what is Link indicator (file type) in tar parser

Question

I want to know the file-type of html file present inside a tar file. I have stored theis tar file in a buffer and i know that i can know the size of the file at location buffer[124] but i want to know is - (1.) I want to know that if the file present in tar is HTML file ?? Any idea how to do that ?? I think that i can know the file form Link Indicator(but i am not sure). could any one explain please how to do that ??

(2.) Once if i am sure that i have html file inside the tar file then i want to staotr the contents of these html file in the tar file.

There are many other files also present in tar file (not only html file-Please attention on this line, so i don't know the loctaion of the html file)

Any idea how to achieve this ??

Or use a parser if you don't trust extensions. Just [don't use regex](http://stackoverflow.com/a/1732454/1056003) ;) — nikolas, Jul 29 '13 at 07:18
@ Joachim sorry i couldn't understand ?? you mean i have to check in link indicator that if the file is of type .html ? could you please tell me what link indicator actually is meant for ?? as i think we can get file extension (.html in my case) from link indicator.Am i right?? Or where you are asking me to check for .html file extension ?? — Sss, Jul 29 '13 at 07:19
I'm not sure what you mean by "link indicator", as files in a tar file doesn't have such a field. I mean that you simply check the filename or each file in the archive, and if it ends with `".html"` (or `".htm"` or other combinations) then you could assume it's a HTML file. — Some programmer dude, Jul 29 '13 at 07:34
by "link indicator" i mean location buffer[156] (field offset in tar parser) Can we get the file type from here ??and i can't archive a file manually i have to write c++ code for getting html file contents inside tar file. and for that i am storing the tar file in a buffer by creating a stream (but that's immaterial here the purpose is to know if any way exist to get html file contents in a tar file by c++ programming , I even can't use strstr function to find html file because it stop working after the first NULL value it encounters) the purpose is to know how to achieve it?? — Sss, Jul 29 '13 at 07:43

score 0 · Answer 1 · answered Jul 29 '13 at 07:52

0

Ah, you mean the typeflag field. No, it's not for that kind of filetype, it's to tell if the file is a regular file, directory, hard link, soft link, device special file, etc.

Your system (if it's POSIX compliant) should have a <tar.h> system header file (usually in /usr/include) that contains these flags. Or you can see the official POSIX specification.

answered Jul 29 '13 at 07:52

Some programmer dude

400,186
35
402
621

so what i think was wrong. Do you know any way to acieve my target ? How to get aware about html file(there are so many other files also present) by c++ code(i can access the tar file using buffer indexes ) ?? – Sss Jul 29 '13 at 07:56
@ShekharSinghSHEKHAWAT The only way is to check the file names, and look for file-name endings typical for HTML files. – Some programmer dude Jul 29 '13 at 08:07
you mean i should check the data from buffer[0 to 100] which contains the file name and if i have .html here then i should store the contents of this file. but i will come across so many other files by doing this. How to skip these extra files (even they have different sizes) in order to get my .html files contents ?? – Sss Jul 29 '13 at 08:21
1

@ShekharSinghSHEKHAWAT Yes that's the only way to do it. If you want a specific file with a specific name, you have to look through the whole archive. It's not really hard, as you only have to read the header of each file, and only need to read the data blocks if it's a file you want to extract. – Some programmer dude Jul 29 '13 at 08:24
Do you think that my thsi algo to skip the file is correct int skip = filesize % 512 ? filesize + 512 - (filesize % 512) : filesize . I will skip this much index(buffer[skip]) until i don't get .html file.(the algo is to skip (512 bytes for data block +size of file which must be the next multiple of 512) Am i werite?? please correct me if i am wrong. – Sss Jul 29 '13 at 08:35

score 0 · Answer 2 · answered Jul 29 '13 at 16:17

@ Joachim thanks for your suggestion finally i have done it the code is as below you can see if you want -

char* StartPosition;

    size_t skip= 0;
    char HtmlFileContents [200000];
    char contents [8000];
    do
    { 
            int SizeOfFile = CreateOctalToInteger(&buffer[skip+124],11);
            size_t distance= ((SizeOfFile%512) ? SizeOfFile + 512 - (SizeOfFile%512) : SizeOfFile );
            skip += distance + 512;
            memcpy(contents,&buffer[skip],100);
            if (StartPosition=strstr(contents,".html"))
            {
                MessageBox(m_hwndPreview,L"finally string is copied",L"BTN WND6",MB_ICONINFORMATION);
                int SizeOfFile = CreateOctalToInteger(&buffer[skip+124],11);
                memcpy(HtmlFileContents,&buffer[skip+512],SizeOfFile);
                break;
            }


    }
    while(strcmp(contents,".html") != NULL);

and it can work for every file stored in tar file . we just need to put the file extension (.html in my case) and this code will give us its contents.

what is Link indicator (file type) in tar parser

2 Answers2