Hi! I found (and fixed) a few Bugs in the file local/bin/sage-preparse.
These are the things I fixed: * The module docstrings disappeared when preparsing because the preparse_file function inserted those numeric_literals definitions before the docstrings. * Now also unicode-docstrings (e.g. u"""foo""") are recognized as docstrings. Also raw docstrings may now use an upper case R as string modifier (R"""foo""" would work now) which is allowed in Python. * Now all coding-comments as specified by Python are found and excluded from preparsing. * I did not fix a bug that occurs when a statement is on the same line where the docstring ends (e.g. """foo"""; print 2^5). It will not be preparsed! I added a TODO-comment on the according line. I don't know if I am allowed to attach files here, so I'm going to host the diff and my fixed version of the file elsewhere: http://spielwiese.hsg-kl.de/~cui/bug-fix-preparse/ Included is the original file as of sage version 4.4 on opensuse 64bit, the file that contains my bugfixes and a diff file of those two. greetings, David Poetzsch-Heffter. PS: If I am allowed to attach files, the named files come also attached with this e-mail. -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
sage-preparse.bugfix
Description: Binary data
8a9,12
> -- David Poetzsch-Heffter (2010): fix trac #????
> (Recognize unicode docstrings, factoring out numeric literals
> won't destroy the docstring search, all allowed encoding
> comments are found)
63a68,72
> # We use this regexp to check if a line is a coding-comment.
> # This regexp is officially specified by python for
> # recognizing coding-comments.
> coding_regexp = re.compile(r"coding[:=]\s*([-\w.]+)")
>
106,113c115,149
<
< #Check to see if a coding is specified in the .sage file.
< #If it is, then we want to copy it over to the new file
< #and not include it in the preprocessing
< if F.startswith('# -*- coding:'):
< end = F.find('\n')
< coding = F[:end+1] if end != -1 else F
< F = F[len(coding):]
---
>
> # It is ** critical ** that we put all inserts made via preparsing
> # after the module docstring, since otherwise the module docstring
> # will disappear
> i = find_position_right_after_module_docstring(F)
> docstring = F[:i]
> F = F[i:]
>
> # If a docstring was found the coding comment should be in the comments
> # preceeding the docstring and is therefore included in the
> # docstring-variable. So we have to look for a coding-comment only if
> # no docstring was found!
> if (i == 0):
> # A coding-comment may appear in the first or in the second
> # line of a file. If there is such a comment, then we want to
> # copy it over to the new file and not include it in the
> # preprocessing.
> lines = F.splitlines()
> if coding_regexp.search(lines[0]):
> coding = lines[0] + "\n"
> F = F[len(coding):]
> elif coding_regexp.search(lines[1]):
> coding = "\n".join(lines[:2]) + "\n"
> F = "\n".join(lines[2:])
> else:
> coding = ''
> elif "\n" not in docstring.strip():
> # Ok, there is one possibility: If the found docstring is only
> # a single line the second line may be a coding comment.
> lines = F.splitlines()
> if coding_regexp.search(lines[0]):
> coding = lines[0] + "\n"
> F = F[len(coding):]
> else:
> coding = ''
116c152
<
---
>
124d159
<
126,127d160
< # It is ** critical ** that we put this after the mdoule docstring, since
< # otherwise the module docstring will disappear.
129,130c162
< i = find_position_right_after_module_docstring(G)
< G = coding + G[:i] + insert + G[i:]
---
> G = docstring + coding + insert + G
163c195,196
< if not (n[0] in ['"',"'"] or n[0:2] in ['r"',"r'"]):
---
> possible_docstring_starts = ['"',"'", 'r"',"r'", 'u"', "u'", 'ur"', "ur'"]
> if not any( map(n.lower().startswith, possible_docstring_starts) ):
171c204,205
< n = n.lstrip('r') # strip leading r if there is one
---
> # strip leading u and/or r if there is one
> n = n.lstrip('u').lstrip('U').lstrip('r').lstrip('R')
177a212,213
> # TODO: If there are statements in the same line where the
> # docstring ends they will not be preparsed!
sage-preparse.orig-4.4
Description: Binary data
