[hageboeck]

This adapts the header cleanup campaign to the user request of leaving deprecation messages visible for much more than one year.
For almost all headers, the removal date is now "ROOT 7", opening a path for escalating the warnings to #error for the first release(s) of ROOT 7.

@ferdymercury thanks for your cleanup campaign! We discussed the proposed timeline and suggest to give users a bit more time, so they don't find themselves in a situation where code does not compile without an obvious way how to fix it.

The proposed solution is:

• The headers will be installed for longer
• In ROOT 7, they will not emit a #warning but an #error
• If a packager decides that they want to remove these headers earlier, they are encouraged to do so. In 6.44, we could even provide a build option called something like -DInstallCompatHeaders=On/Off, which can be used by packagers to clean up the package.

[pcanal]

Thanks (Consider applying the clang-format suggestions)

[ferdymercury]

• If a packager decides that they want to remove these headers earlier, they are encouraged to do so. In 6.44, we could even provide a build option called something like -DInstallCompatHeaders=On/Off, which can be used by packagers to clean up the package.

Having this option sounds like the perfect solution in a world of infinite resources, but, in my opinion, it will be a maintenance nightmare, taking developers much more time than helping the-few-users in the forum not asking first LLMs in finding out how to fix their potential compilation errors.

In any case, the RNotes must be updated accordingly ;)

[github-actions]

Test Results

    24 files      24 suites   4d 2h 44m 38s ⏱️
 3 896 tests  3 896 ✅ 0 💤 0 ❌
83 440 runs  83 440 ✅ 0 💤 0 ❌

Results for commit 83d0a1e.

[hageboeck]

Thanks (Consider applying the clang-format suggestions)

It did consider it, but having the #warning and the following messages in the same line has its benefits (e.g. for grep).

[vepadulano]

I still see here after and e.g. in in Bswapcpy.h

[vepadulano]

This one has an even different formulation "removed later on"

[hageboeck]

This one is indeed a special case, because it's not clear whether it can be removed.

[ferdymercury]

cleanup campaign!

Side note: the remaining ones that do not conform to the the R* T* pattern are:

Buttons.h
Bytes.h
Byteswap.h
Compression.h
DllImport.h
ESTLType.h
Foption.h
Getline.h
GuiTypes.h
HFitInterface.h
KeySymbols.h
MPCode.h
MPSendRecv.h
MessageTypes.h
PoolUtils.h
ThreadLocalStorage.h
root_std_complex.h

but it should be decided whether they should go into ROOT/ or how to approach it.