Re: Fix for 687151, gswin32 starts with 7s delay..., improved log message

Ray,

> From: "Ray Johnston" <[email protected]>
> To: "Igor V. Melichev" <[email protected]>
> Cc: "Alex Cherepanov" <[email protected]>; "Miles Jones"
<[email protected]>; "GS Code Review List"
<[email protected]>
> Sent: Thursday, December 11, 2003 2:38 AM
> Subject: Re: [gs-code-review] Fix for 687151, gswin32 starts with 7s
delay..., improved log message


> I think your code for finding the 'head_size' searching back from
> the end is not correct. You check pattern[i] without updating 'i'.
>
> I think you meant:
>
>        for (i = j - 1; i > 0; i--)
>           if(pattern[i] == '/' || pattern[i] == '\\' || pattern[i] == ':')
>               break;
>       hsize = i + 1;
>
> But with no directory serparator, this will terminate with i==0,
> setting hsize to 1 instead of 0. You probably also meant to have
> the loop run as long as 'i >= 0'.

Right you are. I'm sorry for inaccuracy.

> The code I had initialized hsize=0, so with no directory separator,
> hsize would be left at 0.

OK. I'm sorry for missing it while the first reading.

> Also an error might be reasonable if there is
> a ? or * in the directory portion (head) since that is not implemented.

Agree.

> I am not sure if 'invalidaccess' is the best error to throw for either
> case. Possibly 'ioerror' for the \? an \* case and 'rangecheck' or
> 'limitcheck' for the second case (although 'limitcheck' is not listed
> in the PLRM as a possible error, it tends to be used for implementation
> limitations).

I don't like 'ioerror' here, because it assumes a bad disk content.
'rangecheck' should be fine.

Igor.