Re: [Bug-apl] chunk length specified across bytes

[Elias Mårtenson]

I'd presume that's 4 binary bytes in big-endian form. If so, then you have to do:

      256 256 256 256 ⊥ 0 0 32 0

8192

On 9 September 2015 at 12:01, <address@hidden> wrote:

The result of FIO∆fread FIO∆fopen_ro filename

will return a vector of values, where each number is a byte. Many different file format specifications specify a size within the file that may be a head size, a chunk size, etc. 

Is the correct way to interpret a number that spans many bytes (in this case 0 0 32 0 and 0 0 0 4) to use 

10 ⊥ 0 0 32 0 and 10 ⊥0 0 0 4

Hope that clears up my idea a little

-Alex

-------- Original Message --------
Subject: Re: [Bug-apl] chunk length specified across bytes
From: Elias_Mårtenson <address@hidden>
Date: Tue, September 08, 2015 8:44 pm
To: address@hidden
Cc: "address@hidden" <address@hidden>

I'm not entirely sure how this relates to GNU APL in any way?

On 9 September 2015 at 11:41, <address@hidden> wrote:

Hi Bug APL,

The PNG spec says that the length of a "chunk" (their term, not mine) is specified in the first four bytes of the file. When I read it in, the first four values are :

0 0 32 0

and the length is almost certainly not 32. 

A smaller chunk that was 4 had its first four bytes as:

0 0 0 4

What is the weighting given to these values (if that even is the right question to ask)?

-Alex