[v2] chpasswd.cgi: Fixes bug12755 - v2 with password verification correction

  - Realised that I had not tested the old password beinhg correct or not. Previous check
   gave the same answer irrespective of the output coming from the htpasswd verification.
- This changes the variable used for the system_output result to an array and then
   checks if the first element contains the failure message that htpasswd gives if
   password verification fails.
- Tested out with correct and incorrect old passwords and gave the correct answer in
   both cases. Confirmed also that the check for the user being present works correctly
   for both an existing and new user name, which it did.

Fixes: bug12755
Tested-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
---
 html/cgi-bin/chpasswd.cgi | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
  

Hello Adolf,

Thanks for the patch. Is there no return code that we get from htpasswd instead of parsing the output?

-Michael

> On 7 May 2025, at 13:42, Adolf Belka <adolf.belka@ipfire.org> wrote:
> 
> - Realised that I had not tested the old password beinhg correct or not. Previous check
>   gave the same answer irrespective of the output coming from the htpasswd verification.
> - This changes the variable used for the system_output result to an array and then
>   checks if the first element contains the failure message that htpasswd gives if
>   password verification fails.
> - Tested out with correct and incorrect old passwords and gave the correct answer in
>   both cases. Confirmed also that the check for the user being present works correctly
>   for both an existing and new user name, which it did.
> 
> Fixes: bug12755
> Tested-by: Adolf Belka <adolf.belka@ipfire.org>
> Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
> ---
> html/cgi-bin/chpasswd.cgi | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/html/cgi-bin/chpasswd.cgi b/html/cgi-bin/chpasswd.cgi
> index c00caca20..46c3e02f6 100644
> --- a/html/cgi-bin/chpasswd.cgi
> +++ b/html/cgi-bin/chpasswd.cgi
> @@ -77,11 +77,11 @@ if ($cgiparams{'SUBMIT'} eq $tr{'advproxy chgwebpwd change password'})
>        # Check if a user with this name and password exists in the userdb file
>        # and if it does then change the password to the new one
>        my $user = &General::system_output("grep", "$cgiparams{'USERNAME'}", "$userdb");
> -       my $old_password = &General::system_output("/usr/bin/htpasswd", "-bv", "$userdb", "$cgiparams{'USERNAME'}", "$cgiparams{'OLD_PASSWORD'}");
> +       my @old_password = &General::system_output("/usr/bin/htpasswd", "-bv", "$userdb", "$cgiparams{'USERNAME'}", "$cgiparams{'OLD_PASSWORD'}");
>        if (!$user) {
>                $errormessage = $tr{'advproxy errmsg invalid user'};
>                goto ERROR;
> -       } elsif (!$old_password) {
> +       } elsif (@old_password[0] =~ /password verification failed/) {
>                 $errormessage = $tr{'advproxy errmsg password incorrect'};
>                 goto ERROR;
>        } else {
> -- 
> 2.49.0
> 
>
  

Hi Michael,

On 07/05/2025 14:44, Michael Tremer wrote:
> Hello Adolf,
> 
> Thanks for the patch. Is there no return code that we get from htpasswd instead of parsing the output?

It gives a return code for everything, with numbers of 0 to 7, except for the use of the -v option to verify the password.
This gives

password verification failed

if the existing password for the specified user is not correct and

Password for user fred correct.

if the user specified was fred and the specified password was correct.

It does the above for both the interactive -v and the batch mode using the command line of -bv

I had to use the check for if the string was found in the return variable because if I checked if the string matched the contents of the variable it always failed so I think there is a hidden Carriage Return or something in the output from htpasswd for the verification test.

Regards,

Adolf.

> 
> -Michael
> 
>> On 7 May 2025, at 13:42, Adolf Belka <adolf.belka@ipfire.org> wrote:
>>
>> - Realised that I had not tested the old password beinhg correct or not. Previous check
>>    gave the same answer irrespective of the output coming from the htpasswd verification.
>> - This changes the variable used for the system_output result to an array and then
>>    checks if the first element contains the failure message that htpasswd gives if
>>    password verification fails.
>> - Tested out with correct and incorrect old passwords and gave the correct answer in
>>    both cases. Confirmed also that the check for the user being present works correctly
>>    for both an existing and new user name, which it did.
>>
>> Fixes: bug12755
>> Tested-by: Adolf Belka <adolf.belka@ipfire.org>
>> Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
>> ---
>> html/cgi-bin/chpasswd.cgi | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/html/cgi-bin/chpasswd.cgi b/html/cgi-bin/chpasswd.cgi
>> index c00caca20..46c3e02f6 100644
>> --- a/html/cgi-bin/chpasswd.cgi
>> +++ b/html/cgi-bin/chpasswd.cgi
>> @@ -77,11 +77,11 @@ if ($cgiparams{'SUBMIT'} eq $tr{'advproxy chgwebpwd change password'})
>>         # Check if a user with this name and password exists in the userdb file
>>         # and if it does then change the password to the new one
>>         my $user = &General::system_output("grep", "$cgiparams{'USERNAME'}", "$userdb");
>> -       my $old_password = &General::system_output("/usr/bin/htpasswd", "-bv", "$userdb", "$cgiparams{'USERNAME'}", "$cgiparams{'OLD_PASSWORD'}");
>> +       my @old_password = &General::system_output("/usr/bin/htpasswd", "-bv", "$userdb", "$cgiparams{'USERNAME'}", "$cgiparams{'OLD_PASSWORD'}");
>>         if (!$user) {
>>                 $errormessage = $tr{'advproxy errmsg invalid user'};
>>                 goto ERROR;
>> -       } elsif (!$old_password) {
>> +       } elsif (@old_password[0] =~ /password verification failed/) {
>>                  $errormessage = $tr{'advproxy errmsg password incorrect'};
>>                  goto ERROR;
>>         } else {
>> -- 
>> 2.49.0
>>
>>
> 
>
  

Hi Michael,

On 07/05/2025 15:52, Adolf Belka wrote:
> Hi Michael,
> 
> On 07/05/2025 14:44, Michael Tremer wrote:
>> Hello Adolf,
>>
>> Thanks for the patch. Is there no return code that we get from htpasswd instead of parsing the output?
> 
> It gives a return code for everything, with numbers of 0 to 7, except for the use of the -v option to verify the password.
There might be a status code returned. The man page says

3 if the password was entered interactively and the verification entry didn't match

but elsewhere it does suggest that interactively is not via the -bv option but where you just use -v and manually type the password when requested on the command line.

If the status 3 is a valid status code, how can I access that from the output of the &General::system_output subroutine?

I could give it a try out and if it does work then I could do a v3 patch.

Regards,

Adolf.

> This gives
> 
> password verification failed
> 
> if the existing password for the specified user is not correct and
> 
> Password for user fred correct.
> 
> if the user specified was fred and the specified password was correct.
> 
> It does the above for both the interactive -v and the batch mode using the command line of -bv
> 
> I had to use the check for if the string was found in the return variable because if I checked if the string matched the contents of the variable it always failed so I think there is a hidden Carriage Return or something in the output from htpasswd for the verification test.
> 
> Regards,
> 
> Adolf.
> 
>>
>> -Michael
>>
>>> On 7 May 2025, at 13:42, Adolf Belka <adolf.belka@ipfire.org> wrote:
>>>
>>> - Realised that I had not tested the old password beinhg correct or not. Previous check
>>>    gave the same answer irrespective of the output coming from the htpasswd verification.
>>> - This changes the variable used for the system_output result to an array and then
>>>    checks if the first element contains the failure message that htpasswd gives if
>>>    password verification fails.
>>> - Tested out with correct and incorrect old passwords and gave the correct answer in
>>>    both cases. Confirmed also that the check for the user being present works correctly
>>>    for both an existing and new user name, which it did.
>>>
>>> Fixes: bug12755
>>> Tested-by: Adolf Belka <adolf.belka@ipfire.org>
>>> Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
>>> ---
>>> html/cgi-bin/chpasswd.cgi | 4 ++--
>>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/html/cgi-bin/chpasswd.cgi b/html/cgi-bin/chpasswd.cgi
>>> index c00caca20..46c3e02f6 100644
>>> --- a/html/cgi-bin/chpasswd.cgi
>>> +++ b/html/cgi-bin/chpasswd.cgi
>>> @@ -77,11 +77,11 @@ if ($cgiparams{'SUBMIT'} eq $tr{'advproxy chgwebpwd change password'})
>>>         # Check if a user with this name and password exists in the userdb file
>>>         # and if it does then change the password to the new one
>>>         my $user = &General::system_output("grep", "$cgiparams{'USERNAME'}", "$userdb");
>>> -       my $old_password = &General::system_output("/usr/bin/htpasswd", "-bv", "$userdb", "$cgiparams{'USERNAME'}", "$cgiparams{'OLD_PASSWORD'}");
>>> +       my @old_password = &General::system_output("/usr/bin/htpasswd", "-bv", "$userdb", "$cgiparams{'USERNAME'}", "$cgiparams{'OLD_PASSWORD'}");
>>>         if (!$user) {
>>>                 $errormessage = $tr{'advproxy errmsg invalid user'};
>>>                 goto ERROR;
>>> -       } elsif (!$old_password) {
>>> +       } elsif (@old_password[0] =~ /password verification failed/) {
>>>                  $errormessage = $tr{'advproxy errmsg password incorrect'};
>>>                  goto ERROR;
>>>         } else {
>>> -- 
>>> 2.49.0
>>>
>>>
>>
>>
>