Zsh Mailing List Archive
Messages sorted by:
Reverse Date,
Date,
Thread,
Author
Re: crash in completing code with unicode support
- X-seq: zsh-workers 21644
- From: Wayne Davison <wayned@xxxxxxxxxxxxxxxxxxxxx>
- To: Andrey Borzenkov <arvidjaar@xxxxxxxxxx>
- Subject: Re: crash in completing code with unicode support
- Date: Tue, 16 Aug 2005 14:37:13 -0700
- Cc: zsh-workers@xxxxxxxxxx
- In-reply-to: <200508162327.58597.arvidjaar@xxxxxxxxxx>
- Mailing-list: contact zsh-workers-help@xxxxxxxxxx; run by ezmlm
- References: <237967ef05081518376385563e@xxxxxxxxxxxxxx> <200508161157.59407.arvidjaar@xxxxxxxxxx> <237967ef050816050130fefbf8@xxxxxxxxxxxxxx> <200508162327.58597.arvidjaar@xxxxxxxxxx>
On Tue, Aug 16, 2005 at 11:27:49PM +0400, Andrey Borzenkov wrote:
> p = compprefix + 1;
> if (*p == Meta)
> p += 2;
I haven't tried to fully grok that code yet, but it looks to me like the
non-meta case results in just the character at compprefix being sent to
pattry(), so shouldn't the code really be checking if the compprefix
character is metified or not?
p = compprefix + 1 + (*compprefix == Meta);
Also, the following check (from slightly lower down) needs to be changed
to handle the potential for the pointer to be bumped further forward:
if (p == compprefix + l)
If the above is right, then this line should probably be changed to be:
if (p == compprefix + l + (*compprefix == Meta))
..wayne..
Messages sorted by:
Reverse Date,
Date,
Thread,
Author