[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Crash recovery strategies

From: Daniel Colascione
Subject: Re: Crash recovery strategies
Date: Mon, 4 Jan 2016 07:44:00 -0800
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.4.0

On 01/04/2016 07:40 AM, Eli Zaretskii wrote:
>> From: Daniel Colascione <address@hidden>
>> Date: Sun, 3 Jan 2016 15:51:13 -0800
>> How would moving to C help? Keep in mind that we're talking about fork
>> *and exec*, so the parent and child don't share memory. They can,
>> however, communicate over a pipe.
> Communicating over a pipe means encoding/decoding stuff, which
> allocates memory, temporary buffers, and can potentially run Lisp.
> It's IMO unwise to do that in a crashed session.

Did you read my proposal? I suggested reading a *single byte*. POSIX
guarantees we can do that, even from async-signal-unsafe context. The
byte tells us whether to save, crash, die, or attempt to recover.

man 7 signal

But never mind. This whole ask-the-user-what-to-do thing is too
confusing. Just imagine I'd said we'd always write the save file.

Attachment: signature.asc
Description: OpenPGP digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]