This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

classification
Title: Splitting CJK codecs from pythoncore dll
Type: Stage:
Components: Build Versions: Python 2.5
process
Status: closed Resolution: out of date
Dependencies: Superseder:
Assigned To: hyeshik.chang Nosy List: georg.brandl, hyeshik.chang, kxroberto, loewis
Priority: normal Keywords: patch

Created on 2006-03-14 10:18 by hyeshik.chang, last changed 2022-04-11 14:56 by admin. This issue is now closed.

Files
File name Uploaded Description Edit
splitcjkcodecs.diff hyeshik.chang, 2006-03-14 10:18 first patch
Messages (11)
msg49736 - (view) Author: Hyeshik Chang (hyeshik.chang) * (Python committer) Date: 2006-03-14 10:18
As many of western language natives have requested it,
this patch splits CJK codecs from pythoncore dll. 
python25.dll shrinks from 2024KB to 1364KB by this. 
Even some of CJK users would take some advantages for
domestic py2exe-ed releases.
msg49737 - (view) Author: Martin v. Löwis (loewis) * (Python committer) Date: 2006-03-19 10:00
Logged In: YES 
user_id=21627

I must say that I find this splitting arbitrary. Why are
these modules moved out of pythonxy.dll, when, say,
multibytecodec, mmap, parser, xxsubtype, sha512 stay. I
really wish there was some policy guiding this.

That said, as the author of the code, it is certainly your
decision to make that split that way. So please go ahead and
apply that patch.

Please also add the new modules to Tool/msi/msi.py.
msg49738 - (view) Author: Hyeshik Chang (hyeshik.chang) * (Python committer) Date: 2006-03-19 10:20
Logged In: YES 
user_id=55188

Yes.  I agree about the arbitrariness where to modules go.
I thought that some dll size was concerned looking
"unicodedata" module case; which is only 396KB.  It would be
clearer if we use same scheme for Windows distribution as
specified in Modules/Setup.dist. (what drawbacks are
expected then?)

Thank you for the review!
msg49739 - (view) Author: Martin v. Löwis (loewis) * (Python committer) Date: 2006-03-19 11:56
Logged In: YES 
user_id=21627

The main issue is maintainability. Everytime a module is
added, a project file needs to be created, which is more
tedious than adding a few lines to setup.py (although one
also has to remember editing Modules/Setup.dist); in
addition, Tools/msi/msi.py needs to be edited.

Everytime a "global" change is made to the compilation
options of Python (e.g. adding AMD64 configurations), it
needs to be applied to all project files individually.
msg49740 - (view) Author: Hyeshik Chang (hyeshik.chang) * (Python committer) Date: 2006-03-19 13:22
Logged In: YES 
user_id=55188

Aah.  Then, how about to use a way how the PuTTY project does?
They maintain a relatively simple perl script to generate
various types of build files including MSVC6 dsw, MSVC7.1
sln, Dev-Cpp project, MinGW Makefile, MS nmake Makefile and
Borland C++ project file at a time.  If we make a simliar
thing, keeping PC/VC6 up-to-date would become a bit easy.

I'll try it in my spare time before the first alpha of 2.5.
msg49741 - (view) Author: kxroberto (kxroberto) Date: 2006-03-20 11:24
Logged In: YES 
user_id=972995

good decision. 
otherwise, once we learn the codecs of Marsians, they would
have probably gone into the core as well :-) 
msg49742 - (view) Author: Martin v. Löwis (loewis) * (Python committer) Date: 2006-03-20 20:53
Logged In: YES 
user_id=21627

I personally think the codec from the Marsians is more
important to be part of the core than _hotshot module.
msg49743 - (view) Author: Georg Brandl (georg.brandl) * (Python committer) Date: 2006-04-30 08:33
Logged In: YES 
user_id=849994

What's the status of this? It's marked as Accepted, but not
checked in.
msg49744 - (view) Author: Martin v. Löwis (loewis) * (Python committer) Date: 2006-04-30 09:22
Logged In: YES 
user_id=21627

See my message from 2006-03-19 11:00. It's accepted, and
perky could check it in if he wanted to.
msg49745 - (view) Author: Martin v. Löwis (loewis) * (Python committer) Date: 2007-03-05 13:24
The patch doesn't apply cleanly anymore, so I'm marking it out-of-date now. If it hasn't been committed 6 months from now, I think it should be closed.
msg61271 - (view) Author: Georg Brandl (georg.brandl) * (Python committer) Date: 2008-01-20 11:34
Another year has passed; closing this.
History
Date User Action Args
2022-04-11 14:56:15adminsetgithub: 43033
2008-01-20 11:34:36georg.brandlsetstatus: open -> closed
messages: + msg61271
2006-03-14 10:18:36hyeshik.changcreate