Update to 24.09 troubles ...

[sebres]

Trying to merge mainline 24.09, there is some crossing of hashers (introduced in 24.09) by its registration:

24.09 (0x208):

7-Zip-zstd/CPP/Common/Md5Reg.cpp

Line 44 in e10d992

REGISTER_HASHER(CMd5Hasher, 0x208, "MD5", MD5_DIGEST_SIZE)

7-Zip ZS (0x207):

7-Zip-zstd/CPP/Common/Md5Reg.cpp

Line 43 in 4c42558

REGISTER_HASHER(CMD5Hasher, 0x207, "MD5", MD5_DIGEST_LENGTH)

---

24.09 (0x222 and 0x223):

7-Zip-zstd/CPP/Common/Sha512Reg.cpp

Lines 82 to 83 in e10d992

	REGISTER_SHA512_HASHER (Sha384Hasher, 0x222, "SHA384", SHA512_384_DIGEST_SIZE)
	REGISTER_SHA512_HASHER (Sha512Hasher, 0x223, "SHA512", SHA512_DIGEST_SIZE)

7-Zip ZS (0x209 and 0x208):

7-Zip-zstd/CPP/Common/Sha512Reg.cpp

Line 43 in 4c42558

REGISTER_HASHER(CSHA512Hasher, 0x209, "SHA512", SHA512_DIGEST_LENGTH)

7-Zip-zstd/CPP/Common/Sha384Reg.cpp

Line 43 in 4c42558

REGISTER_HASHER(CSHA384Hasher, 0x208, "SHA384", SHA384_DIGEST_LENGTH)

---

@mcmilk what would you suggest here:

1. drop that modules introduced in 24.09 and use 7-Zip ZS (inclusive registration IDs)
2. drop 7-Zip ZS modules and use modules introduced in 24.09 (inclusive registration IDs of them)
3. drop 7-Zip ZS modules and use modules introduced in 24.09 (but adjust registration IDs to 7-Zip ZS)
4. mix of both, somehow (they deviate by their case-sensitivity and name, CMd5Hasher vs CMD5Hasher, but what is about their numbers)?
5. alternative?

By the way, have you talked to Igor about a possible merge of 7-Zip ZS (in parts or completely)? Is there any chance for that?..

[mcmilk]

We can remove the own hashes, when Igor added them.
The IDs are not important here. I will take also a look to it today.

Edit: I think one hash was not properly added by me last time, so the hashsum was wrong also.

[sebres]

Edit: I think one hash was not properly added by me last time, so the hashsum was wrong also.

Which one?

[mcmilk]

It's listed in issue #336 - the SHA3 thing.

[sebres]

It's listed in issue #336 - the SHA3 thing.

I cannot reproduce it with new branch 24.08 (#377 (comment), probably fixed somewhere by merge or conflicts resolving)... I get proper SHA3-256 in any case, by * and by direct menu items.
So the issue #336 could be closed if it gets merged.

[sebres]

Found it:
4f9c3eb#diff-3efab4a12e4465b384048d18e8a8356e17d932d122f75aa18f56efe815f8c5ceL309-R311

[isarrider]

Thanks folks...
Thats why I bought you 2 beers recently ;)

[MouriNaruto]

I think I should provide my opinions which may help.

According to my 7-Zip mainline merging experiences in NanaZip, maybe 7-Zip ZS can separate the extended implementations to the separate dynamic/static library which follow the 7-Zip interfaces ABI for better maintenance.

I use the different GUID scheme to make my extended implementations will not affected by Igor's changes.

Here is my work in progress subproject for create 7-Zip interface definition without 7-Zip source code because Igor may changes a lot in the future like 22.xx to 23.xx.

• https://github.com/M2Team/NanaZip/blob/main/NanaZip.Codecs/NanaZip.Codecs.h
• https://github.com/M2Team/NanaZip/blob/main/NanaZip.Specification/NanaZip.Specification.SevenZip.h

Kenji Mouri