Android application to manipulate sd image from the site

aaaa

aaaaa

aaa

I missed these latest posts because of the new page, yes the answer was to emphaticly set the terminating 0.
To save you turning back I’ll reprint.

		strcpy_s       (dst_filename,256,src_filename);
		dst_filename[(strlen(src_filename)-3)] ='\0';

strcpy_s

Not strncpy_s ?

aaaa

aaaa

line 140: };
The end of typedef struct adfs_dir_t

Regarding the strncpy, I will have to make a simple test case and forward it as a bug, It could be giving problems. I have had problems with argument handling on one occasion, could be a reason. Any bugfix is obviously a good thing.

aaaa

aaaa

aaaa

The strncpy() function is similar, except that at most n bytes of src are copied. Warning: If there is no null byte among the first n bytes of src, the string placed in dest will not be null-terminated.

Ah, I see now, it is not a bug, I guess the purpose of strncpy is to at least allow you to restrict the amount from a large buffer to fit in a smaller buffer.
But the terminating 0 does have to be set later.

aaaa

The string concept is there in BASIC, I had given up on that in C, You could keep improving/making your own functions, but they will get bigger. A lot of source code has wrapper functions around the basic functions, nearly always have their own error functions.

I see your point though, I will look at the memcpy function more in the future.

aaaaa

aaaa

Like this you can know the length of the string without having to parse it for an hypothetical null terminating character that can be removed by smart functions like strncpy.

Yeh, I like the idea of keeping things simple with one function like memcpy though.
‘The memcpy() function copies n bytes from memory area src to memory
area dest.’
It would do this job wouldn’t it?
I see the wmemcpy() will use wide characters as a count, sort of expansion so that you can use the count of visual characters I suppose.
How much overhead would there be in initialising a buffer to all 0’s?
That would ensure a terminator after placing the string.
I’m a bit rusty but if you use pointer to buffer, doesn’t that maintain the position in the buffer?
*buf++1 = ‘\0’
sort of thing.
There is probably overheads associated with pointers as well.
I’m thinking that a block or buffer copy of 0’s to a newly created buffer might be less work than calculating the length of a string with strlen. strncpy_s will incorporate some sort of strlen calculation so it wont be any better.

aaaa

aaaa

aaaa

aaa

So they should give a valid string on output.

Exactly.

There is a lot of stuff in the standard C library definition (one might even say C itself) that appears to take the path of least resistance effort.

How much overhead would there be in initialising a buffer to all 0’s?

Use calloc() instead of malloc()?

I updated it again, now it can browse directories and output squashed file from the sqar file.

I compiled this on linux tonight, a few new warnings, and squached a file there and back no problems. However I didn’t get the sqar back to anything using squach, should I have got a tar back? I see, just a directory listing- dont recognise [home] yet.
Looking at the output from the compression stage gives the answer. It is using the full hierachy of the drive, rather than the selected directories. So It will have to open /home/ron/Software stages first.
This happened because I used /home/ron/Software/kalc when I made the tar. – My fault.

ron@Asus ~/Software/squach $ ./squach kalc.sqar
$
	[home] 2048 bytes 

ron@Asus ~/Software $ ./squach/squach kalc.sqar kalc-2/2/1
kalc-2/2/1
	README 1 Kb at offset 0 bytes - load ffffffca exe ffffffff  
	kalc/hlp 18 Kb at offset 1536 bytes - load ffffffca exe ffffffff  
	install-sh 3 Kb at offset 20480 bytes - load ffffffca exe ffffffff  
	kalc 195 Kb at offset 24064 bytes - load ffffffca exe ffffffff  
	Makefile/in 0 Kb at offset 224256 bytes - load ffffffca exe ffffffff  

The (linux) gcc warnings are

ron@Asus ~/Software/squach $ gcc -c squach.c -static 
squach.c: In function ‘squash_align’:
squach.c:784:2: warning: format ‘%ld’ expects argument of type ‘long int’, but argument 3 has type ‘size_t’ [-Wformat=]
  printf("lzw clear ofset %d %ld %d\n",ofs_line,ofs_dif,offsets[(ofs_line<<3)+ofs_dif]);
  ^
squach.c: In function ‘compress_tar’:
squach.c:1700:2: warning: too many arguments for format [-Wformat-extra-args]
  printf("%d directories starting at %d\n",n_dirs,sqar_write,dir_rec_start);
  ^

aaaa