NSIS installer is NSIS installer stub (EXE) + NSIS archive. NSIS archive can contain NSIS uninstaller.
NSIS uninstaller is special NSIS uninstall stub + NSIS uninstaller archive.
But NSIS uninstaller stub is produced from NSIS installer stub with special patch data stored in NSIS archive.
When NSIS handler in 7-Zip doesn't have access to NSIS EXE installer stub, then it can't create correct NSIS uninstaller NSIS stub (EXE), so it shows only
NSIS uninstaller archive.
Now there are still problems in 7-Zip code, when NSIS handler works after PE parser. And NSIS code can extract only NSIS archive (without exe stub).
But it's possible to use -tnsis switch to extract uninst.exe:
7z x -tnsis
Last edit: Igor Pavlov 2014-06-16
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Игорь, Вы меня поражаете! :)
Новый ключ, описание которого в справке состоит из одной строки в таблице ключей, а синтаксис ключа вообще не приведен. Кто же догадается как им пользоваться!?
Извините, ворчу по стариковски...
Спасибо.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Please keep this code in future versions.
Even "bloated", it's an absolute must have feature.
IDK about +some Kb for 7z.exe.
Everyone have multi-GB HDDs these days, so size is not a problem at all.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
p7zip have not been updated. I need to upgrade to at least version 9.30
Last edit: Gemorroj 2014-06-19
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Anonymous
-
2014-06-16
Please add, ...
1 ) 7Zip option [Temporary working directory in memory, until a max memory size is selected.].
I still miss than option, ...
For the moment, did RamDisk it's the only way, [until temporary working directory in memory], will be an option inside 7Zip.
My computer have 8 Gig. Ram, Windows 8.1 update 1 x64.
Like a lot the NSIS decompiling functionality. Maybe it is not useful for end users, but for sure is better having this [NSIS.nsi] file, than having nothing.
Also love new usability enhancements as seen in Easy 7-Zip.
Hope next release will give more Easy 7-Zip features integrated as well as some core changes such as RAR5 decompression, multimedia compression, ...
Thanks Igor!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
9.33 is a small step yet, but it is towards a great goal.
As for NSIS decompiling, I think it may be provided optionally (maybe even in a separate DLL). But anyway, since the code has been already written, it should not be discarded for good.
Last edit: Shell 2014-06-16
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Estimated Igor, could you publish here the SHA-1 sum for the installer of all 7-Zip 9.33 versions. It will be useful , to know, that we are working with your real releases and not some forged ones. RAR5 is a must. Thanks for your effort. Remember The Truecrypt WTF
Last edit: gogo 2014-06-16
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thanks a lot @Ricardo for your valuable collaboration. But the post of the hashes must be done by Igor Pavlov, so it could be taken OFFICIALLY. And peace of mind...
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
lol, if we two have the same hashes, then you could be enough sure the server sent us the correct file. You can test by downloading multiple times from different connections and checking the hashes every time. Anyway, if you get a different hash, very probably the download was corrupted and won't even execute.
A difference of hashes would be a concern if you downloaded from a different server other than the one Igor published. Otherwise, if you got it from the official server, a hash difference would mean just a small corruption caused by some sporadic network issue, not a fake or malicious file. This is peace of mind.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
@Ricardo, yes we have both of us, same hashes. But in some form, that coincidence made me to feel more confortable. I always download from www.7-zip.org the official page for security means. But that doesnt mean that could be one "man in the middle" sending to you and me, and thwousands of users, the same files redirecting the download queries, that is one common NSA capability. Thanks again for your great and valuable answers, but trust me... Igor must publish the hashes.... So in case of....
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Well, the NSA can probably also post here as Igor, so you can never be sure. Maybe if Igor visits you personally and brings the hashes – if you can be really sure that it's the real Igor! You have to draw your line of trust somewhere. As soon as someone like the NSA gets involved, you cannot even really trust a root-signed certificate.
Last edit: Daniel Saner 2014-06-19
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thanks a lot for new version.
Many useful must have features were added.
Please check updated language file for Japanese, which I've just made. https://sourceforge.net/p/sevenzip/patches/241/
Last edit: NoAngel 2014-06-17
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
On 7z-Swap4 archive with dummy short "sfx header" 64-bit 7zFM.exe crashes, 32-bit 7zFM.exe opens associated external program (winrar, works).
7z.dll dropped into Far3 ArcLite plugin doesn't like some Split archives: crashes inside in_arc->Open, but the same works in 7zFM.exe. (Previous case also crashed here.)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
1) No idea how it was generated. Certain site uploads video in this way.
2) Any recent 32-bit Far 3 will do, for example latest build: http://farmanager.com/nightly/Far30b3953.x86.20140619.7z
It was already broken in 9.32, but worked ok in 9.30. ArcLite catches crashes so it just doesn't enter archive instead of showing actual crash.
3) Another 'bug' I remember well in 7-zip but unfortunately can't find examples: there are some zip archives (something Chinese I suppose) that have UTF-8 flag set in central directory records, but not in file records (or the other way around...). The rest of parameters are identical there. 7-zip refuses to unpack such archives, but WinRar has no problems there.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
2) I've fixed that BUG. It was BUG in CAB code. Thanks!
7-Zip didn't see that bug, since 7-Zip now uses new code to find archive signatures. That new code placed in 7z.exe (outside of 7z.dll).
3) Yes, 7-Zip doesn't like some collisions in zip headers. Now I'm not ready to change that code. Maybe later.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
NSIS installer is NSIS installer stub (EXE) + NSIS archive. NSIS archive can contain NSIS uninstaller.
NSIS uninstaller is special NSIS uninstall stub + NSIS uninstaller archive.
But NSIS uninstaller stub is produced from NSIS installer stub with special patch data stored in NSIS archive.
When NSIS handler in 7-Zip doesn't have access to NSIS EXE installer stub, then it can't create correct NSIS uninstaller NSIS stub (EXE), so it shows only
NSIS uninstaller archive.
Now there are still problems in 7-Zip code, when NSIS handler works after PE parser. And NSIS code can extract only NSIS archive (without exe stub).
But it's possible to use -tnsis switch to extract uninst.exe:
7z x -tnsis
Last edit: Igor Pavlov 2014-06-16
Thank you, Igor. Very good switch. It is a pity that in the manual about {archive_type} = NSIS not written.
Also it's possible to use -stx{Type} switch:
-stxpe
So 7-Zip will not use PE parser. It can help for another archive types inside EXE file.
Last edit: Igor Pavlov 2014-06-16
Игорь, Вы меня поражаете! :)
Новый ключ, описание которого в справке состоит из одной строки в таблице ключей, а синтаксис ключа вообще не приведен. Кто же догадается как им пользоваться!?
Извините, ворчу по стариковски...
Спасибо.
Please keep this code in future versions.
Even "bloated", it's an absolute must have feature.
IDK about +some Kb for 7z.exe.
Everyone have multi-GB HDDs these days, so size is not a problem at all.
Linux plz.... =(
p7zip? no?
p7zip have not been updated. I need to upgrade to at least version 9.30
Last edit: Gemorroj 2014-06-19
Please add, ...
1 ) 7Zip option [Temporary working directory in memory, until a max memory size is selected.].
I still miss than option, ...
For the moment, did RamDisk it's the only way, [until temporary working directory in memory], will be an option inside 7Zip.
My computer have 8 Gig. Ram, Windows 8.1 update 1 x64.
2 ) 7Zip option [Convert compressed archived files].
I have also, still miss than option, ...
Convert from ".zip to .7z", ...
Thanks in advance, ...
Like a lot the NSIS decompiling functionality. Maybe it is not useful for end users, but for sure is better having this [NSIS.nsi] file, than having nothing.
Also love new usability enhancements as seen in Easy 7-Zip.
Hope next release will give more Easy 7-Zip features integrated as well as some core changes such as RAR5 decompression, multimedia compression, ...
Thanks Igor!
+1
9.33is a small step yet, but it is towards a great goal.As for NSIS decompiling, I think it may be provided optionally (maybe even in a separate DLL). But anyway, since the code has been already written, it should not be discarded for good.
Last edit: Shell 2014-06-16
Estimated Igor, could you publish here the SHA-1 sum for the installer of all 7-Zip 9.33 versions. It will be useful , to know, that we are working with your real releases and not some forged ones. RAR5 is a must. Thanks for your effort. Remember The Truecrypt WTF
Last edit: gogo 2014-06-16
7z933-x64.msi
SHA512: 90A17997ED17A25DFDDC18FF8AEF1558AFC2AD0A3C8443854FEDAC706B2DAF9DE274EF55D9737174D94EE556F5283E5E9A397AC8BD19B85F584C43667D2F32EF
SHA256: 48218C2171C8FEE49A00F3C13072DED3B26607835CBECD215A9ED09CCD4B7B46
SHA1: FDC8F5BFEB2099159DA77CE3F0E79886B0A8866D
7z933.msi
SHA512: DA6645C0AD4CB476F0151F0B452CA7CBD2CEAAACC9E5422F1471B7E0B1B95820FE204EE8E687B862D35C09B71C7E7E53DFD7B025F3B0AC2B93A9B77B58884228
SHA256: 98978873A3D66949E68F1D77142AA128A1AB9423744E284567C91FF9D6AAF09B
SHA1: B1B0BD61D7187DE8298CA879F5E616CDCC193CE2
7z933.exe
SHA512: 74CF0DABF13E8E6AD725F2E89481F3CD87D7DDC714E2177658682070F239D74E7E3F0FEFC3C1302D4B487F06C4EAC620226FCC0960F94AA34FD4634865739909
SHA256: E1420B32F8D861D75DF66D3948266FF26524B12FFCAF83F7CB3ADEFA0BC05856
SHA1: 346BD9F93867EE98AE44CF01904393514C48D41D
Thanks a lot @Ricardo for your valuable collaboration. But the post of the hashes must be done by Igor Pavlov, so it could be taken OFFICIALLY. And peace of mind...
lol, if we two have the same hashes, then you could be enough sure the server sent us the correct file. You can test by downloading multiple times from different connections and checking the hashes every time. Anyway, if you get a different hash, very probably the download was corrupted and won't even execute.
A difference of hashes would be a concern if you downloaded from a different server other than the one Igor published. Otherwise, if you got it from the official server, a hash difference would mean just a small corruption caused by some sporadic network issue, not a fake or malicious file. This is peace of mind.
@Ricardo, yes we have both of us, same hashes. But in some form, that coincidence made me to feel more confortable. I always download from www.7-zip.org the official page for security means. But that doesnt mean that could be one "man in the middle" sending to you and me, and thwousands of users, the same files redirecting the download queries, that is one common NSA capability. Thanks again for your great and valuable answers, but trust me... Igor must publish the hashes.... So in case of....
Well, the NSA can probably also post here as Igor, so you can never be sure. Maybe if Igor visits you personally and brings the hashes – if you can be really sure that it's the real Igor! You have to draw your line of trust somewhere. As soon as someone like the NSA gets involved, you cannot even really trust a root-signed certificate.
Last edit: Daniel Saner 2014-06-19
Hello, Igor!
Thanks a lot for new version.
Many useful must have features were added.
Please check updated language file for Japanese, which I've just made.
https://sourceforge.net/p/sevenzip/patches/241/
Last edit: NoAngel 2014-06-17
https://sourceforge.net/p/sevenzip/bugs/1418/
a bug
This version is far from stable I'd have to say.
On 7z-Swap4 archive with dummy short "sfx header" 64-bit 7zFM.exe crashes, 32-bit 7zFM.exe opens associated external program (winrar, works).
7z.dll dropped into Far3 ArcLite plugin doesn't like some Split archives: crashes inside in_arc->Open, but the same works in 7zFM.exe. (Previous case also crashed here.)
Please write exact instructions to reproduce swap4 problem.
Or upload bad archive to some server.
http://anonfiles.com/file/d1da2dba1afc71170456fbcb1b4c5223
http://anonfiles.com/file/92004c1f2d258119032dd84816cf26ce
Both demonstrate first problem, after unpacking you can see the second.
If the second one isn't from internal bug, but rather API change – can you post the updated headers for direct usage of 7z.dll?
1) I'll fix the bug with that swap4 archive.
How did you create that archive?
2) About ArcLite. Please describe exact FAR version (version number, 32-bit / 64-bit).
Did it work with 7z.dll from 7-Zip 9.32?
1) No idea how it was generated. Certain site uploads video in this way.
2) Any recent 32-bit Far 3 will do, for example latest build: http://farmanager.com/nightly/Far30b3953.x86.20140619.7z
It was already broken in 9.32, but worked ok in 9.30. ArcLite catches crashes so it just doesn't enter archive instead of showing actual crash.
3) Another 'bug' I remember well in 7-zip but unfortunately can't find examples: there are some zip archives (something Chinese I suppose) that have UTF-8 flag set in central directory records, but not in file records (or the other way around...). The rest of parameters are identical there. 7-zip refuses to unpack such archives, but WinRar has no problems there.
2) I've fixed that BUG. It was BUG in CAB code. Thanks!
7-Zip didn't see that bug, since 7-Zip now uses new code to find archive signatures. That new code placed in 7z.exe (outside of 7z.dll).
3) Yes, 7-Zip doesn't like some collisions in zip headers. Now I'm not ready to change that code. Maybe later.