[bug] Obscure error in CMakeDeps on malformed requirement string passed to self.cpp_info.requires

[transegn]

Describe the bug

Conan produces a cryptic "too many values to unpack (expected 2)" error

======== Testing the package ========
Removing previously existing 'test_package' build folder: /__w/1/s/packages/sim_blocks/test_package/build/gcc-11-x86_64-release
foo/1.2.3 (test package): Test package build: build/gcc-11-x86_64-release
foo/1.2.3 (test package): Test package build folder: /__w/1/s/packages/sim_blocks/test_package/build/gcc-11-x86_64-release
foo/1.2.3 (test package): Writing generators to /__w/1/s/packages/sim_blocks/test_package/build/gcc-11-x86_64-release/generators
foo/1.2.3 (test package): Generator 'CMakeDeps' calling 'generate()'
ERROR: Error in generator 'CMakeDeps': error generating context for 'foo/1.2.3': too many values to unpack (expected 2)

when self.cpp_info.requires contains malformed requirement strings with multiple :: separators. The error occurs deep in the CMakeDeps generator and doesn't clearly indicate that the issue is with malformed requirement syntax in the conanfile. A useful error pointing out the malformed string would have helped a lot with catching the underlying issue here!

Version Info:

version: 2.9.1
conan_path: /home/tom/.pyenv/versions/conan2env/bin/conan
python
  version: 3.10.12
  sys_version: 3.10.12 (main, Jul 30 2024, 13:07:00) [GCC 11.4.0]
  sys_executable: /home/tom/.pyenv/versions/3.10.12/envs/conan2env/bin/python
  is_frozen: False
  architecture: x86_64
system
  version: #63~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Tue Apr 22 19:00:15 UTC 2
  platform: Linux-6.8.0-60-generic-x86_64-with-glibc2.35
  system: Linux
  release: 6.8.0-60-generic
  cpu: x86_64

How to reproduce it

1. Start from conan new cmake_exe -d name=foo -d version=1.2.3
2. Add generators = "CMakeDeps", "CMakeToolchain" to test_package/conanfile.py
3. Add the following the conanfile.py

   requires = "gtest/1.14.0"
   def package_info(self):
       self.cpp_info.requires = [
           "gtest::this"  # <- How we encountered bug: Missed comma here
           "is::problem"
       ]

4. conan create .
5. Observer error

======== Testing the package ========
foo/1.2.3 (test package): Test package build: 
foo/1.2.3 (test package): Test package build folder: ./test_package
foo/1.2.3 (test package): Writing generators to ./test_package
foo/1.2.3 (test package): Generator 'CMakeDeps' calling 'generate()'
ERROR: Error in generator 'CMakeDeps': error generating context for 'foo/1.2.3': too many values to unpack (expected 2)

[memsharded]

Hi @transegn

Thanks for your feedback.

While it is true that the error message is not very clear, this wouldn't be a bug, but expected behavior. The inputs were incorrect, and the generator CMakeDeps managed to identify it and report the package foo/1.2.3 had some error.

The issue is that Python is very flexible, and the exposed API surface via conanfile.py is very large. Conan cannot be checking for all possible errors that could happen to report them adequately. Because users could do:

• Introduce an incorrect cpp_info.requires type, such as a dictionary or a custom object
• Introduce any incorrect value as requires items, not only in content/format of the value, but also different types.
• Introduce many different variants of format errors in the values, such as gtest:this, gtest;this (semi-colon), gtest:::this, etc, etc.

Every single check that is done in the code will be executed for every cpp_info.requires and for every cpp_info.components[comp].requires component information, for every package package_info() method in the dependency graph, once per instance of the package in the dependency graph (so repeated nodes such as tool-requires and test-requires will execute once per node), with an extra overhead, to try to protect against very occasional user errors.

As always, this is a trade-off, in this case a trade-off between performance (and code complexity and maintenance) and UX for user protection against syntax errors.

So I am afraid this cannot be classified as a bug to be fixed with a .requires validity check.

But I agree it would be worth to try to improve the specific message of the consumer side. I'll have a quick look if the error message in that specific consumer generator CMakeDeps might be improved to add some extra hint about the specific string. Thanks for your reproduction steps, that really help!

[memsharded]

Closed by #18552 for next Conan 2.18.
The error message will print now the specific offending string with the double "::". Thanks for the feedback!