Closed mouse07410 closed 8 years ago
LOL, CI was not building against OpenSSL due to a typo so this was missed. I think most of these are the same bug, looking at it now.
What format is the date/time above?
Just the raw ASN.1 time string. The output of X509_Certificate::to_string could be better.
The BER infinite loop is because of a strange interaction with DataSource_Stream - the peek
function which is used to guess DER vs PEM resets the EOF state, so the cli never realizes EOF already happened. Not really sure how to fix this one yet, thanks for mentioning the problem.
... I think most of these are the same bug, looking at it now ...
Thanks! I confirm that 55b8fb5 with 27a6d11 fixes the tests:
. . . . .
Tests complete ran 528317 tests all tests ok
What format is the date/time above?
Just the raw ASN.1 time string. The output of X509_Certificate::to_string could be better.
Could you change it then...? When a user invokes botan
CLI to print out the cert, he probably would like to see what the expiration data/time is, and does not have the immediate skills to convert the raw ASN.1 to DTG...
...the peek function which is used to guess DER vs PEM resets the EOF state, so the cli never realizes EOF already happened...
But... with the (new?) --ber
flag is there still a need to guess DER vs. PEM? A simpler logic would be "--ber
=> DER, otherwise PEM", and it would avoid the need to peek
...?
Could you change it then...?
LOL, you must be joking. But if not:
I have only so many (completely unpaid) hours to devote to this project and the plain truth is when selecting a task to work on I'm quite unlikely to spend time on an issue like this [*], which is purely cosmetic and can be improved in a somewhat drive by manner by any skilled and motivated programmer. In contrast there are many things, such as reviewing complicated patches or redesigning major parts of the library, which can only be done by me - either I do them or they will not happen.
[*] Unless it starts bugging me personally, of course, in which case I'll fix it that day. ;)
The output of X509_Certificate::to_string could be much better (more or less matching openssl x509 -text
would be fine), but until/unless it starts bugging me personally it remains just a todo. (But, it was missing in the todo, so I added it).
But... with the (new?) --ber flag is there still a need to guess DER vs. PEM?
Surely so if this flag were passed to the library in any way, but it just exists in the cli to control text vs binary IO on the file.
Closing, as tests are fixed now. The DataSource_Stream EOF issue should have its own ticket.
Mac OS X 10.11.6 El Capitan. Xcode-8.0. Installed OpenSSL-1.0.2j (via Macports). Current Botan master, configured:
Some compilation warnings:
Several tests fail:
Also,
cert_info
seems to misbehave, especially when given--ber
flag:What format is the date/time above?
And with
--ber
it gets into infinite loop: