1 """SCons.Scanner.LaTeX
2
3 This module implements the dependency scanner for LaTeX code.
4
5 """
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30 __revision__ = "src/engine/SCons/Scanner/LaTeX.py 5134 2010/08/16 23:02:40 bdeegan"
31
32 import os.path
33 import re
34
35 import SCons.Scanner
36 import SCons.Util
37
38
39 TexGraphics = ['.eps', '.ps']
40 LatexGraphics = ['.pdf', '.png', '.jpg', '.gif', '.tif']
41
42
45 _null = _Null
46
47
48
49
50
51
53 try:
54 save = env['ENV'][var]
55 except KeyError:
56 save = _null
57 env.PrependENVPath(var, abspath)
58 try:
59 if SCons.Util.is_List(env[var]):
60 env.PrependENVPath(var, [os.path.abspath(str(p)) for p in env[var]])
61 else:
62
63 env.PrependENVPath(var, [os.path.abspath(p) for p in str(env[var]).split(os.pathsep)])
64 except KeyError:
65 pass
66
67
68
69
70
71 if SCons.Util.is_List(env['ENV'][var]):
72 env['ENV'][var] = os.pathsep.join(env['ENV'][var])
73
74 env['ENV'][var] = env['ENV'][var] + os.pathsep
75
76 return save
77
79 """A class to bind a specific *PATH variable name to a function that
80 will return all of the *path directories."""
82 self.variable = variable
83 - def __call__(self, env, dir=None, target=None, source=None, argument=None):
93
94
95
97 """Return a prototype Scanner instance for scanning LaTeX source files
98 when built with latex.
99 """
100 ds = LaTeX(name = "LaTeXScanner",
101 suffixes = '$LATEXSUFFIXES',
102
103 graphics_extensions = TexGraphics,
104 recursive = 0)
105 return ds
106
108 """Return a prototype Scanner instance for scanning LaTeX source files
109 when built with pdflatex.
110 """
111 ds = LaTeX(name = "PDFLaTeXScanner",
112 suffixes = '$LATEXSUFFIXES',
113
114 graphics_extensions = LatexGraphics,
115 recursive = 0)
116 return ds
117
118 -class LaTeX(SCons.Scanner.Base):
119 """Class for scanning LaTeX files for included files.
120
121 Unlike most scanners, which use regular expressions that just
122 return the included file name, this returns a tuple consisting
123 of the keyword for the inclusion ("include", "includegraphics",
124 "input", or "bibliography"), and then the file name itself.
125 Based on a quick look at LaTeX documentation, it seems that we
126 should append .tex suffix for the "include" keywords, append .tex if
127 there is no extension for the "input" keyword, and need to add .bib
128 for the "bibliography" keyword that does not accept extensions by itself.
129
130 Finally, if there is no extension for an "includegraphics" keyword
131 latex will append .ps or .eps to find the file, while pdftex may use .pdf,
132 .jpg, .tif, .mps, or .png.
133
134 The actual subset and search order may be altered by
135 DeclareGraphicsExtensions command. This complication is ignored.
136 The default order corresponds to experimentation with teTeX
137 $ latex --version
138 pdfeTeX 3.141592-1.21a-2.2 (Web2C 7.5.4)
139 kpathsea version 3.5.4
140 The order is:
141 ['.eps', '.ps'] for latex
142 ['.png', '.pdf', '.jpg', '.tif'].
143
144 Another difference is that the search path is determined by the type
145 of the file being searched:
146 env['TEXINPUTS'] for "input" and "include" keywords
147 env['TEXINPUTS'] for "includegraphics" keyword
148 env['TEXINPUTS'] for "lstinputlisting" keyword
149 env['BIBINPUTS'] for "bibliography" keyword
150 env['BSTINPUTS'] for "bibliographystyle" keyword
151
152 FIXME: also look for the class or style in document[class|style]{}
153 FIXME: also look for the argument of bibliographystyle{}
154 """
155 keyword_paths = {'include': 'TEXINPUTS',
156 'input': 'TEXINPUTS',
157 'includegraphics': 'TEXINPUTS',
158 'bibliography': 'BIBINPUTS',
159 'bibliographystyle': 'BSTINPUTS',
160 'usepackage': 'TEXINPUTS',
161 'lstinputlisting': 'TEXINPUTS'}
162 env_variables = SCons.Util.unique(list(keyword_paths.values()))
163
164 - def __init__(self, name, suffixes, graphics_extensions, *args, **kw):
165
166
167
168
169
170
171
172 regex = r'^[^%\n]*\\(include|includegraphics(?:\s*\[[^\]]+\])?|lstinputlisting(?:\[[^\]]+\])?|input|bibliography|usepackage)\s*{([^}]*)}'
173 self.cre = re.compile(regex, re.M)
174 self.comment_re = re.compile(r'^((?:(?:\\%)|[^%\n])*)(.*)$', re.M)
175
176 self.graphics_extensions = graphics_extensions
177
178 def _scan(node, env, path=(), self=self):
179 node = node.rfile()
180 if not node.exists():
181 return []
182 return self.scan_recurse(node, path)
183
184 class FindMultiPathDirs(object):
185 """The stock FindPathDirs function has the wrong granularity:
186 it is called once per target, while we need the path that depends
187 on what kind of included files is being searched. This wrapper
188 hides multiple instances of FindPathDirs, one per the LaTeX path
189 variable in the environment. When invoked, the function calculates
190 and returns all the required paths as a dictionary (converted into
191 a tuple to become hashable). Then the scan function converts it
192 back and uses a dictionary of tuples rather than a single tuple
193 of paths.
194 """
195 def __init__(self, dictionary):
196 self.dictionary = {}
197 for k,n in dictionary.items():
198 self.dictionary[k] = ( SCons.Scanner.FindPathDirs(n),
199 FindENVPathDirs(n) )
200
201 def __call__(self, env, dir=None, target=None, source=None,
202 argument=None):
203 di = {}
204 for k,(c,cENV) in self.dictionary.items():
205 di[k] = ( c(env, dir=None, target=None, source=None,
206 argument=None) ,
207 cENV(env, dir=None, target=None, source=None,
208 argument=None) )
209
210 return tuple(di.items())
211
212 class LaTeXScanCheck(object):
213 """Skip all but LaTeX source files, i.e., do not scan *.eps,
214 *.pdf, *.jpg, etc.
215 """
216 def __init__(self, suffixes):
217 self.suffixes = suffixes
218 def __call__(self, node, env):
219 current = not node.has_builder() or node.is_up_to_date()
220 scannable = node.get_suffix() in env.subst_list(self.suffixes)[0]
221
222 return scannable and current
223
224 kw['function'] = _scan
225 kw['path_function'] = FindMultiPathDirs(LaTeX.keyword_paths)
226 kw['recursive'] = 0
227 kw['skeys'] = suffixes
228 kw['scan_check'] = LaTeXScanCheck(suffixes)
229 kw['name'] = name
230
231 SCons.Scanner.Base.__init__(self, *args, **kw)
232
234 filename = include[1]
235 if include[0] == 'input':
236 base, ext = os.path.splitext( filename )
237 if ext == "":
238 return [filename + '.tex']
239 if (include[0] == 'include'):
240 return [filename + '.tex']
241 if include[0] == 'bibliography':
242 base, ext = os.path.splitext( filename )
243 if ext == "":
244 return [filename + '.bib']
245 if include[0] == 'usepackage':
246 base, ext = os.path.splitext( filename )
247 if ext == "":
248 return [filename + '.sty']
249 if include[0] == 'includegraphics':
250 base, ext = os.path.splitext( filename )
251 if ext == "":
252
253
254
255
256
257 return [filename+e for e in self.graphics_extensions]
258 return [filename]
259
262
264 try:
265 sub_path = path[include[0]]
266 except (IndexError, KeyError):
267 sub_path = ()
268 try_names = self._latex_names(include)
269 for n in try_names:
270
271 i = SCons.Node.FS.find_file(n, (source_dir,) + sub_path[0])
272 if i:
273 return i, include
274
275 i = SCons.Node.FS.find_file(n, (source_dir,) + sub_path[1])
276 if i:
277 return i, include
278 return i, include
279
280 - def canonical_text(self, text):
281 """Standardize an input TeX-file contents.
282
283 Currently:
284 * removes comments, unwrapping comment-wrapped lines.
285 """
286 out = []
287 line_continues_a_comment = False
288 for line in text.splitlines():
289 line,comment = self.comment_re.findall(line)[0]
290 if line_continues_a_comment == True:
291 out[-1] = out[-1] + line.lstrip()
292 else:
293 out.append(line)
294 line_continues_a_comment = len(comment) > 0
295 return '\n'.join(out).rstrip()+'\n'
296
297 - def scan(self, node):
298
299
300
301
302
303
304
305 noopt_cre = re.compile('\s*\[.*$')
306 if node.includes != None:
307 includes = node.includes
308 else:
309 text = self.canonical_text(node.get_text_contents())
310 includes = self.cre.findall(text)
311
312
313
314
315
316
317
318
319
320 split_includes = []
321 for include in includes:
322 inc_type = noopt_cre.sub('', include[0])
323 inc_list = include[1].split(',')
324 for j in range(len(inc_list)):
325 split_includes.append( (inc_type, inc_list[j]) )
326
327 includes = split_includes
328 node.includes = includes
329
330 return includes
331
333 """ do a recursive scan of the top level target file
334 This lets us search for included files based on the
335 directory of the main file just as latex does"""
336
337 path_dict = dict(list(path))
338
339 queue = []
340 queue.extend( self.scan(node) )
341 seen = {}
342
343
344
345
346
347
348
349
350 nodes = []
351 source_dir = node.get_dir()
352
353 while queue:
354
355 include = queue.pop()
356 try:
357 if seen[include[1]] == 1:
358 continue
359 except KeyError:
360 seen[include[1]] = 1
361
362
363
364
365 n, i = self.find_include(include, source_dir, path_dict)
366 if n is None:
367
368
369 if include[0] != 'usepackage':
370 SCons.Warnings.warn(SCons.Warnings.DependencyWarning,
371 "No dependency generated for file: %s (included from: %s) -- file not found" % (i, node))
372 else:
373 sortkey = self.sort_key(n)
374 nodes.append((sortkey, n))
375
376 queue.extend( self.scan(n) )
377
378 return [pair[1] for pair in sorted(nodes)]
379
380
381
382
383
384
385