Nemcio/Fast-Export
1
0
Fork 0
mirror of https://github.com/frej/fast-export.git synced 2025-11-02 09:15:48 +01:00
Code Issues Packages Projects Releases Activity

Fix: Largefiles ignored #141

Browse Source 
Import mercurial large files as ordinary files into git

The basic idea to this fix is based on
https://github.com/planestraveler/fast-export/tree/add-lfs-support-v2
from PR #65

Closes #141
This commit is contained in:
Frank Zingsheim
2024-09-11 15:48:26 +02:00
committed by Frank Zingsheim
parent 0afd336d6f
commit bd707b5d6e
3 changed files with 94 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
hg-fast-export.py
View File

@@ -1,6 +1,7 @@
#!/usr/bin/env python3
# Copyright (c) 2007, 2008 Rocco Rutte <pdmef@gmx.net> and others.
# Copyright (c) 2025 Siemens
# License: MIT <http://www.opensource.org/licenses/mit-license.php>
from hg2git import setup_repo,fixup_user,get_branch,get_changeset
@@ -11,6 +12,7 @@ import sys
import os
from binascii import hexlify
import pluginloader
from hgext.largefiles import lfutil
# silly regex to catch Signed-off-by lines in log message
sob_re=re.compile(b'^Signed-[Oo]ff-[Bb]y: (.+)$')
@@ -162,6 +164,32 @@ def refresh_gitmodules(ctx):
wr(b'M 100644 inline .gitmodules')
wr_data(gitmodules)
def is_largefile(filename):
return filename[:6] == b'.hglf/'
def largefile_orig_name(filename):
return filename[6:]
def largefile_data(ctx, file, filename):
lf_file_ctx=ctx.filectx(file)
lf_hash=lf_file_ctx.data().strip(b'\n')
sys.stderr.write("Detected large file hash %s\n" % lf_hash.decode())
#should detect where the large files are located
file_with_data = lfutil.findfile(ctx.repo(), lf_hash)
if file_with_data is None:
# Autodownloading from the mercurial repository would be an issue as there
# is a good chance that we may need to input some username and password.
# This will surely break fast-export as there will be some unexpected
# output.
sys.stderr.write("Large file wasn't found in local cache.\n")
sys.stderr.write("Please clone with --all-largefiles\n")
sys.stderr.write("or pull all large files with 'hg lfpull --rev "
"\"all()\"'\n")
# closing in the middle of import will revert everything to the last checkpoint
sys.exit(3)
with open(os.path.normpath(file_with_data), 'rb') as file_with_data_handle:
return file_with_data_handle.read()
def export_file_contents(ctx,manifest,files,hgtags,encoding='',plugins={}):
count=0
max=len(files)
@@ -183,8 +211,12 @@ def export_file_contents(ctx,manifest,files,hgtags,encoding='',plugins={}):
b'Ignoring file %s which cannot be tracked by git\n' % filename
)
continue
file_ctx=ctx.filectx(file)
d=file_ctx.data()
if is_largefile(filename):
filename = largefile_orig_name(filename)
d = largefile_data(ctx, file, filename)
else:
file_ctx=ctx.filectx(file)
d=file_ctx.data()
if plugins and plugins['file_data_filters']:
file_data = {'filename':filename,'file_ctx':file_ctx,'data':d}
@@ -327,6 +359,8 @@ def export_commit(ui,repo,revision,old_marks,max,count,authors,
filename=strip_leading_slash(filename)
if filename==b'.hgsub':
remove_gitmodules(ctx)
if is_largefile(filename):
filename=largefile_orig_name(filename)
wr(b'D %s' % filename)
export_file_contents(ctx,man,modified,hgtags,fn_encoding,plugins)