no checks in ident string

[ewongbb]

I'm currently trying to convert a large mercurial repo to git using fast-export. After running it for some time, fast-export crashes with:

fatal: Missing > in ident string: test user <test@test.com 123456778 +1200
with the following traceback:
Traceback (most recent call last):
File "/home/cc/projs/fast-export/hg-fast-export.py", line 442, in
notes=options.notes,encoding=encoding))
File "/home/cc/projs/fast-export/hg-fast-export.py", line 364, in hg2git
c=export_commit(ui,repo,rev,old_marks,max,c,authors,sob,brmap,hgtags,notes,encoding)
File "/home/cc/projs/fast-export/hg-fast-export.py", line 226, in export_commit
export_file_contents(ctx,man,changed,hgtags,encoding)
File "/home/cc/projs/fast-export/hg-fast-export.py", line 140, in export_file_contents
wr(d)
File "/home/cc/projs/fast-export/hg-fast-export.py", line 37, in wr
wr_no_nl(msg)
File "/home/cc/projs/fast-export/hg-fast-export.py", line 34, in wr_no_nl
sys.stdout.write(msg)
IOError: [Errno 32] Broken pipe

[ewongbb]

my current 'fix'/hack is:
diff --git a/hg-fast-export.py b/hg-fast-export.py
index 0c683c5..2f1c90b 100755
--- a/hg-fast-export.py
+++ b/hg-fast-export.py
@@ -184,6 +184,8 @@ def export_commit(ui,repo,revision,old_marks,max,count,authors,sob,brmap,hgtags,
wr('mark :%d' % (revision+1))
if sob:
wr('author %s %d %s' % (get_author(desc,user,authors),time,timezone))

• if '<<' in user:
• user.replace('<<', '<')
  wr('committer %s %d %s' % (user,time,timezone))
  wr('data %d' % (len(desc)+1)) # wtf?
  wr(desc)

of course, with the different possible ways to break that, it'd have to be a bit better at
checking the ident syntax, et. al. (but right now it fixes my issue)

[ewongbb]

nvm.. the fix didn't work.

Should've been |user = user.replace('<<', '<')

[frej]

What you want to use is the author map (-A) option (Look in the
README).

Fast-export is not in the business of doing cute transformations in
order to try to guess the correct author string. The policy is to keep
things simple. If you have too many malformed identities to deal with
them manually, write an external tool which produces the author map.

[frej]

Won't fix

[c4lliope]

Picking up this long-dead thread,
I had a similar issue parsing Mozilla's source;
a couple <<address@place> cases and similar.

I am mirroring this code base, and in this code base I made a couple changes;
see the change log.

Open an issue here should you experience similar issues,
and link the code base you need help on.

[c4lliope]

Once I searched around, I realized https://github.com/glandium/git-cinnabar handles these cases nicely,
and is a really speedy program.

I'm mirroring the code here.