[doNOTmerge] check if filesize increase depends on comment title

[ferdymercury]

This Pull request:

Changes or fixes:

See #18459 (comment)

[ferdymercury]

No, it does not depend on the number of functions.
Trying now removing things.
Checking now if the const intmight be the problem.

[github-actions]

Test Results

     6 files       6 suites   22h 0m 36s ⏱️
 2 694 tests  2 694 ✅ 0 💤 0 ❌
16 042 runs  16 042 ✅ 0 💤 0 ❌

Results for commit 27ae438.

♻️ This comment has been updated with latest results.

[ferdymercury]

@linev @pcanal this little change is the one that leads to a change in the compressed file size:

27ae438

[linev]

@pcanal

Is it by chance class checksum which changes after changing in class methods?

[ferdymercury]

@pcanal

Is it by chance class checksum which changes after changing in class methods?

Note that if I add many other functions the filesize does not change. Only if a parameter "const int" instead of "int" appears.

[linev]

For me checksum remains.

[ferdymercury]

EDIT: False alarm, it seems it's not related to the const int.
Now trying with the modified comment descriptor.

[linev]

@ferdymercury
Yes, changed comments can be a reason. They stored with streamer infos in the file.
Probably just modify your original PR and do not touch comments there .

[ferdymercury]

@ferdymercury Yes, changed comments can be a reason. They stored with streamer infos in the file. Probably just modify your original PR and do not touch comments there .

But I need there MENU comments anyway, so I will need to still modify the roottests ? Or only comments of members are stored but not comments of functions ?

[linev]

Or only comments of members are stored but not comments of functions ?

Yes, only persistent members comments stored in the file.
Class methods are not listed in streamer infos.

[linev]

You can check here that is stored in streamer infos of hsimple.root file.

[pcanal]

Is it by chance class checksum which changes after changing in class methods?

The CheckSum should only change if you change the type (or name) of any persistent data member. The comment should have no effect (hopefully)