See http://bugzilla.lyx.org/show_bug.cgi?id=1990.
lyx2lyx does not revert pagebreak and lyxline insets and can therefore
create invalid .lyx files. The attached patch fixes that by extending
convert_vspace() to revert_breaks(), making it symmetric to
convert_breaks().
I have tested it with file attached to the bug report. The test file
contains all cases that I can think of, and it works.
Lars, can I commit this?
Georg
diff -p -r -U 3 -X excl.tmp lyx-1.4-clean/lib/lyx2lyx/ChangeLog lyx-1.4-cvs/lib/lyx2lyx/ChangeLog
--- lyx-1.4-clean/lib/lyx2lyx/ChangeLog 2005-08-19 10:47:40.000000000 +0200
+++ lyx-1.4-cvs/lib/lyx2lyx/ChangeLog 2005-08-28 17:28:20.000000000 +0200
@@ -1,3 +1,8 @@
+2005-08-28 Georg Baum <[EMAIL PROTECTED]>
+
+ * lyx_1_4.py (convert_vspace): rename to revert_breaks and revert
+ \lyxline and \newpage, too
+
2005-08-18 José Matos <[EMAIL PROTECTED]>
* LyX.py (read, write): add preamble as data member of LyX_Base,
diff -p -r -U 3 -X excl.tmp lyx-1.4-clean/lib/lyx2lyx/lyx_1_4.py lyx-1.4-cvs/lib/lyx2lyx/lyx_1_4.py
--- lyx-1.4-clean/lib/lyx2lyx/lyx_1_4.py 2005-08-19 10:47:41.000000000 +0200
+++ lyx-1.4-cvs/lib/lyx2lyx/lyx_1_4.py 2005-08-28 17:22:58.000000000 +0200
@@ -907,92 +907,258 @@ def convert_ertbackslash(body, i, ert):
return i
-def convert_vspace(file):
+# get all paragraph parameters. They can be all on one line or on several lines.
+# lines[i] must be the first parameter line
+def get_par_params(lines, i):
+ par_params = ('added_space_bottom', 'added_space_top', 'align',
+ 'labelwidthstring', 'line_bottom', 'line_top', 'noindent',
+ 'pagebreak_bottom', 'pagebreak_top', 'paragraph_spacing',
+ 'start_of_appendix')
+ # We cannot check for '\\' only because paragraphs may start e.g.
+ # with '\\backslash'
+ params = ''
+ while lines[i][:1] == '\\' and split(lines[i][1:])[0] in par_params:
+ params = params + ' ' + strip(lines[i])
+ i = i + 1
+ return strip(params)
+
+
+# Change vspace insets, page breaks and lyxlines to paragraph options
+# (if possible) or ERT
+def revert_breaks(file):
# Get default spaceamount
i = find_token(file.header, '\\defskip', 0)
if i == -1:
- defskipamount = 'medskip'
+ defskipamount = 'medskip'
else:
- defskipamount = split(file.header[i])[1]
+ defskipamount = split(file.header[i])[1]
+
+ keys = {"\\begin_inset" : "vspace", "\\lyxline" : "lyxline",
+ "\\newpage" : "newpage"}
+ keywords_top = {"vspace" : "\\added_space_top", "lyxline" : "\\line_top",
+ "newpage" : "\\pagebreak_top"}
+ keywords_bot = {"vspace" : "\\added_space_bottom", "lyxline" : "\\line_bottom",
+ "newpage" : "\\pagebreak_bottom"}
+ tokens = ["\\begin_inset VSpace", "\\lyxline", "\\newpage"]
# Convert the insets
i = 0
while 1:
- i = find_token(file.body, '\\begin_inset VSpace', i)
+ i = find_tokens(file.body, tokens, i)
if i == -1:
return
- spaceamount = split(file.body[i])[2]
+ lines = list()
+ insets = list()
+ lines.append(split(file.body[i]))
+ insets.append(keys[lines[0][0]])
+
+ # Are we at the beginning of a paragraph?
+ paragraph_start = 1
+ start = get_paragraph(file.body, i) + 1
+ params = get_par_params(file.body, start)
+ # Paragraph parameters may be on one or more lines.
+ # Find the start of the real paragraph text.
+ while file.body[start][:1] == '\\' and split(file.body[start])[0] in params:
+ start = start + 1
+ for k in range(start, i):
+ if is_nonempty_line(file.body[k]):
+ paragraph_start = 0
+ break
+
+ # Are we at the end of a paragraph?
+ paragraph_end = 1
+ if insets[0] == "vspace":
+ j = find_end_of_inset(file.body, i)
+ if j == -1:
+ file.warning("Malformed lyx file: Missing '\\end_inset'.")
+ i = i + 1
+ continue
+ else:
+ j = i
+ end = get_next_paragraph(file.body, i)
+ for k in range(j + 1, end):
+ if is_nonempty_line(file.body[k]):
+ paragraph_end = 0
+ break
+
+ # Detect paragraphs created by convert_breaks()
+ before = 0
+ after = 0
+ if paragraph_start:
+ # Make a copy of the paragraph, since we need to strip empty lines
+ paragraph = list()
+ for k in range(start, end):
+ if is_nonempty_line(file.body[k]):
+ if (len(paragraph) > 0 and
+ find_token(file.body, "\\end_inset", k) == k):
+ paragraph[len(paragraph) - 1] = paragraph[len(paragraph) - 1] + ' ' + file.body[k]
+ else:
+ paragraph.append(file.body[k])
+
+ # Do we have a second inset to convert in this paragraph?
+ if ((len(paragraph) == 2 or len(paragraph) == 3) and
+ find_tokens(paragraph, tokens, 1) == 1):
+ lines.append(split(paragraph[1]))
+ insets.append(keys[lines[1][0]])
+ del paragraph[1]
+ # Do we have a third inset to convert in this paragraph?
+ if (len(paragraph) == 2 and
+ find_tokens(paragraph, tokens, 1) == 1):
+ lines.append(split(paragraph[1]))
+ insets.append(keys[lines[2][0]])
+ del paragraph[1]
+
+ if len(paragraph) == 1:
+ # We have only insets that need to be converted in this paragraph
+ if ((len(insets) == 3 and insets[0] == "newpage" and
+ insets[1] == "vspace" and insets[2] == "lyxline") or
+ (len(insets) == 2 and
+ ((insets[0] == "newpage" and insets[1] == "vspace") or
+ (insets[0] == "newpage" and insets[1] == "lyxline") or
+ (insets[0] == "vspace" and insets[1] == "lyxline")))):
+ # This paragraph has been created before a paragraph by
+ # convert_breaks()
+ before = 1
+ paragraph_start = 1
+ paragraph_end = 1
+ if ((len(insets) == 3 and insets[0] == "lyxline" and
+ insets[1] == "vspace" and insets[2] == "newpage") or
+ (len(insets) == 2 and
+ ((insets[0] == "lyxline" and insets[1] == "vspace") or
+ (insets[0] == "lyxline" and insets[1] == "newpage") or
+ (insets[0] == "vspace" and insets[1] == "newpage")))):
+ # This paragraph has been created after a paragraph by
+ # convert_breaks()
+ paragraph_start = 1
+ paragraph_end = 1
+ after = 1
+
+ # Determine space amount for vspace insets
+ spaceamount = list()
+ arguments = list()
+ for k in range(len(lines)):
+ if insets[k] == "vspace":
+ spaceamount.append(lines[k][2])
+ arguments.append(' ' + spaceamount[k] + ' ')
+ else:
+ spaceamount.append('')
+ arguments.append(' ')
- # Are we at the beginning or end of a paragraph?
- paragraph_start = 1
- start = get_paragraph(file.body, i) + 1
- for k in range(start, i):
- if is_nonempty_line(file.body[k]):
- paragraph_start = 0
- break
- paragraph_end = 1
- j = find_end_of_inset(file.body, i)
- if j == -1:
- file.warning("Malformed LyX file: Missing '\\end_inset'.")
- i = i + 1
- continue
- end = get_next_paragraph(file.body, i)
- for k in range(j + 1, end):
- if is_nonempty_line(file.body[k]):
- paragraph_end = 0
- break
-
- # Convert to paragraph formatting if we are at the beginning or end
- # of a paragraph and the resulting paragraph would not be empty
- if ((paragraph_start and not paragraph_end) or
- (paragraph_end and not paragraph_start)):
- # The order is important: del and insert invalidate some indices
- del file.body[j]
- del file.body[i]
- if paragraph_start:
- file.body.insert(start, '\\added_space_top ' + spaceamount + ' ')
- else:
- file.body.insert(start, '\\added_space_bottom ' + spaceamount + ' ')
- continue
+ if paragraph_start and paragraph_end:
+ # We are in a paragraph of our own.
+ # We must not delete this paragraph if it has parameters
+ if params == '':
+ # First try to merge with the previous paragraph.
+ # We try the previous paragraph first because we would
+ # otherwise need ERT for two subsequent vspaces.
+ prev_par = get_paragraph(file.body, start - 2) + 1
+ if prev_par > 0 and not before:
+ prev_params = get_par_params(file.body, prev_par + 1)
+ ert = 0
+ n = len(insets)
+ for k in range(n):
+ if keywords_bot[insets[k]] in prev_params:
+ ert = 1
+ break
+ if not ert:
+ for k in range(len(insets)):
+ file.body.insert(prev_par + 1,
+ keywords_bot[insets[k]] + arguments[k])
+ del file.body[start-1+n:end+n]
+ i = start + n
+ continue
+ # Then try next paragraph
+ next_par = end
+ if next_par > 0 and not after:
+ next_params = get_par_params(file.body, next_par + 1)
+ ert = 0
+ n = len(insets)
+ for k in range(n):
+ if keywords_top[insets[k]] in next_params:
+ ert = 1
+ break
+ if not ert:
+ for k in range(len(insets)):
+ file.body.insert(next_par + 1,
+ keywords_top[insets[k]] + arguments[k])
+ del file.body[start-1:end]
+ i = start
+ continue
+ else:
+ # Convert to paragraph formatting if we are at the beginning or end
+ # of a paragraph and the resulting paragraph would not be empty
+ # The order is important: del and insert invalidate some indices
+ if paragraph_start:
+ if j != i:
+ del file.body[j]
+ del file.body[i]
+ file.body.insert(start, keywords_top[insets[0]] + arguments[0])
+ continue
+ elif paragraph_end:
+ if j != i:
+ del file.body[j]
+ del file.body[i]
+ file.body.insert(start, keywords_bot[insets[0]] + arguments[0])
+ continue
- # Convert to ERT
- file.body[i:i+1] = ['\\begin_inset ERT', 'status Collapsed', '',
- '\\layout Standard', '', '\\backslash ']
- i = i + 6
- if spaceamount[-1] == '*':
- spaceamount = spaceamount[:-1]
- keep = 1
- else:
- keep = 0
+ # Convert the first inset to ERT.
+ # The others are converted in the next loop runs (if they exist)
+ if insets[0] == "vspace":
+ file.body[i:i+1] = ['\\begin_inset ERT', 'status Collapsed', '',
+ '\\layout Standard', '', '\\backslash ']
+ i = i + 6
+ if spaceamount[0][-1] == '*':
+ spaceamount[0] = spaceamount[0][:-1]
+ keep = 1
+ else:
+ keep = 0
- # Replace defskip by the actual value
- if spaceamount == 'defskip':
- spaceamount = defskipamount
-
- # LaTeX does not know \\smallskip* etc
- if keep:
- if spaceamount == 'smallskip':
- spaceamount = '\\smallskipamount'
- elif spaceamount == 'medskip':
- spaceamount = '\\medskipamount'
- elif spaceamount == 'bigskip':
- spaceamount = '\\bigskipamount'
- elif spaceamount == 'vfill':
- spaceamount = '\\fill'
-
- # Finally output the LaTeX code
- if (spaceamount == 'smallskip' or spaceamount == 'medskip' or
- spaceamount == 'bigskip' or spaceamount == 'vfill'):
- file.body.insert(i, spaceamount)
- else :
- if keep:
- file.body.insert(i, 'vspace*{')
- else:
- file.body.insert(i, 'vspace{')
- i = convert_ertbackslash(file.body, i, spaceamount)
- file.body[i] = file.body[i] + '}'
- i = i + 1
+ # Replace defskip by the actual value
+ if spaceamount[0] == 'defskip':
+ spaceamount[0] = defskipamount
+
+ # LaTeX does not know \\smallskip* etc
+ if keep:
+ if spaceamount[0] == 'smallskip':
+ spaceamount[0] = '\\smallskipamount'
+ elif spaceamount[0] == 'medskip':
+ spaceamount[0] = '\\medskipamount'
+ elif spaceamount[0] == 'bigskip':
+ spaceamount[0] = '\\bigskipamount'
+ elif spaceamount[0] == 'vfill':
+ spaceamount[0] = '\\fill'
+
+ # Finally output the LaTeX code
+ if (spaceamount[0] == 'smallskip' or spaceamount[0] == 'medskip' or
+ spaceamount[0] == 'bigskip' or spaceamount[0] == 'vfill'):
+ file.body.insert(i, spaceamount[0] + '{}')
+ else :
+ if keep:
+ file.body.insert(i, 'vspace*{')
+ else:
+ file.body.insert(i, 'vspace{')
+ i = convert_ertbackslash(file.body, i, spaceamount[0])
+ file.body[i] = file.body[i] + '}'
+ i = i + 1
+ elif insets[0] == "lyxline":
+ file.body[i:i+1] = ['\\begin_inset ERT', 'status Collapsed', '',
+ '\\layout Standard', '', '\\backslash',
+ 'lyxline{}', '', '\\end_inset', '']
+ i = i + 10
+ # We use \providecommand so that we don't get an error if native
+ # lyxlines are used (LyX writes first its own preamble and then
+ # the user specified one)
+ add_to_preamble(file,
+ ['% Commands inserted by lyx2lyx for lyxlines',
+ '\\providecommand{\\lyxline}[1]{',
+ ' {#1 \\vspace{1ex} \\hrule width \\columnwidth \\vspace{1ex}}'
+ '}'])
+ elif insets[0] == "newpage":
+ file.body[i:i+1] = ['\\begin_inset ERT', 'status Collapsed', '',
+ '\\layout Standard', '', '\\backslash',
+ 'newpage{}', '', '\\end_inset', '']
+ i = i + 10
# Convert a LyX length into a LaTeX length
