Progress goes over 100%

[ManuelBerrueta]

Using the latest release at this time (v3.6) for Linux. I was running the tool to enumerate virtual hosts for a lab and the tool is going way over the 100% mark:

[firefart]

Can you please try the dev branch and see if the problem exists there too?

go install github.com/OJ/gobuster/v3@dev

[ManuelBerrueta]

Hello, I tried to install it and got an issue with the library: /usr/libexec/gcc/x86_64-linux-gnu/13/cc1: error while loading shared libraries: /lib/x86_64-linux-gnu/libisl.so.23: invalid ELF header. I would usually explore more about the issue, but time constraints!

[NITHINSPACETIME]

Hello, I tried to install it and got an issue with the library: /usr/libexec/gcc/x86_64-linux-gnu/13/cc1: error while loading shared libraries: /lib/x86_64-linux-gnu/libisl.so.23: invalid ELF header. I would usually explore more about the issue, but time constraints!

this error occurs while the library file /lib/x86_64-linux-gnu/libisl.so.23 is corrupted or invalid i guess... To solve this error try checking if this file exists : ls -l /lib/x86_64-linux-gnu/libisl.so.23

if it does then , try insatlling : sudo apt update sudo apt install --reinstall libisl23

if anything happenes ask !!!

[Freeedim]

Hi there!

I am running into an issue that might be related to this, so I haven't filed a new issue; please tell me if I should.

The issue happens with v3.6 when I use --wordlist-offset and -x like in the command:
gobuster dir -u $url -w 1000-words.txt -r -x pdf,txt --wordlist-offset $X

If $X == 0, the progress shows the number of attempts against 3000 (I assume it comes from the 1000 words in the wordlist, times 3 because each word is tried with two different extensions and with no extension).

I struggled to find how to only try the second half of the wordlist:

• If I try $X = 1501, I have an error "offset is greater than the number of lines in the wordlist".
• If I try $X = 501, gobuster correctly tries the expected 1500 tests, but displays progress in a weird way:
  • the progress is shown against 3000 total attempts (fine);
  • the initial value is 500 (where, against 3000, I would have expected 1501);
  • the final value, when gobuster stops and exits, is 2000 / 3000 (66.67%), which is a bit odd, although consistent with the 1500 tests.

In the end, the work is done, but in my use case (many VMs with different offsets on very long wordlists), it is very misleading and difficult to know when a given VM starts to overlap the tests already done by another VM that had a greater offset.

My suggestion is to either:

• make shown progress only rely on the wordlist's number of words, and somehow indicate that each will result in N actual tests; or
• make shown progress fully rely on the actual number of tests including, in case there is a --wordlist-offset and some -x file extensions, a starting value being the actual number of tests that were skipped instead of the number of words that were skipped.

Additionally, I would find it great to be able to ignore the N last words of a wordlist, or maybe replace --wordlist-offset by something like --wordlist-slice <range>, where <range> might be "100-200", "100-" or "-200".

[firefart]

@Freeedim can you please try the dev branch? go install github.com/OJ/gobuster/v3@dev

There are a few updates to the progress handling in this version, curious if this problem still exists in the new code