bug-gnulib
[Top][All Lists]
Advanced

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

Re: abort() in regex.c


From: Vladimir 'φ-coder/phcoder' Serbinenko
Subject: Re: abort() in regex.c
Date: Mon, 26 Mar 2012 12:07:21 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:8.0) Gecko/20120216 Icedove/8.0

On 26.03.2012 11:21, Bruno Haible wrote:
> Hi Vladimir,
>
>> In GRUB everything runs in the same environment and so
>> abort() can only be mapped to complete bootloader termination. It's
>> unreasonable to do such an action in case of regexp error.
> Have you seen this abort() happen?
No, but I've seen its use resulting in link error
> abort() marks the end of an execution branch in case of
>   - hardware error,
>   - a program bug that is easy to determine.
>
> abort() is not used in conditions that can be expected to happen in
> production environments.
> When you are in a situation where abort() is called, trying to continue
> program execution (via "unknown regexp error", for example) only makes the
> determination of the program bug harder.
Usually I'd agree. However in GRUB case there is no dump created (we
have no FS writing) so debug value of abort() is the same as normal printf.
On the other hand abort()ing will most likely result in an unbootable
system instead of a glitch in glob expansion (primary use of regex in
GRUB). Having the whole system unbootable would make the debugging harder.
> Bruno
>
>


-- 
Regards
Vladimir 'φ-coder/phcoder' Serbinenko


Attachment: signature.asc
Description: OpenPGP digital signature


reply via email to

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