> available. Apologies if this causes any problems (does Inform need
> them?).

Inform uses malloc when the game includes dynamic features, such as
lists or indexed-text values. There is a certain amount of
pre-allocated memory (the "Use dynamic memory allocation of..."
option), and beyond that, it's malloc time. See 19.2.

--Z

--
"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the borogoves..."
*
If the Bush administration hasn't subjected you to searches without a warrant,
it's for one reason: they don't feel like it. Not because you're an American.

[aaroni...@gmail.com]

This still fails with my WIP, presumably because it uses indexed text
and lists. Great to see a Git update, though! The extra speed is
rapidly becoming a lifesaver as IF becomes more complex under the
hood.

--Aaron

[Simon Baldwin]

FYI, in builds with non-Unicode Glks, terp.c carefully converts L1/2
to a safe-to-print value or '?', but then discards this and prints
L1/2 anyway. Lines 775 and 1030.

Also, terp.c doesn't check glk_gestalt(gestalt_Unicode,0) before
calling any Glk unicode functions.

[iain.m...@gmail.com]

On Jan 6, 10:48 pm, aaronius.r...@gmail.com wrote:
> On Jan 6, 12:12 pm, iain.merr...@gmail.com wrote:
>
> > I haven't yet implemented the "malloc" and "mfree" opcodes -- instead,
> > I've just indicated via the "gestalt" opcode that they're not
> > available. Apologies if this causes any problems (does Inform need
> > them?). As soon as I get around to implementing these opcodes, I'll
> > release another Git update.
>

> This still fails with my WIP, presumably because it uses indexed text
> and lists. Great to see a Git update, though! The extra speed is
> rapidly becoming a lifesaver as IF becomes more complex under the
> hood.

Oh well, watch this space for 1.2.1!

Iain

[iain.m...@gmail.com]

Thanks for the tip -- I'll fix both of these.

Iain