mORMot Open Source

je_code

Member

Registered: 2025-04-17

Posts: 3

'mormot2tests.dpr' - 'Core crypto' - Regression Tests Failure

Following the Installation Guide for mORMot2 on Windows (Windows 10 64-bit), I have opened and compiled 'mormot2tests.dpr' with my Delphi 12.2 Athens IDE.

When I ran the newly created executable, all of the regression tests except 36 of them run successfully. The 36 tests that failed belong to the 'Core crypto' section. The following statements were found in the 'log' file:

000000000000000B  ! rotat mormot2tests 2.3.10743 TSynLog 2.3.10743 {2 9.60 15.55 3.3GB/8GB 4a651701}
000000000000004B  ! EXC   ERsaException {Message:"TRsa.GenerateNew failed as rgrTimeout"} [Main] at 50c9c3   {2 9.60 15.55 3.3GB/8GB 4a651701}
0000000000000C90  ! EXC   ECryptCert {Message:"TCryptCertX509.GeneratePrivateKey(xkaRsa) failed"} [Main] at 50c9c3   {2 9.60 15.55 3.3GB/8GB 4a651701}

How do I resolve this issue?

ab

Administrator

From: France

Registered: 2010-06-21

Posts: 14,966

Website

Re: 'mormot2tests.dpr' - 'Core crypto' - Regression Tests Failure

The timeout is about the fact that the RSA generation takes too much time.
This generation takes 100% of a core.

What is the CPU? You may have a very slow/overloaded system.

Try to run only

mormot2tests /test crypto._rsa

and report if it works (with the numbers).

For reference, here are my results on a Windows 7 VM, compiled with Delphji 12.2 Athens, on both Win32 and Win64:
https://gist.github.com/synopse/837dcf1 … 4ce40ee4ac

je_code

Member

Registered: 2025-04-17

Posts: 3

Re: 'mormot2tests.dpr' - 'Core crypto' - Regression Tests Failure

Thanks for the information.

I forgot to mention that I was running the tests on a VirtualBox Virtual Machine with the following processor: Intel(R) Xeon(R) CPU E3-1226 v3 @ 3.30GHz, 3292 Mhz, 2 Core(s), 2 Logical Processor(s).

Per your suggestion, I run the limited set of tests. They completed without any issues. Here are my results:

   mORMot2 Regression Tests
  --------------------------


1. Core units

 1.4. Core crypto:
  - RSA: 1,307,460 assertions passed  12.49s
     10 RS256 sign in 1.12s i.e. 8/s, aver. 112.08ms
     100 RS256 verify in 203.47ms i.e. 491/s, aver. 2.03ms
     10 PS256 sign in 1.12s i.e. 8/s, aver. 112.04ms
     100 PS256 verify in 206.23ms i.e. 484/s, aver. 2.06ms
  Total failed: 0 / 1,307,460 - Core crypto PASSED  12.49s


2. ORM


3. SOA


Software version tested: 2.3.10743 (2025-04-17 15:02:39)

Windows 10 64bit 22H2 (10.0.19045.5737) [ms1252 4GB 4A651701]
    2 x Intel(R) Xeon(R) CPU E3-1226 v3 @ 3.30GHz [8MB] (x86)
    on VBOX - 1 Oracle VirtualBox Version 7.1.8 VGA BIOS
Using mORMot 2.3.10743 17 Apr 2025
    TSqlite3LibraryStatic 3.46.1 with internal MM
Generated with: Delphi 12.2 Athens 32 bit Win compiler

Time elapsed for all tests: 12.49s
Performed 18 Apr 2025, 09:37:21 by user1 on DESKTOP-82SJUE9

Total assertions failed for all test suits:  0 / 1,307,460
! All tests passed successfully.

ab

Administrator

From: France

Registered: 2010-06-21

Posts: 14,966

Website

Re: 'mormot2tests.dpr' - 'Core crypto' - Regression Tests Failure

Weird.

My Core i5 seems much faster than this CPU.
The whole tests took 12 second on your VM, while in mine it took 600 ms...

je_code

Member

Registered: 2025-04-17

Posts: 3

Re: 'mormot2tests.dpr' - 'Core crypto' - Regression Tests Failure

The speed difference might be partially explained by the age of my CPU. It was released in 2014.

I ran the same partial test on my 'host' PC. My results are still slower than yours but much better than my VM ones:

   mORMot2 Regression Tests
  --------------------------


1. Core units

 1.4. Core crypto:
  - RSA: 1,307,458 assertions passed  1.85s
     10 RS256 sign in 481.14ms i.e. 20/s, aver. 48.11ms
     100 RS256 verify in 83.16ms i.e. 1.1K/s, aver. 831.66us
     10 PS256 sign in 449.52ms i.e. 22/s, aver. 44.95ms
     100 PS256 verify in 83.84ms i.e. 1.1K/s, aver. 838.46us
  Total failed: 0 / 1,307,458 - Core crypto PASSED  1.85s


2. ORM


3. SOA


Software version tested: 2.3.10743 (2025-04-17 15:02:39)

Windows 10 64bit 22H2 (10.0.19045.5737) [ms1252 3.9GB 4A651701]
    4 x Intel(R) Xeon(R) CPU E3-1226 v3 @ 3.30GHz [8MB] (x86)
    on Dell Inc. Precision T1700 00
Using mORMot 2.3.10743 17 Apr 2025 and OpenSSL 1.1.1b  26 Feb 2019
    TSqlite3LibraryStatic 3.46.1 with internal MM
Generated with: Delphi 12.2 Athens 32 bit Win compiler

Time elapsed for all tests: 1.85s
Performed 18 Apr 2025, 10:54:26 by user1 on PC109NEW

Total assertions failed for all test suits:  0 / 1,307,458
! All tests passed successfully.

Since I was able to successfully run the partial tests on my VM, should I ignore the failed tests that occurred during while running the full set of tests?

ab

Administrator

From: France

Registered: 2010-06-21

Posts: 14,966

Website

Re: 'mormot2tests.dpr' - 'Core crypto' - Regression Tests Failure

Yes, I just suspect you did not have any luck at that time.

Creating a RSA keypair is not deterministic at all (it searches for primes by incrementing a random value), so it could be very quick or very slow, from time to time.