#!/usr/bin/env python # encoding: utf8 # # wmllint -- check WML for conformance to the most recent dialect # # By Eric S. Raymond April 2007. # # PURPOSE # # wmllint is a tool for porting and debugging WML (Wesnoth Markup Language). # It can do much of the work of upgrading old content to a new version of # Battle for Wesnoth, automatically. # # While the script is at it, it checks for various incorrect and dodgy WML # constructs, including: # * unbalanced tags # * strings that need a translation mark and should not have them # * strings that have a translation mark and should not # * translatable strings containing macro references # * filter references by id= not matched by an actual unit # * abilities or traits without matching special notes, or vice-versa # * consistency between recruit= and recruitment_pattern= instances # * unknown unit types in recruitment lists # * double space after punctuation in translatable strings. # * unknown races or movement types in units # * unknown base units # * misspellings in message and description strings # # BASIC PROCEDURE # # Takes any number of directories as arguments. Each directory is converted. # If no directories are specified, acts on the current directory. # # The recommended procedure is this: # 1. Run it with --dryrun first to see what it will do. # 2. If the messages look good, run without --dryrun; the old content # will be left in backup files with a -bak extension. # 3. Eyeball the changes with the --diff option. # 4. Use wmlscope, with a directory list including the Wesnoth mainline WML # as first argument, to check that you have no unresolved references. # 5. Test the conversion. # 6. Use either --clean to remove the -bak files or --revert to # undo the conversion. # # Compatibility with pre-1.4 versions has been removed; to port very old # content, first run wmllint-1.4 to convert to "1.4", before running this # one. # # Standalone terrain mask files *must* have a .mask extension on their name # or they'll have an incorrect usage=map generated into them. # # MAGIC COMMENTS # # Note: You can shut wmllint up about custom terrains by having a comment # on the same line that includes the string "wmllint: ignore" or # "wmllint: noconvert". The same magic comments will also disable checking # of translation marks, which may alternatively be toggled for line blocks # with "wmllint: markcheck off" and "wmllint: markcheck on". # # You can also prevent description insertions with "wmllint: no-icon". # # You can force otherwise undeclared characters to be recognized with # a magic comment containing the string "wmllint: recognize". # The rest of the line is stripped and treated as the name of a character # who should be recognized in descriptions. This will be useful, # for example, if your scenario follows a continue so there are # characters present who were not explicitly recalled. It may # also be useful if you have wrapped unit-creation or recall markup in # non-core macros and wmllint cannot recognize it. # # If you use custom macros to create (or recall) a named unit, you can tell # wmllint which field contains the id with the string, "wmllint: whofield # ". After the macro's last use, *make sure* to remove it # with "wmllint: whofield clear "; if "clear" is not followed by a # macro name, all macros will be cleared from the list. (If the # string was not a non-zero number, wmllint will also attempt to remove the # specified macro.) # # For macros used in multiple scenarios to field characters, you can tell # wmllint to recognize them with another magic comment: # wmllint: who is # should be the macro name; would be the ids of the # recalled (or created) characters, comma-separated if there were more than # one. If more characters attach to a macro in later scenarios, they can be # appended with additional "who" magic comments (though this requires scenarios # to be numbered or otherwise sorted to work properly, as wmllint traverses # files in alphabetical order). If there is a character that drops out, # preceding that person's entry with a double-minus will cause that character # to be removed, e.g., "-- Garak". # # At the end of the campaign, *make sure* you use: # wmllint: unwho ALL # This will clear the list of "who" magic comments and tell wmllint to stop # checking for them in later files and directories. If a specific macro is # not used in subsequent scenarios, put "unwho " in its last scenario. # # Similarly, it is possible to explicitly declare a unit's usage class # with a magic comment that looks like this: # wmllint: usage of is # Note that must be a string wrapped in ASCII doublequotes. This # declaration will be useful if you are declaring units with macros that # include a substitutable formal in the unit name; there were examples in # UtBS, but they have since been converted to use the [base_unit] tag. # # If a mismatch between a recruit list and recruitment pattern involves a # usage type outside the five core types, the warning message will include a # note that a non-standard usage class is involved. If you are using custom # usage classes and would like wmllint to be aware of them, you can insert # the magic comment, "wmllint: usagetype ". This comment will take a # comma-separated list, and can also be pluralized to "usagetypes". # # You can disable stack-based malformation checks with a comment # containing "wmllint: validate-off" and re-enable with "wmllint: validate-on". # # You can prevent filename conversions with a comment containing # "wmllint: noconvert" on the same line as the filename. # # You can suppress complaints about files without an initial textdomain line # by embedding the magic comment "# wmllint: no translatables" in the file. # of course, it's a good idea to be sure this assertion is actually true. # # You can skip checks on unbalanced WML (e.g. in a macro definition) by # bracketing it with "wmllint: unbalanced-on" and "wmllint: unbalanced-off". # Note that this will also disable stack-based validation on the span # of lines they enclose. # # You can suppress warnings about newlines in messages (and attempts to # repair them) with "wmllint: display on", and re-enable them with # "wmllint: display off". The repair attempts (only) may also be # suppressed with the --stringfreeze option. # # A special comment "# wmllint: notecheck off" will disable checking unit types # for consistency between abilities/weapon specials and usage of special notes # macros in their descriptions. # The comment "# wmllint: notecheck on" will re-enable this check. # # A special comment "# wmllint: deathcheck off" will disable the check whether # dying units speak in their death events. # The comment "# wmllint: deathcheck on" will re-enable this check. # # A magic comment of the form "wmllint: general spellings word1 # word2..." will declare the tokens word1, word2, etc. to be # acceptable spellings for anywhere in the Wesnoth tree that the # spellchecker should never flag. If the keyword "general" is # replaced by "local", the spelling exceptions apply only in the # current file. If the keyword "general" is replaced by "directory", # the spelling exceptions apply to all files below the parent # directory. # # A comment containing "no spellcheck" disables spellchecking on the # line where it occurs. # # A comment of the form # # #wmllint: match {ABILITY_FOO} with {SPECIAL_NOTES_IOO} # # will declare an ability macro and a special-notes macro to be tied # together for reference-checking purposes. # # In 1.11.5 and 1.11.6 respectively, the special ellipses for leaders and units # without a ZoC are automatically handled by the C++ engine. wmllint warns if such # ellipses are found in your WML and asks to remove them; at the same time it # notifies of custom ellipses that may need to be updated (updating means that # -nozoc, -leader and -leader-nozoc variations of your ellipse need to be added). # You can disable this sanity check for the current line with the comment # "# wmllint: no ellipsecheck". # # DEVELOPER INFORMATION # # All conversion logic for lifting WML and maps from older versions of the # markup to newer ones should live here. This includes resource path changes # and renames, also map format conversions. # # Note: Lift logic for pre-1.4 versions has been removed; if you need it, # use wmllint-1.4 to lift before running this one. I did this for a policy # reason; I wanted to kill off the --oldversion switch. It will *not* # be restored; in future, changes to WML syntax *must* be forward # compatible in such a way that tags from old versions can be # unambiguously recognized (this will save everybody heartburn). As a # virtuous side effect, this featurectomy cuts wmllint's code # complexity by over 50%, improves performance by about 33%, and # banishes some annoying behaviors related to the 1.2 map-conversion # code. # import sys, os, re, getopt, string, copy, difflib, time, gzip from wesnoth.wmltools import * from wesnoth.wmliterator import * # Changes meant to be done on maps and .cfg lines. mapchanges = ( ("^Voha", "^Voa"), ("^Voh", "^Vo"), ("^Vhms", "^Vhha"), ("^Vhm", "^Vhh"), ("^Vcha", "^Vca"), ("^Vch", "^Vc"), ("^Vcm", "^Vc"), ("Ggf,", "Gg^Emf"), ("Qv,", "Mv"), ) # Base terrain aliasing changes. aliaschanges = ( # 1.11.8: ("Ch", "Ct"), ("Ds", "Dt"), ("Hh", "Ht"), ("Mm", "Mt"), ("Ss", "St"), ("Uu", "Ut"), ("Ww", "Wst"), ("Wo", "Wdt"), ("Wwr", "Wrt"), ("^Uf", "Uft"), # Vi -> Vit in 1.11.8, Vit -> Vt in 1.11.9. ("Vit", "Vt"), # 1.11.9: ("Vi", "Vt"), ) # Global changes meant to be done on all lines. Suppressed by noconvert. linechanges = ( ("canrecruit=1", "canrecruit=yes"), ("canrecruit=0", "canrecruit=no"), # Fix a common typo ("agression=", "aggression="), # These changed just after 1.5.0 ("[special_filter]", "[filter_attack]"), ("[wml_filter]", "[filter_wml]"), ("[unit_filter]", "[filter]"), ("[secondary_unit_filter]", "[filter_second]"), ("[attack_filter]", "[filter_attack]"), ("[secondary_attack_filter]", "[filter_second_attack]"), ("[special_filter_second]", "[filter_second_attack]"), ("[/special_filter]", "[/filter_attack]"), ("[/wml_filter]", "[/filter_wml]"), ("[/unit_filter]", "[/filter]"), ("[/secondary_unit_filter]", "[/filter_second]"), ("[/attack_filter]", "[/filter_attack]"), ("[/secondary_attack_filter]", "[/filter_second_attack]"), ("[/special_filter_second]", "[/filter_second_attack]"), ("grassland=", "flat="), ("tundra=", "frozen="), ("cavewall=", "impassable="), ("canyon=", "unwalkable="), # This changed after 1.5.2 ("advanceto=", "advances_to="), # This changed after 1.5.5, to enable mechanical spellchecking ("sabre", "saber"), ("nr-sad.ogg", "sad.ogg"), # Changed after 1.5.7 ("[debug_message]", "[wml_message]"), ("[/debug_message]", "[/wml_message]"), # Changed just before 1.5.9 ("portraits/Alex_Jarocha-Ernst/drake-burner.png", "portraits/drakes/burner.png"), ("portraits/Alex_Jarocha-Ernst/drake-clasher.png", "portraits/drakes/clasher.png"), ("portraits/Alex_Jarocha-Ernst/drake-fighter.png", "portraits/drakes/fighter.png"), ("portraits/Alex_Jarocha-Ernst/drake-glider.png", "portraits/drakes/glider.png"), ("portraits/Alex_Jarocha-Ernst/ghoul.png", "portraits/undead/ghoul.png"), ("portraits/Alex_Jarocha-Ernst/mermaid-initiate.png", "portraits/merfolk/initiate.png"), ("portraits/Alex_Jarocha-Ernst/merman-fighter.png", "portraits/merfolk/fighter.png"), ("portraits/Alex_Jarocha-Ernst/merman-hunter.png", "portraits/merfolk/hunter.png"), ("portraits/Alex_Jarocha-Ernst/naga-fighter.png", "portraits/nagas/fighter.png"), ("portraits/Alex_Jarocha-Ernst/nagini-fighter.png", "portraits/nagas/fighter+female.png"), ("portraits/Alex_Jarocha-Ernst/orcish-assassin.png", "portraits/orcs/assassin.png"), ("portraits/Emilien_Rotival/human-general.png", "portraits/humans/general.png"), ("portraits/Emilien_Rotival/human-heavyinfantry.png", "portraits/humans/heavy-infantry.png"), ("portraits/Emilien_Rotival/human-ironmauler.png", "portraits/humans/iron-mauler.png"), ("portraits/Emilien_Rotival/human-lieutenant.png", "portraits/humans/lieutenant.png"), ("portraits/Emilien_Rotival/human-marshal.png", "portraits/humans/marshal.png"), ("portraits/Emilien_Rotival/human-peasant.png", "portraits/humans/peasant.png"), ("portraits/Emilien_Rotival/human-pikeman.png", "portraits/humans/pikeman.png"), ("portraits/Emilien_Rotival/human-royalguard.png", "portraits/humans/royal-guard.png"), ("portraits/Emilien_Rotival/human-sergeant.png", "portraits/humans/sergeant.png"), ("portraits/Emilien_Rotival/human-spearman.png", "portraits/humans/spearman.png"), ("portraits/Emilien_Rotival/human-swordsman.png", "portraits/humans/swordsman.png"), ("portraits/Emilien_Rotival/transparent/human-general.png", "portraits/humans/transparent/general.png"), ("portraits/Emilien_Rotival/transparent/human-heavyinfantry.png", "portraits/humans/transparent/heavy-infantry.png"), ("portraits/Emilien_Rotival/transparent/human-ironmauler.png", "portraits/humans/transparent/iron-mauler.png"), ("portraits/Emilien_Rotival/transparent/human-lieutenant.png", "portraits/humans/transparent/lieutenant.png"), ("portraits/Emilien_Rotival/transparent/human-marshal.png", "portraits/humans/transparent/marshal.png"), ("portraits/Emilien_Rotival/transparent/human-marshal-2.png", "portraits/humans/transparent/marshal-2.png"), ("portraits/Emilien_Rotival/transparent/human-peasant.png", "portraits/humans/transparent/peasant.png"), ("portraits/Emilien_Rotival/transparent/human-pikeman.png", "portraits/humans/transparent/pikeman.png"), ("portraits/Emilien_Rotival/transparent/human-royalguard.png", "portraits/humans/transparent/royal-guard.png"), ("portraits/Emilien_Rotival/transparent/human-sergeant.png", "portraits/humans/transparent/sergeant.png"), ("portraits/Emilien_Rotival/transparent/human-spearman.png", "portraits/humans/transparent/spearman.png"), ("portraits/Emilien_Rotival/transparent/human-swordsman.png", "portraits/humans/transparent/swordsman.png"), ("portraits/James_Woo/assassin.png", "portraits/humans/assassin.png"), ("portraits/James_Woo/dwarf-guard.png", "portraits/dwarves/guard.png"), ("portraits/James_Woo/orc-warlord.png", "portraits/orcs/warlord.png"), ("portraits/James_Woo/orc-warlord2.png", "portraits/orcs/warlord2.png"), ("portraits/James_Woo/orc-warlord3.png", "portraits/orcs/warlord3.png"), ("portraits/James_Woo/orc-warlord4.png", "portraits/orcs/warlord4.png"), ("portraits/James_Woo/orc-warlord5.png", "portraits/orcs/warlord5.png"), ("portraits/James_Woo/troll.png", "portraits/trolls/troll.png"), ("portraits/Jason_Lutes/human-bandit.png", "portraits/humans/bandit.png"), ("portraits/Jason_Lutes/human-grand-knight.png", "portraits/humans/grand-knight.png"), ("portraits/Jason_Lutes/human-halberdier.png", "portraits/humans/halberdier.png"), ("portraits/Jason_Lutes/human-highwayman.png", "portraits/humans/highwayman.png"), ("portraits/Jason_Lutes/human-horseman.png", "portraits/humans/horseman.png"), ("portraits/Jason_Lutes/human-javelineer.png", "portraits/humans/javelineer.png"), ("portraits/Jason_Lutes/human-knight.png", "portraits/humans/knight.png"), ("portraits/Jason_Lutes/human-lancer.png", "portraits/humans/lancer.png"), ("portraits/Jason_Lutes/human-paladin.png", "portraits/humans/paladin.png"), ("portraits/Jason_Lutes/human-thug.png", "portraits/humans/thug.png"), ("portraits/Kitty/elvish-archer.png", "portraits/elves/archer.png"), ("portraits/Kitty/elvish-archer+female.png", "portraits/elves/archer+female.png"), ("portraits/Kitty/elvish-captain.png", "portraits/elves/captain.png"), ("portraits/Kitty/elvish-druid.png", "portraits/elves/druid.png"), ("portraits/Kitty/elvish-fighter.png", "portraits/elves/fighter.png"), ("portraits/Kitty/elvish-hero.png", "portraits/elves/hero.png"), ("portraits/Kitty/elvish-high-lord.png", "portraits/elves/high-lord.png"), ("portraits/Kitty/elvish-lady.png", "portraits/elves/lady.png"), ("portraits/Kitty/elvish-lord.png", "portraits/elves/lord.png"), ("portraits/Kitty/elvish-marksman.png", "portraits/elves/marksman.png"), ("portraits/Kitty/elvish-marksman+female.png", "portraits/elves/marksman+female.png"), ("portraits/Kitty/elvish-ranger.png", "portraits/elves/ranger.png"), ("portraits/Kitty/elvish-ranger+female.png", "portraits/elves/ranger+female.png"), ("portraits/Kitty/elvish-scout.png", "portraits/elves/scout.png"), ("portraits/Kitty/elvish-shaman.png", "portraits/elves/shaman.png"), ("portraits/Kitty/elvish-shyde.png", "portraits/elves/shyde.png"), ("portraits/Kitty/elvish-sorceress.png", "portraits/elves/sorceress.png"), ("portraits/Kitty/human-dark-adept.png", "portraits/humans/dark-adept.png"), ("portraits/Kitty/human-dark-adept+female.png", "portraits/humans/dark-adept+female.png"), ("portraits/Kitty/human-mage.png", "portraits/humans/mage.png"), ("portraits/Kitty/human-mage+female.png", "portraits/humans/mage+female.png"), ("portraits/Kitty/human-mage-arch.png", "portraits/humans/mage-arch.png"), ("portraits/Kitty/human-mage-arch+female.png", "portraits/humans/mage-arch+female.png"), ("portraits/Kitty/human-mage-light.png", "portraits/humans/mage-light.png"), ("portraits/Kitty/human-mage-light+female.png", "portraits/humans/mage-light+female.png"), ("portraits/Kitty/human-mage-red.png", "portraits/humans/mage-red.png"), ("portraits/Kitty/human-mage-red+female.png", "portraits/humans/mage-red+female.png"), ("portraits/Kitty/human-mage-silver.png", "portraits/humans/mage-silver.png"), ("portraits/Kitty/human-mage-silver+female.png", "portraits/humans/mage-silver+female.png"), ("portraits/Kitty/human-mage-white.png", "portraits/humans/mage-white.png"), ("portraits/Kitty/human-mage-white+female.png", "portraits/humans/mage-white+female.png"), ("portraits/Kitty/human-necromancer.png", "portraits/humans/necromancer.png"), ("portraits/Kitty/human-necromancer+female.png", "portraits/humans/necromancer+female.png"), ("portraits/Kitty/troll-whelp.png", "portraits/trolls/whelp.png"), ("portraits/Kitty/undead-lich.png", "portraits/undead/lich.png"), ("portraits/Kitty/transparent/elvish-archer.png", "portraits/elves/transparent/archer.png"), ("portraits/Kitty/transparent/elvish-archer+female.png", "portraits/elves/transparent/archer+female.png"), ("portraits/Kitty/transparent/elvish-captain.png", "portraits/elves/transparent/captain.png"), ("portraits/Kitty/transparent/elvish-druid.png", "portraits/elves/transparent/druid.png"), ("portraits/Kitty/transparent/elvish-fighter.png", "portraits/elves/transparent/fighter.png"), ("portraits/Kitty/transparent/elvish-hero.png", "portraits/elves/transparent/hero.png"), ("portraits/Kitty/transparent/elvish-high-lord.png", "portraits/elves/transparent/high-lord.png"), ("portraits/Kitty/transparent/elvish-lady.png", "portraits/elves/transparent/lady.png"), ("portraits/Kitty/transparent/elvish-lord.png", "portraits/elves/transparent/lord.png"), ("portraits/Kitty/transparent/elvish-marksman.png", "portraits/elves/transparent/marksman.png"), ("portraits/Kitty/transparent/elvish-marksman+female.png", "portraits/elves/transparent/marksman+female.png"), ("portraits/Kitty/transparent/elvish-ranger.png", "portraits/elves/transparent/ranger.png"), ("portraits/Kitty/transparent/elvish-ranger+female.png", "portraits/elves/transparent/ranger+female.png"), ("portraits/Kitty/transparent/elvish-scout.png", "portraits/elves/transparent/scout.png"), ("portraits/Kitty/transparent/elvish-shaman.png", "portraits/elves/transparent/shaman.png"), ("portraits/Kitty/transparent/elvish-shyde.png", "portraits/elves/transparent/shyde.png"), ("portraits/Kitty/transparent/elvish-sorceress.png", "portraits/elves/transparent/sorceress.png"), ("portraits/Kitty/transparent/human-dark-adept.png", "portraits/humans/transparent/dark-adept.png"), ("portraits/Kitty/transparent/human-dark-adept+female.png", "portraits/humans/transparent/dark-adept+female.png"), ("portraits/Kitty/transparent/human-mage.png", "portraits/humans/transparent/mage.png"), ("portraits/Kitty/transparent/human-mage+female.png", "portraits/humans/transparent/mage+female.png"), ("portraits/Kitty/transparent/human-mage-arch.png", "portraits/humans/transparent/mage-arch.png"), ("portraits/Kitty/transparent/human-mage-arch+female.png", "portraits/humans/transparent/mage-arch+female.png"), ("portraits/Kitty/transparent/human-mage-light.png", "portraits/humans/transparent/mage-light.png"), ("portraits/Kitty/transparent/human-mage-light+female.png", "portraits/humans/transparent/mage-light+female.png"), ("portraits/Kitty/transparent/human-mage-red.png", "portraits/humans/transparent/mage-red.png"), ("portraits/Kitty/transparent/human-mage-red+female.png", "portraits/humans/transparent/mage-red+female.png"), ("portraits/Kitty/transparent/human-mage-silver.png", "portraits/humans/transparent/mage-silver.png"), ("portraits/Kitty/transparent/human-mage-silver+female.png", "portraits/humans/transparent/mage-silver+female.png"), ("portraits/Kitty/transparent/human-mage-white.png", "portraits/humans/transparent/mage-white.png"), ("portraits/Kitty/transparent/human-mage-white+female.png", "portraits/humans/transparent/mage-white+female.png"), ("portraits/Kitty/transparent/human-necromancer.png", "portraits/humans/transparent/necromancer.png"), ("portraits/Kitty/transparent/human-necromancer+female.png", "portraits/humans/transparent/necromancer+female.png"), ("portraits/Kitty/transparent/troll-whelp.png", "portraits/trolls/transparent/whelp.png"), ("portraits/Kitty/transparent/undead-lich.png", "portraits/undead/transparent/lich.png"), ("portraits/Nicholas_Kerpan/human-poacher.png", "portraits/humans/poacher.png"), ("portraits/Nicholas_Kerpan/human-thief.png", "portraits/humans/thief.png"), ("portraits/Other/brown-lich.png", "portraits/undead/brown-lich.png"), ("portraits/Other/cavalryman.png", "portraits/humans/cavalryman.png"), ("portraits/Other/human-masterbowman.png", "portraits/humans/master-bowman.png"), ("portraits/Other/scorpion.png", "portraits/monsters/scorpion.png"), ("portraits/Other/sea-serpent.png", "portraits/monsters/sea-serpent.png"), ("portraits/Pekka_Aikio/human-bowman.png", "portraits/humans/bowman.png"), ("portraits/Pekka_Aikio/human-longbowman.png", "portraits/humans/longbowman.png"), ("portraits/Philip_Barber/dwarf-dragonguard.png", "portraits/dwarves/dragonguard.png"), ("portraits/Philip_Barber/dwarf-fighter.png", "portraits/dwarves/fighter.png"), ("portraits/Philip_Barber/dwarf-lord.png", "portraits/dwarves/lord.png"), ("portraits/Philip_Barber/dwarf-thunderer.png", "portraits/dwarves/thunderer.png"), ("portraits/Philip_Barber/saurian-augur.png", "portraits/saurians/augur.png"), ("portraits/Philip_Barber/saurian-skirmisher.png", "portraits/saurians/skirmisher.png"), ("portraits/Philip_Barber/undead-death-knight.png", "portraits/undead/death-knight.png"), ("portraits/Philip_Barber/transparent/dwarf-dragonguard.png", "portraits/dwarves/transparent/dragonguard.png"), ("portraits/Philip_Barber/transparent/dwarf-fighter.png", "portraits/dwarves/transparent/fighter.png"), ("portraits/Philip_Barber/transparent/dwarf-lord.png", "portraits/dwarves/transparent/lord.png"), ("portraits/Philip_Barber/transparent/dwarf-thunderer.png", "portraits/dwarves/transparent/thunderer.png"), ("portraits/Philip_Barber/transparent/saurian-augur.png", "portraits/saurians/transparent/augur.png"), ("portraits/Philip_Barber/transparent/saurian-skirmisher.png", "portraits/saurians/transparent/skirmisher.png"), ("portraits/Philip_Barber/transparent/undead-death-knight.png", "portraits/undead/transparent/death-knight.png"), # Changed just before 1.5.11 ("titlescreen/landscapebattlefield.jpg", "story/landscape-battlefield.jpg"), ("titlescreen/landscapebridge.jpg", "story/landscape-bridge.jpg"), ("titlescreen/landscapecastle.jpg", "story/landscape-castle.jpg"), ("LABEL_PERSISTANT", "LABEL_PERSISTENT"), # Changed just before 1.5.13 ("targetting", "targeting"), # Changed just after 1.7 fork ("[stone]", "[petrify]"), ("[unstone]", "[unpetrify]"), ("[/stone]", "[/petrify]"), ("[/unstone]", "[/unpetrify]"), ("WEAPON_SPECIAL_STONE", "WEAPON_SPECIAL_PETRIFY"), ("SPECIAL_NOTE_STONE", "SPECIAL_NOTE_PETRIFY"), (".stoned", ".petrified"), ("stoned=", "petrified="), # Changed at rev 37390 ("swing=", "value_second="), # Changed just before 1.7.3 ("Drake Gladiator", "Drake Thrasher"), ("gladiator-", "thrasher-"), ("Drake Slasher", "Drake Arbiter"), ("slasher-", "arbiter-"), # Changes after 1.7.5 ("portraits/nagas/fighter+female.png", "portraits/nagas/fighter.png"), # Changes after 1.8rc1 ("portraits/orcs/warlord.png", "portraits/orcs/transparent/warlord.png"), #("portraits/orcs/warlord2.png","portraits/orcs/transparent/warlord.png"), // see 1.11.5 ("portraits/orcs/warlord3.png","portraits/orcs/transparent/grunt-2.png"), #("portraits/orcs/warlord4.png","portraits/orcs/transparent/grunt-2.png"), // see 1.11.5 ("portraits/orcs/warlord5.png","portraits/orcs/transparent/grunt-3.png"), # Changes just before 1.9.0 ("flat/grass-r8", "flat/grass6"), ("flat/grass-r7", "flat/grass5"), ("flat/grass-r6", "flat/grass6"), ("flat/grass-r5", "flat/grass5"), ("flat/grass-r4", "flat/grass4"), ("flat/grass-r3", "flat/grass3"), ("flat/grass-r2", "flat/grass2"), ("flat/grass-r1", "flat/grass1"), ("second_value=", "value_second="), # Correct earlier wmllint error (".stones", ".petrifies"), ("stones=", "petrifies="), # Changes just before 1.9.1 ("[colour_adjust]", "[color_adjust]"), ("[/colour_adjust]", "[/color_adjust]"), ("colour=", "color="), ("colour_lock=", "color_lock="), # Changes just before 1.9.2 ("[removeitem]", "[remove_item]"), ("[/removeitem]", "[/remove_item]"), # Changes just before 1.11.0 ("viewing_side", "side"), ("duration=level", "duration=scenario"), # Note: this may be removed after 1.11.2, so an actual duration=level can be implemented # Changed before 1.11.5 to incorporate 1.9.0 portraits ("portraits/orcs/warlord2.png","portraits/orcs/transparent/grunt-5.png"), ("portraits/orcs/warlord4.png","portraits/orcs/transparent/grunt-6.png"), # Changed before 1.11.8 ("misc/schedule-dawn.png","misc/time-schedules/default/schedule-dawn.png"), ("misc/schedule-morning.png","misc/time-schedules/default/schedule-morning.png"), ("misc/schedule-afternoon.png","misc/time-schedules/default/schedule-afternoon.png"), ("misc/schedule-dusk.png","misc/time-schedules/default/schedule-dusk.png"), ("misc/schedule-firstwatch.png","misc/time-schedules/default/schedule-firstwatch.png"), ("misc/schedule-secondwatch.png","misc/time-schedules/default/schedule-secondwatch.png"), ("misc/schedule-indoors.png","misc/time-schedules/schedule-indoors.png"), ("misc/schedule-underground.png","misc/time-schedules/schedule-underground.png"), ("misc/schedule-underground-illum.png","misc/time-schedules/schedule-underground-illum.png"), ("misc/tod-schedule-24hrs.png","misc/time-schedules/tod-schedule-24hrs.png") ) def validate_on_pop(tagstack, closer, filename, lineno): "Validate the stack at the time a new close tag is seen." (tag, attributes) = tagstack[-1] ancestors = [x[0] for x in tagstack] if verbose >= 3: print '"%s", line %d: closing %s I see %s with %s' % (filename, lineno, closer, tag, attributes) # Detect a malformation that will cause the game to barf while attempting # to deserialize an empty unit. The final "and attributes" is a blatant # hack; some campaigns like to generate entire side declarations with # macros. if "scenario" in ancestors and closer == "side" and "type" not in attributes and ("no_leader" not in attributes or attributes["no_leader"] != "yes") and "multiplayer" not in ancestors and attributes: print '"%s", line %d: [side] without type attribute' % (filename, lineno) # This assumes that conversion will always happen in units/ files. if "units" not in filename and closer == "unit" and "race" in attributes: print '"%s", line %d: [unit] needs hand fixup to [unit_type]' % \ (filename, lineno) if closer in ["campaign", "race"] and "id" not in attributes: print '"%s", line %d: %s requires an ID attribute but has none' % \ (filename, lineno, closer) if closer == "terrain" and attributes.get("heals") in ("true", "false"): print '"%s", line %d: heals attribute no longer takes a boolean' % \ (filename, lineno) if closer == "unit" and attributes.get("id") is not None and attributes.get("type") is not None and attributes.get("side") is None and not "side" in ancestors: print '"%s", line %d: unit declaration without side attribute' % \ (filename, lineno) if closer == "theme" and "id" not in attributes: if "name" in attributes: print '"%s", line %d: using [theme]name= instead of [theme]id= is deprecated' % (filename, lineno) else: print '"%s", line %d: [theme] needs an id attribute' % (filename, lineno) # Check for user-visible themes that lack a UI name or description. if closer == "theme" and ("hidden" not in attributes or attributes["hidden"] not in ("yes", "true")): for attr in ("name", "description"): if attr not in attributes: print '"%s", line %d: [theme] needs a %s attribute unless hidden=yes' % \ (filename, lineno, attr) if closer == "filter_side": ancestor = False if "gold" in ancestors: ancestor = "gold" elif "modify_ai" in ancestors: ancestor = "modify_ai" if ancestor: print '"%s", line %d: %s should have an inline SSF instead of using [filter_side]' % \ (filename, lineno, ancestor) if closer == "effect": if attributes.get("unit_type") is not None: print '"%s", line %d: use [effect][filter]type= instead of [effect]unit_type=' % \ (filename, lineno) if attributes.get("unit_gender") is not None: print '"%s", line %d: use [effect][filter]gender= instead of [effect]unit_gender=' % \ (filename, lineno) if missingside and closer in ["set_recruit", "allow_recruit", "disallow_recruit", "store_gold"] and "side" not in attributes: print '"%s", line %d: %s without "side" attribute is now applied to all sides' % \ (filename, lineno, closer) def within(tag): "Did the specified tag lead one of our enclosing contexts?" if type(tag) == type(()): # Can take a list. for t in tag: if within(t): return True else: return False else: return tag in [x[0] for x in tagstack] def under(tag): "Did the specified tag lead the latest context?" if type(tag) == type(()): # Can take a list. for t in tag: if within(t): return True else: return False elif tagstack: return tag == tagstack[-1][0] else: return False def standard_unit_filter(): "Are we within the syntactic context of a standard unit filter?" # It's under("message") rather than within("message") because # [message] can contain [option] markup with menu item description= # attributes that should not be altered. return within(("filter", "filter_second", "filter_adjacent", "filter_opponent", "unit_filter", "secondary_unit_filter", "special_filter", "special_filter_second", "neighbor_unit_filter", "recall", "teleport", "kill", "unstone", "store_unit", "have_unit", "scroll_to_unit", "role", "hide_unit", "unhide_unit", "protect_unit", "target", "avoid")) \ or under("message") # Sanity checking # Associations for the ability sanity checks. notepairs = [ ("movement_type=undeadspirit", "{SPECIAL_NOTES_SPIRIT}"), ("type=arcane", "{SPECIAL_NOTES_ARCANE}"), ("{ABILITY_HEALS}", "{SPECIAL_NOTES_HEALS}"), ("{ABILITY_EXTRA_HEAL}", "{SPECIAL_NOTES_EXTRA_HEAL}"), ("{ABILITY_UNPOISON}", "{SPECIAL_NOTES_UNPOISON}"), ("{ABILITY_CURES}", "{SPECIAL_NOTES_CURES}"), ("{ABILITY_REGENERATES}", "{SPECIAL_NOTES_REGENERATES}"), ("{ABILITY_STEADFAST}", "{SPECIAL_NOTES_STEADFAST}"), ("{ABILITY_LEADERSHIP_LEVEL_", "{SPECIAL_NOTES_LEADERSHIP}"), # No } deliberately ("{ABILITY_SKIRMISHER}", "{SPECIAL_NOTES_SKIRMISHER}"), ("{ABILITY_ILLUMINATES}", "{SPECIAL_NOTES_ILLUMINATES}"), ("{ABILITY_TELEPORT}", "{SPECIAL_NOTES_TELEPORT}"), ("{ABILITY_AMBUSH}", "{SPECIAL_NOTES_AMBUSH}"), ("{ABILITY_NIGHTSTALK}", "{SPECIAL_NOTES_NIGHTSTALK}"), ("{ABILITY_CONCEALMENT}", "{SPECIAL_NOTES_CONCEALMENT}"), ("{ABILITY_SUBMERGE}", "{SPECIAL_NOTES_SUBMERGE}"), ("{ABILITY_FEEDING}", "{SPECIAL_NOTES_FEEDING}"), ("{WEAPON_SPECIAL_BERSERK}", "{SPECIAL_NOTES_BERSERK}"), ("{WEAPON_SPECIAL_BACKSTAB}", "{SPECIAL_NOTES_BACKSTAB}"), ("{WEAPON_SPECIAL_PLAGUE", "{SPECIAL_NOTES_PLAGUE}"), # No } deliberately ("{WEAPON_SPECIAL_SLOW}", "{SPECIAL_NOTES_SLOW}"), ("{WEAPON_SPECIAL_PETRIFY}", "{SPECIAL_NOTES_PETRIFY}"), ("{WEAPON_SPECIAL_MARKSMAN}", "{SPECIAL_NOTES_MARKSMAN}"), ("{WEAPON_SPECIAL_MAGICAL}", "{SPECIAL_NOTES_MAGICAL}"), ("{WEAPON_SPECIAL_SWARM}", "{SPECIAL_NOTES_SWARM}"), ("{WEAPON_SPECIAL_CHARGE}", "{SPECIAL_NOTES_CHARGE}"), ("{WEAPON_SPECIAL_DRAIN}", "{SPECIAL_NOTES_DRAIN}"), ("{WEAPON_SPECIAL_FIRSTSTRIKE}", "{SPECIAL_NOTES_FIRSTSTRIKE}"), ("{WEAPON_SPECIAL_POISON}", "{SPECIAL_NOTES_POISON}"), ("{WEAPON_SPECIAL_STUN}", "{SPECIAL_NOTES_STUN}"), ] # This dictionary will pair macros with the characters they recall or create, # but must be populated by the magic comment, "#wmllint: who ... is ...". whopairs = {} # This dictionary pairs macros with the id field of the characters they recall # or create, and is populated by the comment, "wmllint: whofield <#>." whomacros = {} # This dictionary pairs the ids of stored units with their variable name. storedids = {} # This list of the standard recruitable usage types can be appended with the # magic comment, "#wmllint: usagetype[s]". usage_types = ["scout", "fighter", "mixed fighter", "archer", "healer"] # These are accumulated by sanity_check() and examined by consistency_check() unit_types = [] derived_units = [] usage = {} sides = [] advances = [] movetypes = [] unit_movetypes = [] races = [] unit_races = [] nextrefs = [] scenario_to_filename = {} # Attributes that should have translation marks translatables = re.compile( \ "^abbrev$|" \ "^cannot_use_message$|" \ "^caption$|" \ "^current_player$|" \ "^currently_doing_description$|" \ "^description$|" \ "^description_inactive$|" \ "^editor_name$|" \ "^end_text$|" \ "^difficulty_descriptions$|" \ "^female_name_inactive$|" \ "^female_names$|" \ "^help_topic_text$|" \ "^label$|" \ "^male_names$|" \ "^message$|" \ "^name$|" \ "^name_inactive$|" \ "^new_game_title$|" \ "^note$|" \ "^option_description$|" \ "^option_name$|" \ "^order$|" \ "^plural_name$|" \ "^prefix$|" \ "^set_description$|" \ "^source$|" \ "^story$|" \ "^summary$|" \ "^victory_string$|" \ "^defeat_string$|" \ "^gold_carryover_string$|" \ "^notes_string$|" \ "^text$|" \ "^title$|" \ "^title2$|" \ "^tooltip$|" \ "^translator_comment$|" \ "^user_team_name$|" \ "^type_.[a-z]*$|" \ "^range_[a-z]*$") # This is a list of mainline campaigns, used to convert UMC from # "data/campaigns" to "data/add-ons" while not clobbering mainline. mainline = ("An_Orcish_Incursion", "Dead_Water", "Delfadors_Memoirs", "Descent_Into_Darkness", "Eastern_Invasion", "Heir_To_The_Throne", "Legend_of_Wesmere", "Liberty", "Northern_Rebirth", "Sceptre_of_Fire", "Son_Of_The_Black_Eye", "The_Hammer_of_Thursagan", "The_Rise_Of_Wesnoth", "The_South_Guard", "tutorial", "Two_Brothers", "Under_the_Burning_Suns", ) spellcheck_these = (\ "cannot_use_message=", "caption=", "description=", "description_inactive=", "editor_name=", "end_text=", "help_topic_text=", "message=", "note=", "story=", "summary=", "text=", "title=", "title2=", "tooltip=", "user_team_name=", ) # Declare a few common English contractions and ejaculations that pyenchant # inexplicably knows nothing of. declared_spellings = {"GLOBAL":["I'm", "I've", "I'd", "I'll", "heh", "ack", # Fantasy/SF/occult jargon that we need "aerie", "aeon", "aide-de-camp", "axe", "ballista", "bided", "crafters", "glaive", "greatsword", "hellspawn", "hurrah", "morningstar", "numbskulls", "overmatched", "spearman", "stygian", "teleport", "teleportation", "teleported", "terraform", "wildlands", # game jargon "melee", "arcane", "day/night", "gameplay", "hitpoint", "hitpoints", "FFA", "multiplayer", "playtesting", "respawn", "respawns", "WML", "HP", "XP", "AI", "ZOC", "YW", "L0", "L1", "L2", "L3", "MC", # archaisms "faugh", "hewn", "leapt", "dreamt", "spilt", "grandmam", "grandsire", "grandsires", "scry", "scrying", "scryed", "woodscraft", "princeling", "wilderlands", "ensorcels", "unlooked", "naphtha", "naïve", # Sceptre of Fire gets spelled with -re. "sceptre", ]} pango_conversions = (("~", "", ""), ("@", "", ""), ("#", "", ""), ("*", "", ""), ("`", "", ""), ) def pangostrip(message): "Strip Pango markup out of a string." # This is all known Pango convenience tags for tag in ("b", "big", "i", "s", "sub", "sup", "small", "tt", "u"): message = message.replace("<%s>" % tag, "").replace("" % tag, "") # Now remove general span tags message = re.sub("]*>", "", message) # And Pango specials; message = re.sub("&[a-z]+;", "", message) return message def pangoize(message, filename, line): "Pango conversion of old-style Wesnoth markup." if '&' in message: amper = message.find('&') if message[amper:amper+1].isspace(): message = message[:amper] + "&" + message[amper+1:] rgb = re.search("(?:<|<)([0-9]+),([0-9]+),([0-9]+)(>|>)", message) if rgb: r = int(rgb.group(1)) g = int(rgb.group(2)) b = int(rgb.group(3)) if ( r or g or b) > 255: print '"%s", line %d: RGB color value over 255 requires manual fix (%s).' % (filename, line, rgb.group()) else: hexed = hex(r).replace('0x', '0')[-2:] + hex(g).replace('0x', '0')[-2:] + hex(b).replace('0x', '0')[-2:] print '"%s", line %d: color spec (%s) requires manual fix (, ).' % (filename, line, rgb.group(), hexed) # Hack old-style Wesnoth markup for (oldstyle, newstart, newend) in pango_conversions: if oldstyle not in message: continue where = message.find(oldstyle) if message[where - 1] != '"': # Start of string only continue if message.strip()[-1] != '"': print '"%s", line %d: %s highlight at start of multiline string requires manual fix.' % (filename, line, oldstyle) continue if '+' in message: print '"%s", line %d: %s highlight in composite string requires manual fix.' % (filename, line, oldstyle) continue # This is the common, simple case we can fix automatically message = message[:where] + newstart + message[where + 1:] endq = message.rfind('"') message = message[:endq] + newend + message[endq:] # Check for unescaped < and > if "<" in message or ">" in message: reduced = pangostrip(message) if "<" in reduced or ">" in reduced: if message == reduced: # No pango markup here = message.find('<') if message[here:here+4] != "<": message = message[:here] + "<" + message[here+1:] here = message.find('>') if message[here:here+4] != ">": message = message[:here] + ">" + message[here+1:] else: print '"%s", line %d: < or > in pango string requires manual fix.' % (filename, line) return message class WmllintIterator(WmlIterator): "Fold an Emacs-compatible error reporter into WmlIterator." def printError(self, *misc): """Emit an error locator compatible with Emacs compilation mode.""" if not hasattr(self, 'lineno') or self.lineno == -1: print >>sys.stderr, '"%s":' % self.fname else: print >>sys.stderr, '"%s", line %d:' % (self.fname, self.lineno+1), for item in misc: print >>sys.stderr, item, print >>sys.stderr #terminate line def local_sanity_check(filename, nav, key, prefix, value, comment): "Sanity checks that don't require file context or globals." errlead = '"%s", line %d: ' % (filename, nav.lineno+1) ancestors = nav.ancestors() in_definition = "#define" in ancestors in_call = [x for x in ancestors if x.startswith("{")] ignored = "wmllint: ignore" in nav.text parent = None if ancestors: parent = ancestors[-1] ancestors = ancestors[:-1] # Check for things marked translated that aren't strings if "_" in nav.text and not ignored: m = re.search(r'[=(]\s*_\s+("?)', nav.text) if m and not m.group(1): print errlead + 'translatability mark before non-string' # Most tags are not allowed with [part] if ("[part]" in ancestors or parent == "[part]") and isOpener(nav.element): # FIXME: this should be part of wmliterator's functionality if isExtender(nav.element): actual_tag = "[" + nav.element[2:] else: actual_tag = nav.element if actual_tag not in ("[part]", "[background_layer]", "[image]", "[insert_tag]", "[if]", "[then]", "[else]", "[switch]", "[case]", "[variable]", "[deprecated_message]"): print errlead + '%s not permitted within [part] tag' % nav.element # Most tags are not permitted inside [if] if (len(ancestors) >= 1 and parent == "[if]") or \ (len(ancestors) >= 2 and parent == "#ifdef" and ancestors[-1] == "[if]"): if isOpener(nav.element) and nav.element not in ("[and]", "[else]", "[frame]", "[have_location]", "[have_unit]", "[not]", "[or]", "[then]", "[variable]") and not nav.element.endswith("_frame]") and not nav.element.startswith("[filter"): print errlead + 'illegal child of [if]:', nav.element # Check for fluky credit parts if parent == "[entry]": if key == "email" and " " in value: print errlead + 'space in email name' # Check for various things that shouldn't be outside an [ai] tag if not in_definition and not in_call and not "[ai]" in nav.ancestors() and not ignored: if key in ("number_of_possible_recruits_to_force_recruit", "recruitment_ignore_bad_movement", "recruitment_ignore_bad_combat", "recruitment_pattern", "villages_per_scout", "leader_value", "village_value", "aggression", "caution", "attack_depth", "grouping", "advancements"): print errlead + key + " outside [ai] scope" # Bad [recruit] attribute if parent in ("[allow_recruit]", "[disallow_recruit]") and key == "recruit": print errlead + "recruit= should be type=" # Check [binary_path] and [textdomain] paths if parent == '[textdomain]' and key == 'path' and not '/translations' in value: print errlead + 'no reference to "/translations" directory in textdomain path' if parent == '[binary_path]' and key == 'path': if '/external' in value or '/public' in value: print errlead + '"/external" or "/public" image directories should no longer be used' # Accumulate data to check for missing next scenarios if parent == '[campaign]': if key == "first_scenario" and value != "null": nextrefs.append((filename, nav.lineno, value)) if parent == '[scenario]' or parent == None: if key == "next_scenario" and value != "null": nextrefs.append((filename, nav.lineno, value)) if key == 'id': scenario_to_filename[value] = filename def global_sanity_check(filename, lines): "Perform sanity and consistency checks on input files." # Sanity-check abilities and traits against notes macros. # Note: This check is disabled on units derived via [base_unit]. # Also, build dictionaries of unit movement types and races in_unit_type = None notecheck = True trait_note = dict(notepairs) note_trait = {p[1]:p[0] for p in notepairs} for nav in WmllintIterator(lines, filename): if "wmllint: notecheck off" in nav.text: notecheck = False continue elif "wmllint: notecheck on" in nav.text: notecheck = True #print "Element = %s, text = %s" % (nav.element, `nav.text`) if nav.element == "[unit_type]": unit_race = "" unit_id = "" base_unit = "" traits = [] notes = [] has_special_notes = False in_unit_type = nav.lineno + 1 hitpoints_specified = False continue elif nav.element == "[/unit_type]": #print '"%s", %d: unit has traits %s and notes %s' \ # % (filename, in_unit_type, traits, notes) if unit_id and base_unit: derived_units.append((filename, nav.lineno + 1, unit_id, base_unit)) if unit_id and not base_unit: missing_notes = [] for trait in traits: tn = trait_note[trait] if tn not in notes and tn not in missing_notes: missing_notes.append(tn) missing_traits = [] for note in notes: nt = note_trait[note] if nt not in traits and nt not in missing_traits: missing_traits.append(nt) if (notes or traits) and not has_special_notes: missing_notes = ["{SPECIAL_NOTES}"] + missing_notes # If the unit didn't specify hitpoints, there is some wacky # stuff going on (possibly pseudo-[base_unit] behavior via # macro generation) so disable some of the consistency checks. if not hitpoints_specified: continue if notecheck and missing_notes: print '"%s", line %d: unit %s is missing notes +%s' \ % (filename, in_unit_type, unit_id, "+".join(missing_notes)) if missing_traits: print '"%s", line %d: unit %s is missing traits %s' \ % (filename, in_unit_type, unit_id, "+".join(missing_traits)) if notecheck and not (notes or traits) and has_special_notes: print '"%s", line %d: unit %s has superfluous {SPECIAL_NOTES}' \ % (filename, in_unit_type, unit_id) if not "[theme]" in nav.ancestors() and not "[base_unit]" in nav.ancestors() and not unit_race: print '"%s", line %d: unit %s has no race' \ % (filename, in_unit_type, unit_id) in_unit_type = None traits = [] notes = [] unit_id = "" base_unit = "" has_special_notes = False unit_race = None if '[unit_type]' in nav.ancestors() and not "[filter_attack]" in nav.ancestors(): try: (key, prefix, value, comment) = parse_attribute(nav.text) if key == "id": if value[0] == "_": value = value[1:].strip() if not unit_id and not "[base_unit]" in nav.ancestors(): unit_id = value unit_types.append(unit_id) if not base_unit and "[base_unit]" in nav.ancestors(): base_unit = value elif key == "hitpoints": hitpoints_specified = True elif key == "usage": assert(unit_id) usage[unit_id] = value elif key == "movement_type": if '{' not in value: assert(unit_id) unit_movetypes.append((unit_id, filename, nav.lineno + 1, value)) elif key == "race": if '{' not in value: assert(unit_id or base_unit) unit_race = value unit_races.append((unit_id, filename, nav.lineno + 1, unit_race)) elif key == "advances_to": assert(unit_id or base_unit) advancements = value if advancements.strip() != "null": advances.append((unit_id, filename, nav.lineno + 1, advancements)) except TypeError: pass precomment = nav.text if '#' in nav.text: precomment = nav.text[:nav.text.find("#")] if "{SPECIAL_NOTES}" in precomment: has_special_notes = True for (p, q) in notepairs: if p in precomment: traits.append(p) if q in precomment: notes.append(q) # Detect units that speak in their death events filter_subject = None die_event = False deathcheck = True for nav in WmllintIterator(lines, filename): if "wmllint: deathcheck off" in nav.text: deathcheck = False continue elif "wmllint: deathcheck on" in nav.text: deathcheck = True if "[/event]" in nav.text: filter_subject = None die_event = False elif not nav.ancestors(): continue elif "[event]" in nav.ancestors(): parent = nav.ancestors()[-1] if parent == "[event]": # Check if it's a death event fields = parse_attribute(nav.text) if fields: (key, prefix, value, comment) = fields if key == 'name' and value == 'die': die_event = True elif die_event and not filter_subject and parent == "[filter]": # Check to see if it has a filter subject if "id" in nav.text: try: (key,prefix,value,comment) = parse_attribute(nav.text) filter_subject = value except TypeError: pass elif die_event and filter_subject and parent == "[message]": # Who is speaking? fields = parse_attribute(nav.text) if fields: (key, prefix, value, comment) = fields if key in ("id", "speaker"): if deathcheck and ((value == filter_subject) or (value == "unit")): print '"%s", line %d: %s speaks in his/her "die" event rather than "last breath"' \ % (filename, nav.lineno+1, value) # Collect information on defined movement types and races for nav in WmllintIterator(lines, filename): above = nav.ancestors() if above and above[-1] in ("[movetype]", "[race]"): try: (key, prefix, value, comment) = parse_attribute(nav.text) if above[-1] == "[movetype]" and key == 'name': movetypes.append(value) if above[-1] == "[race]" and key == 'id': races.append(value) except TypeError: pass # Sanity-check recruit and recruitment_pattern. # This code has a limitation; if there are multiple instances of # recruit and recruitment_pattern (as can happen if these lists # vary by EASY/NORMAL/HARD level) this code will only record the # last of each for later consistency checking. in_side = False in_ai = in_subunit = False recruit = {} in_generator = False sidecount = 0 recruitment_pattern = {} ifdef_stack = [None] for i in xrange(len(lines)): line = lines[i].strip() if line.startswith("#ifdef") or line.startswith("#ifhave") or line.startswith("#ifver"): ifdef_stack.append(line.split()[1]) continue if line.startswith("#ifndef") or line.startswith("#ifnhave") or line.startswith("#ifnver"): ifdef_stack.append("!" + line.split()[1]) continue if line.startswith("#else"): if ifdef_stack[-1].startswith("!"): ifdef_stack.append(ifdef_stack[-1][1:]) else: ifdef_stack.append("!" + ifdef_stack[-1]) continue if line.startswith("#endif"): ifdef_stack.pop() continue if "[generator]" in lines[i]: in_generator = True continue elif "[/generator]" in lines[i]: in_generator = False continue elif "[side]" in lines[i]: in_side = True sidecount += 1 continue elif "[/side]" in lines[i]: if recruit or recruitment_pattern: sides.append((filename, recruit, recruitment_pattern)) in_side = False recruit = {} recruitment_pattern = {} continue elif in_side and "[ai]" in lines[i]: in_ai = True continue elif in_side and "[unit]" in lines[i]: in_subunit = True continue elif in_side and "[/ai]" in lines[i]: in_ai = False continue elif in_side and "[/unit]" in lines[i]: in_subunit = False continue if "wmllint: skip-side" in lines[i]: sidecount += 1 if not in_side or in_subunit or '=' not in lines[i]: continue try: (key, prefix, value, comment) = parse_attribute(lines[i]) if key in ("recruit", "extra_recruit") and value: recruit[ifdef_stack[-1]] = (i+1, [x.strip() for x in value.split(",")]) elif key == "recruitment_pattern" and value: if not in_ai: print '"%s", line %d: recruitment_pattern outside [ai]' \ % (filename, i+1) else: recruitment_pattern[ifdef_stack[-1]] = (i+1, [x.strip() for x in value.split(",")]) elif key == "side" and not in_ai: try: if not in_generator and sidecount != int(value): print '"%s", line %d: side number %s is out of sequence (%d expected)' \ % (filename, i+1, value, sidecount) except ValueError: pass # Ignore ill-formed integer literals except TypeError: pass # Sanity check ellipses # Starting from 1.11.5, units with canrecruit=yes gain automatically a leader ellipse # Starting from 1.11.6, units without a ZoC gain automatically a nozoc ellipse # Check if ellipse= was used and warn if so # Do not warn if misc/ellipse-hero was used, since it isn't automatically assigned by C++ # and it's assigned/removed with IS_HERO/MAKE_HERO/UNMAKE_HERO # magic comment wmllint: no ellipsecheck deactivates this check for the current line in_effect = False in_unit = False in_side = False in_unit_type = False for i in xrange(len(lines)): if "[effect]" in lines[i]: in_effect = True elif "[/effect]" in lines[i]: in_effect = False elif "[unit]" in lines[i]: in_unit = True elif "[/unit]" in lines[i]: in_unit = False elif "[side]" in lines[i]: in_side = True elif "[/side]" in lines[i]: in_side = False elif "[unit_type]" in lines[i]: in_unit_type = True elif "[/unit_type]" in lines[i]: in_unit_type = False # ellipsecheck magic comment allows to deactivate the ellipse sanity check if "wmllint: no ellipsecheck" not in lines[i]: if in_effect: try: (key, prefix, value, comment) = parse_attribute(lines[i]) if key == "ellipse" and value in ("misc/ellipse-nozoc","misc/ellipse-leader"): print '"%s", line %d: [effect] apply_to=ellipse needs to be removed' % (filename, i+1) elif key == "ellipse" and value not in ("none","misc/ellipse","misc/ellipse-hero"): print '"%s", line %d: custom ellipse %s may need to be updated' % (filename, i+1, value) except TypeError: # this is needed to handle tags, that parse_attribute cannot split pass elif in_unit or in_side or in_unit_type: try: (key, prefix, value, comment) = parse_attribute(lines[i]) if key == "ellipse" and value in ("misc/ellipse-nozoc","misc/ellipse-leader"): print '"%s", line %d: %s=%s needs to be removed' % (filename, i+1, key, value) elif key == "ellipse" and value not in ("none","misc/ellipse","misc/ellipse-hero"): print '"%s", line %d: custom ellipse %s may need to be updated' % (filename, i+1, value) except TypeError: # this is needed to handle tags, that parse_attribute cannot split pass # Interpret various magic comments for i in xrange(len(lines)): # Interpret magic comments for setting the usage pattern of units. # This coped with some wacky UtBS units that were defined with # variant-spawning macros. The prototype comment looks like this: #wmllint: usage of "Desert Fighter" is fighter m = re.search('# *wmllint: usage of "([^"]*)" is +(.*)', lines[i]) if m: usage[m.group(1)] = m.group(2).strip() unit_types.append(m.group(1)) # Magic comment for adding non-standard usage types m = re.search('# *wmllint: usagetypes? +(.*)', lines[i]) if m: for newusage in m.group(1).split(","): usage_types.append(newusage.strip()) # Accumulate global spelling exceptions words = re.search("wmllint: general spellings? (.*)", lines[i]) if words: for word in words.group(1).split(): declared_spellings["GLOBAL"].append(word.lower()) words = re.search("wmllint: directory spellings? (.*)", lines[i]) if words: fdir = os.path.dirname(filename) if fdir not in declared_spellings: declared_spellings[fdir] = [] for word in words.group(1).split(): declared_spellings[fdir].append(word.lower()) # Consistency-check the id= attributes in [side], [unit], [recall], # and [message] scopes, also correctness-check translation marks and look # for double spaces at end of sentence. present = [] in_scenario = False in_multiplayer = False subtag_depth = 0 in_person = False in_trait = False ignore_id = False in_object = False in_stage = False in_cfg = False in_goal = False in_set_menu_item = False in_clear_menu_item = False in_facet = False in_sound_source = False in_remove_sound_source = False in_message = False in_option = False #last index is true: we're currently directly in an [event] #this avoids complaints about unknown [event]id=something, but keeps the check #in case some [filter]id= comes in this [event] directly_in_event = [] in_time_area = False in_store = False in_unstore = False in_not = False in_clear = False storeid = None storevar = None ignoreable = False preamble_seen = False sentence_end = re.compile("(?<=[.!?;:]) +") capitalization_error = re.compile("(?<=[.!?]) +[a-z]") markcheck = True translation_mark = re.compile(r'_ *"') for i in xrange(len(lines)): if '[' in lines[i]: preamble_seen = True # This logic looks odd because a scenario can be conditionally # wrapped in both [scenario] and [multiplayer]; we mustn't count # either as a subtag even if it occurs inside the other, otherwise # this code might see id= declarations as being at the wrong depth. if "[scenario]" in lines[i]: in_scenario = True preamble_seen = False elif "[/scenario]" in lines[i]: in_scenario = False elif "[multiplayer]" in lines[i]: in_multiplayer = True preamble_seen = False elif "[/multiplayer]" in lines[i]: in_multiplayer = False else: if re.search(r"\[[a-z]", lines[i]): subtag_depth += 1 if "[/" in lines[i]: subtag_depth -= 1 if "[event]" in lines[i]: directly_in_event.append(True) elif re.search(r"\[[a-z]", lines[i]): directly_in_event.append(False) elif "[/" in lines[i]: if len(directly_in_event) > 0: directly_in_event.pop() # Ordinary subtag flags begin here if "[trait]" in lines[i]: in_trait = True elif "[/trait]" in lines[i]: in_trait = False elif "[object]" in lines[i]: in_object = True elif "[/object]" in lines[i]: in_object = False elif "[stage]" in lines[i]: in_stage = True elif "[/stage]" in lines[i]: in_stage = False elif "[cfg]" in lines[i]: in_cfg = True elif "[/cfg]" in lines[i]: in_cfg = False elif "[goal]" in lines[i]: in_goal = True elif "[/goal]" in lines[i]: in_goal = False elif "[set_menu_item]" in lines[i]: in_set_menu_item = True elif "[/set_menu_item]" in lines[i]: in_set_menu_item = False elif "[clear_menu_item]" in lines[i]: in_clear_menu_item = True elif "[/clear_menu_item]" in lines[i]: in_clear_menu_item = False elif "[facet]" in lines[i]: in_facet = True elif "[/facet]" in lines[i]: in_facet = False elif "[sound_source]" in lines[i]: in_sound_source = True elif "[/sound_source]" in lines[i]: in_sound_source = False elif "[remove_sound_source]" in lines[i]: in_remove_sound_source = True elif "[/remove_sound_source]" in lines[i]: in_remove_sound_source = False elif "[message]" in lines[i]: in_message = True elif "[/message]" in lines[i]: in_message = False elif "[/option]" in lines[i]: in_option = False elif "[option]" in lines[i]: in_option = True elif "[time_area]" in lines[i]: in_time_area = True elif "[/time_area]" in lines[i]: in_time_area = False elif "[label]" in lines[i] or "[chamber]" in lines[i] or "[time]" in lines[i]: ignore_id = True elif "[/label]" in lines[i] or "[/chamber]" in lines[i] or "[/time]" in lines[i]: ignore_id = False elif "[kill]" in lines[i] or "[effect]" in lines[i] or "[move_unit_fake]" in lines[i] or "[scroll_to_unit]" in lines[i]: ignoreable = True elif "[/kill]" in lines[i] or "[/effect]" in lines[i] or "[/move_unit_fake]" in lines[i] or "[/scroll_to_unit]" in lines[i]: ignoreable = False elif "[side]" in lines[i] or "[unit]" in lines[i] or "[recall]" in lines[i]: in_person = True continue elif "[/side]" in lines[i] or "[/unit]" in lines[i] or "[/recall]" in lines[i]: in_person = False elif "[store_unit]" in lines[i]: in_store = True elif "[/store_unit]" in lines[i]: if storeid and storevar: storedids.update({storevar: storeid}) in_store = False storeid = storevar = None elif "[unstore_unit]" in lines[i]: in_unstore = True elif "[/unstore_unit]" in lines[i]: in_unstore = False elif "[not]" in lines[i]: in_not = True elif "[/not]" in lines[i]: in_not = False elif "[clear_variable]" in lines[i]: in_clear = True elif "[/clear_variable]" in lines[i]: in_clear = False if "wmllint: markcheck off" in lines[i]: markcheck = False elif "wmllint: markcheck on" in lines[i]: markcheck = True elif 'wmllint: who ' in lines[i]: try: fields = lines[i].split("wmllint: who ", 1)[1].split(" is ", 1) if len(fields) == 2: mac = string_strip(fields[0].strip()).strip('{}') if whopairs.has_key(mac): whopairs[mac] = whopairs[mac] + ", " + fields[1].strip() else: whopairs.update({mac: fields[1].strip()}) except IndexError: pass elif 'wmllint: unwho ' in lines[i]: unmac = lines[i].split("wmllint: unwho ", 1)[1].strip() if string_strip(unmac).upper() == 'ALL': whopairs.clear() else: try: del whopairs[string_strip(unmac).strip('{}')] except KeyError: print >>sys.stderr, '%s, line %s: magic comment "unwho %s" does not match any current keys: %s' \ % (filename, i+1, unmac, str(whopairs.keys())[1:-1]) elif 'wmllint: whofield' in lines[i]: fields = re.search(r'wmllint: whofield\s+([^\s]+)(\s+is)?\s*([^\s]*)', lines[i]) if fields: if fields.group(1).startswith('clear'): if fields.group(3) in whomacros.keys(): del whomacros[fields.group(3)] else: whomacros.clear() elif re.match(r'[1-9][0-9]*$', fields.group(3)): whomacros.update({fields.group(1): int(fields.group(3))}) else: try: del whomacros[fields.group(1)] except KeyError: print >>sys.stderr, '%s, line %s: magic comment "whofield %s" should be followed by a number: %s' \ % (filename, i+1, unmac, fields.group(3)) # Parse recruit/recall macros to recognize characters. This section # assumes that such a macro is the first item on a line. leadmac = re.match(r'{[^}\s]+.', lines[i].lstrip()) if leadmac: macname = leadmac.group()[1:-1] # Recognize macro pairings from "wmllint: who" magic # comments. if macname in whopairs.keys(): for who in whopairs[macname].split(","): if who.strip().startswith("--"): try: present.remove(who.replace('--', '', 1).strip()) except: ValueError else: present.append(who.strip()) elif not leadmac.group().endswith('}'): # Update 1.4's {LOYAL_UNIT} macro to {NAMED_LOYAL_UNIT}. Do # this here rather than hack_syntax so the character can be # recognized. if macname == 'LOYAL_UNIT': (args, brack, paren) = parse_macroref(0, leadmac.string) if len(args) == 7: lines[i] = lines[i].replace('{LOYAL_UNIT', '{NAMED_LOYAL_UNIT', 1) # Auto-recognize the people in the {NAMED_*UNIT} macros. if re.match(r'NAMED_[A-Z_]*UNIT$', macname): (args, brack, paren) = parse_macroref(0, leadmac.string) if len(args) >= 7 and \ re.match(r'([0-9]+|[^\s]*\$[^\s]*side[^\s]*|{[^\s]*SIDE[^\s]*})$', args[1]) and \ re.match(r'([0-9]+|[^\s]*\$[^\s]*x[^\s]*|{[^\s]*X[^\s]*})$', args[3]) and \ re.match(r'([0-9]+|[^\s]*\$[^\s]*y[^\s]*|{[^\s]*Y[^\s]*})$', args[4]) and \ len(args[5]) > 0: present.append(args[5]) elif macname == 'RECALL': (args, brack, paren) = parse_macroref(0, leadmac.string) if len(args) == 2 and brack == 0: present.append(args[1]) elif macname == 'RECALL_XY': (args, brack, paren) = parse_macroref(0, leadmac.string) if len(args) == 4: present.append(args[1]) elif macname == 'CLEAR_VARIABLE': (args, brack, paren) = parse_macroref(0, leadmac.string) for arg in [x.lstrip() for x in args[1].split(',')]: if arg in storedids.keys(): del storedids[arg] elif macname in whomacros.keys(): (args, brack, paren) = parse_macroref(0, leadmac.string) present.append(args[whomacros[macname]]) m = re.search("# *wmllint: recognize +(.*)", lines[i]) if m: present.append(string_strip(m.group(1)).strip()) if '=' not in lines[i] or ignoreable: continue parseable = False try: (key, prefix, value, comment) = parse_attribute(lines[i]) parseable = True except TypeError: pass if parseable: if "wmllint: ignore" in comment: continue # Recognize units when unstored if (in_scenario or in_multiplayer) and in_store: if key == 'id' and not in_not: if not storeid == None: storeid == storeid + ',' + value else: storeid = value elif key == 'variable' and '{' not in value: storevar = value elif in_unstore: if key == 'variable': value = value.split("[$")[0] if value in storedids.keys(): for unit_id in storedids[value].split(','): present.append(unit_id.lstrip()) del storedids[value] elif key == 'name' and in_clear: for val in value.split(','): val = val.lstrip() if val in storedids.keys(): del storedids[val] has_tr_mark = translation_mark.search(value) if key == 'role': present.append(value) if has_tr_mark: # FIXME: This test is rather bogus as is. # Doing a better job would require tokenizing to pick up the # string boundaries. I'd do it, but AI0867 is already working # on a parser-based wmllint. if '{' in value and "+" not in value and value.find('{') > value.find("_"): print '"%s", line %d: macro reference in translatable string'\ % (filename, i+1) #if future and re.search("[.,!?] ", lines[i]): # print '"%s", line %d: extraneous space in translatable string'\ # % (filename, i+1) # Check correctness of translation marks and descriptions if key.startswith("#"): # FIXME: parse_attribute is confused. pass elif key.startswith("{"): pass elif key == 'letter': # May be led with _s for void pass elif key in ('name', 'male_name', 'female_name', 'value'): # FIXME: check this someday pass elif key == "variation_name": if markcheck and not has_tr_mark: print '"%s", line %d: %s should be renamed as variation_id and/or marked as translatable' \ % (filename, i+1, key) elif translatables.search(key): if markcheck and has_tr_mark and lines[i].find("\"\"")>-1: print '"%s", line %d: %s doesn`t need translation mark (translatable string is empty)' \ % (filename, i+1, key) lines[i] = lines[i].replace("=_","=") if markcheck and not value.startswith("$") and not value.startswith("{") and not re.match(" +", value) and not has_tr_mark and lines[i].find("\"\"")==-1 and not ("wmllint: ignore" in comment or "wmllint: noconvert" in comment): print '"%s", line %d: %s needs translation mark' \ % (filename, i+1, key) lines[i] = lines[i].replace('=', "=_ ") nv = sentence_end.sub(" ", value) if nv != value: print '"%s", line %d: double space after sentence end' \ % (filename, i+1) if not stringfreeze: lines[i] = sentence_end.sub(" ", lines[i]) if capitalization_error.search(lines[i]): print '"%s", line %d: probable capitalization or punctuation error' \ % (filename, i+1) if key == "message" and in_message and not in_option and not ("wmllint: ignore" in comment or "wmllint: noconvert" in comment): lines[i] = pangoize(lines[i], filename, i) else: if (in_scenario or in_multiplayer) and key == "id": if in_person: present.append(value) elif value and value[0] in ("$", "{"): continue elif preamble_seen and subtag_depth > 0 and not ignore_id and not in_object and not in_cfg and not in_facet and not in_sound_source and not in_remove_sound_source and not in_stage and not in_goal and not in_set_menu_item and not in_clear_menu_item and not directly_in_event[-1] and not in_time_area and not in_trait: ids = value.split(",") for j in xrange(len(ids)): # removal of leading whitespace of items in comma-separated lists # is usually supported in the mainline wesnoth lua scripts # not sure about trailing one # also, do not complain about ids if they're referred to a menu item being cleared if ids[j].lstrip() not in present: print '"%s", line %d: unknown \'%s\' referred to by id' \ % (filename, i+1, ids[j]) if (in_scenario or in_multiplayer) and key == "speaker": if value not in present and value not in ("narrator", "unit", "second_unit") and value[0] not in ("$", "{"): print '"%s", line %d: unknown speaker \'%s\' of [message]' \ % (filename, i+1, value) if markcheck and has_tr_mark and not ("wmllint: ignore" in comment or "wmllint: noconvert" in comment): print '"%s", line %d: %s should not have a translation mark' \ % (filename, i+1, key) lines[i] = prefix + value.replace("_", "", 1) + comment + '\n' # Now that we know who's present, register all these names as spellings declared_spellings[filename] = [x.lower() for x in present if len(x) > 0] # Check for textdomain strings; should be exactly one, on line 1 textdomains = [] no_text = False for i in xrange(len(lines)): if "#textdomain" in lines[i]: textdomains.append(i+1) elif "wmllint: no translatables": no_text = True if not no_text: if not textdomains: print '"%s", line 1: no textdomain string' % filename elif textdomains[0] == 1: # Multiples are OK if first is on line 1 pass elif len(textdomains) > 1: print '"%s", line %d: multiple textdomain strings on lines %s' % \ (filename, textdomains[0], ", ".join(map(str, textdomains))) else: w = textdomains[0] print '"%s", line %d: single textdomain declaration not on line 1.' % \ (filename, w) lines = [lines[w-1].lstrip()] + lines[:w-1] + lines[w:] return lines def condition_match(p, q): "Do two condition-states match?" # The empty condition state is represented by None if p is None or q is None or (p == q): return True # Past this point it's all about handling cases with negation sp = p np = False if sp.startswith("!"): sp = sp[1:] np = True sq = q nq = False if sq.startswith("!"): sq = sp[1:] nq == True return (sp != sq) and (np != nq) def consistency_check(): "Consistency-check state information picked up by sanity_check" derivedlist = [x[2] for x in derived_units] baselist = [x[3] for x in derived_units] derivations = dict(zip(derivedlist, baselist)) for (filename, recruitdict, patterndict) in sides: for (rdifficulty, (rl, recruit)) in recruitdict.items(): utypes = [] for rtype in recruit: base = rtype if rtype not in unit_types: # Assume WML coder knew what he was doing if macro reference if not rtype.startswith("{"): print '"%s", line %d: %s is not a known unit type' % (filename, rl, rtype) continue elif rtype not in usage: if rtype in derivedlist: base = derivations[rtype] else: print '"%s", line %d: %s has no usage type' % \ (filename, rl, rtype) continue if not base in usage: print '"%s", line %d: %s has unknown base %s' % \ (filename, rl, rtype, base) continue else: utype = usage[base] utypes.append(utype) for (pdifficulty, (pl, recruit_pattern)) in patterndict.items(): if condition_match(pdifficulty, rdifficulty): if utype not in recruit_pattern: rshow = '' if rdifficulty is not None: rshow = 'At ' + rdifficulty + ', ' ushow = '' if utype not in usage_types: ushow = ', a non-standard usage class' pshow = '' if pdifficulty is not None: pshow = ' ' + pdifficulty print '"%s", line %d: %s%s (%s%s) doesn\'t match the%s recruitment pattern (%s) for its side' % (filename, rl, rshow, rtype, utype, ushow, pshow, ", ".join(recruit_pattern)) # We have a list of all the usage types recruited at this difficulty # in utypes. Use it to check the matching pattern, if any. Suppress # this check if the recruit line is a macroexpansion. if recruit and not recruit[0].startswith("{"): for (pdifficulty, (pl, recruitment_pattern)) in patterndict.items(): if condition_match(pdifficulty, rdifficulty): for utype in recruitment_pattern: if utype not in utypes: rshow = '.' if rdifficulty is not None: rshow = ' at difficulty ' + rdifficulty + '.' ushow = '' if utype not in usage_types: ushow = ' (a non-standard usage class)' print '"%s", line %d: no %s%s units recruitable%s' % (filename, pl, utype, ushow, rshow) if movetypes: for (unit_id, filename, line, movetype) in unit_movetypes: if movetype not in movetypes: print '"%s", line %d: %s has unknown movement type' \ % (filename, line, unit_id) if races: for (unit_id, filename, line, race) in unit_races: if race not in races: print '"%s", line %d: %s has unknown race' \ % (filename, line, unit_id) # Should we be checking the transitive closure of derivation? # It's not clear whether [base_unit] works when the base is itself derived. for (filename, line, unit_type, base_unit) in derived_units: if base_unit not in unit_types: print '"%s", line %d: derivation of %s from %s does not resolve' \ % (filename, line, unit_type, base_unit) # Check that all advancements are known units for (unit_id, filename, lineno, advancements) in advances: advancements = map(string.strip, advancements.split(",")) bad_advancements = [x for x in advancements if x not in (unit_types+derivedlist)] if bad_advancements: print '"%s", line %d: %s has unknown advancements %s' \ % (filename, lineno, unit_id, bad_advancements) # Check next-scenario pointers #print "Scenario ID map", scenario_to_filename for (filename, lineno, value) in nextrefs: if value not in scenario_to_filename: print '"%s", line %d: unresolved scenario reference %s' % \ (filename, lineno, value) # Report stored units never unstored or cleared for store in storedids.keys(): print 'wmllint: stored unit "%s" not unstored or cleared from "%s"' % (storedids[store], store) # Syntax transformations leading_ws = re.compile(r"^\s*") def leader(s): "Return a copy of the leading whitespace in the argument." return leading_ws.match(s).group(0) def hack_syntax(filename, lines): # Syntax transformations go here. This gets called once per WML file; # the name of the file is passed as filename, text of the file as the # array of strings in lines. Modify lines in place as needed; # changes will be detected by the caller. # # Deal with a few Windows-specific problems for the sake of cross- # platform harmony. First, the use of backslashes in file paths. for i in xrange(len(lines)): if "no-syntax-rewrite" in lines[i]: break if lines[i].lstrip().startswith("#"): pass # Looking out for "#" used for color markup precomment = re.split(r'\s#', lines[i], 1)[0] comment = lines[i][len(precomment):] if '\\' in precomment: while re.search(r'(? %s -- please use frontslash (/) for cross-platform compatibility' \ % (filename, i+1, backslash.group(), fronted) # Then get rid of the 'userdata/' headache. if 'userdata/' in precomment: while re.search(r'user(data/)?data/[ac]', precomment): userdata = re.search(r'(?:\.\./)?user(?:data/)?(data/[ac][^/]*/?)', precomment) precomment = precomment[:userdata.start()] + userdata.group(1) + precomment[userdata.end():] print '"%s", line %d: %s -> %s -- DO NOT PREFIX PATHS WITH "userdata/"' \ % (filename, i+1, userdata.group(), userdata.group(1)) lines[i] = precomment + comment # Ensure that every attack has a translatable description. for i in xrange(len(lines)): if "no-syntax-rewrite" in lines[i]: break elif "[attack]" in lines[i]: j = i; have_description = False while '[/attack]' not in lines[j]: if lines[j].strip().startswith("description"): have_description = True j += 1 if not have_description: j = i while '[/attack]' not in lines[j]: fields = lines[j].strip().split('#') syntactic = fields[0] comment = "" if len(fields) > 1: comment = fields[1] if syntactic.strip().startswith("name"): description = syntactic.split("=")[1].strip() if not description.startswith('"'): description = '"' + description + '"\n' # Skip the insertion if this is a dummy declaration # or one modifying an attack inherited from a base unit. if "no-icon" not in comment: new_line = leader(syntactic) + "description=_"+description if verbose: print '"%s", line %d: inserting %s' % (filename, i+1, repr(new_line)) lines.insert(j+1, new_line) j += 1 j += 1 # Ensure that every speaker=narrator block without an image uses # wesnoth-icon.png as an image. need_image = in_message = False for i in xrange(len(lines)): if "no-syntax-rewrite" in lines[i]: break precomment = lines[i].split("#")[0] if '[message]' in precomment: in_message = True if "speaker=narrator" in precomment: need_image = True elif precomment.strip().startswith("image"): need_image = False elif '[/message]' in precomment: if need_image: # This line presumes the code has been through wmlindent if verbose: print '"%s", line %d: inserting "image=wesnoth-icon.png"'%(filename, i+1) lines.insert(i, leader(precomment) + baseindent + "image=wesnoth-icon.png\n") need_image = in_message = False # Hack tracking-map macros from 1.4 and earlier. The idea is to lose # all assumptions about colors in the names for i in xrange(len(lines)): if "no-syntax-rewrite" in lines[i]: break if lines[i].lstrip().startswith("#"): pass elif "{DOT_CENTERED" in lines[i]: lines[i] = lines[i].replace("DOT_CENTERED", "NEW_JOURNEY") elif "{DOT_WHITE_CENTERED" in lines[i]: lines[i] = lines[i].replace("DOT_WHITE_CENTERED", "OLD_JOURNEY") elif "{CROSS_CENTERED" in lines[i]: lines[i] = lines[i].replace("CROSS_CENTERED", "NEW_BATTLE") elif "{CROSS_WHITE_CENTERED" in lines[i]: lines[i] = lines[i].replace("CROSS_WHITE_CENTERED", "OLD_BATTLE") elif "{FLAG_RED_CENTERED" in lines[i]: lines[i] = lines[i].replace("FLAG_RED_CENTERED", "NEW_REST") elif "{FLAG_WHITE_CENTERED" in lines[i]: lines[i] = lines[i].replace("FLAG_WHITE_CENTERED", "OLD_REST") elif "{DOT " in lines[i] or "CROSS" in lines[i]: m = re.search("{(DOT|CROSS) ([0-9]+) ([0-9]+)}", lines[i]) if m: n = m.group(1) if n == "DOT": n = "NEW_JOURNEY" if n == "CROSS": n = "NEW_BATTLE" x = int(m.group(2)) + 5 y = int(m.group(3)) + 5 lines[i] = lines[i][:m.start(0)] +("{%s %d %d}" % (n, x, y)) + lines[i][m.end(0):] # Fix bare strings containing single quotes; these confuse wesnoth-mode.el for i in xrange(len(lines)): if "no-syntax-rewrite" in lines[i]: break elif lines[i].count("'") % 2 == 1: try: (key, prefix, value, comment) = parse_attribute(lines[i]) if "'" in value and value[0].isalpha() and value[-1].isalpha() and not '"'+value+'"' in lines[i]: newtext = prefix + '"' + value + '"' + comment + "\n" if lines[i] != newtext: lines[i] = newtext if verbose: print '"%s", line %d: quote-enclosing attribute value.'%(filename, i+1) except TypeError: pass # Palette transformation for 1.7: for i in xrange(len(lines)): if "no-syntax-rewrite" in lines[i]: break if lines[i].lstrip().startswith("#"): pass # RC -> PAL elif "RC" in lines[i]: lines[i] = re.sub(r"~RC\(([^=\)]*)=([^)]*)\)",r"~PAL(\1>\2)",lines[i]) # Rename the terrain definition tag in_standing_anim = False for i in xrange(len(lines)): if "no-syntax-rewrite" in lines[i]: break if lines[i].lstrip().startswith("#"): pass # Ugh...relies on code having been wmlindented lines[i] = re.sub(r"^\[terrain\]", "[terrain_type]", lines[i]) lines[i] = re.sub(r"^\[/terrain\]", "[/terrain_type]", lines[i]) if "[standing_anim]" in lines[i]: in_standing_anim = True if "[/standing_anim]" in lines[i]: in_standing_anim = False if in_standing_anim: lines[i] = re.sub(r"terrain([^_])", r"terrain_type\1", lines[i]) # Rename two attributes in [set_variable] in_set_variable = False for i in xrange(len(lines)): if "no-syntax-rewrite" in lines[i]: break if lines[i].lstrip().startswith("#"): pass if "[set_variable]" in lines[i]: in_set_variable = True if "[/set_variable]" in lines[i]: in_set_variable = False if in_set_variable: lines[i] = re.sub(r"format(?=\s*=)", r"value", lines[i]) lines[i] = re.sub(r"random(?=\s*=)", r"rand", lines[i]) # campaigns directory becomes add-ons in_binary_path = in_textdomain = False for i in xrange(len(lines)): if "no-syntax-rewrite" in lines[i]: break if lines[i].lstrip().startswith("#"): pass # This is done on every line if "campaigns/" in lines[i]: lines[i] = lines[i].replace("{~campaigns/", "{~add-ons/") lines[i] = lines[i].replace("{~/campaigns/", "{~add-ons/") lines[i] = lines[i].replace("{@campaigns/", "{~add-ons/") # Convert UMC to data/add-ons without clobbering mainline. Each path # is checked against a list of mainline campaigns. UMC paths are # updated to "data/add-ons/"; mainline path strings are unaltered. x = 0 for dc in re.finditer(r"data/campaigns/(\w[\w'&+-]*)", lines[i]): if dc.group(1) in mainline: continue # Because start() and end() are based on the original position # of each iteration, while each replacement shortens the line # by two characters, we must subtract an increment that grows # with each substitution. lines[i] = lines[i][:dc.start()-x] + 'data/add-ons/' + dc.group(1) + lines[i][dc.end()-x:] x = x+2 print '"%s", line %d: data/campaigns/%s -> data/add-ons/%s'\ %(filename, i+1, dc.group(1), dc.group(1)) elif "@add-ons/" in lines[i]: lines[i] = lines[i].replace("{@add-ons/", "{~add-ons/") # Occasionally authors try to use '~' with [textdomain] or [binary_path]. if "[binary_path]" in lines[i]: in_binary_path = True if "[/binary_path]" in lines[i]: in_binary_path = False if "[textdomain]" in lines[i]: in_textdomain = True if "[/textdomain]" in lines[i]: in_textdomain = False if in_binary_path or in_textdomain: if '~' in lines[i]: tilde = re.search('(^\s*path) *= *([^#]{0,5})(~/?(data/)?add-ons/)', lines[i]) if tilde: lines[i] = tilde.group(1) + '=' + tilde.group(2) + 'data/add-ons/' + lines[i][tilde.end():] print '"%s", line %d: %s -> data/add-ons/ -- [textdomain] and [binary_path] paths do not accept "~" for userdata'\ % (filename, i+1, tilde.group(3)) # some tags do no longer support default side=1 attribute but may use [filter_side] # with a SSF instead # (since 1.9.5, 1.9.6) if missingside: side_one_tags_allowing_filter_side = ( ("remove_shroud"), ("place_shroud"), ("gold"), ("modify_side"), ("modify_ai") ) outside_of_theme_wml = True # theme wml contains a [gold] tag - exclude that case in_side_one_tag = False side_one_tag_needs_side_one = True for i in range(len(lines)): if "no-syntax-rewrite" in lines[i]: break precomment = lines[i].split("#")[0] if outside_of_theme_wml: if "[theme]" in precomment: outside_of_theme_wml = False else: if "[/theme]" in precomment: outside_of_theme_wml = True if outside_of_theme_wml: if not in_side_one_tag: for j in range(len(side_one_tags_allowing_filter_side)): if "[" + side_one_tags_allowing_filter_side[j] + "]" in precomment: in_side_one_tag = True else: if side_one_tag_needs_side_one: if "side=" in precomment: side_one_tag_needs_side_one = False if "[filter_side]" in precomment: side_one_tag_needs_side_one = False for j in range(len(side_one_tags_allowing_filter_side)): if "[/" + side_one_tags_allowing_filter_side[j] + "]" in precomment: if side_one_tag_needs_side_one: if verbose: print '"%s", line %d: [%s] without "side" attribute is now applied to all sides'%(filename, i+1, side_one_tags_allowing_filter_side[j]) #lines.insert(i, leader(precomment) + baseindent + "side=1\n") in_side_one_tag = False side_one_tag_needs_side_one = True break # More syntax transformations would go here. return lines def maptransform(filename, baseline, inmap, y): # Transform lines in maps for i in xrange(len(inmap[y])): for (old, new) in mapchanges: inmap[y][i] = inmap[y][i].replace(old, new) # Generic machinery starts here def is_map(filename): "Is this file a map?" return filename.endswith(".map") if 0: # Not used, as there are currently no defined map transforms class maptransform_error: "Error object to be thrown by maptransform." def __init__(self, infile, inline, type): self.infile = infile self.inline = inline self.type = type def __repr__(self): return '"%s", line %d: %s' % (self.infile, self.inline, self.type) def maptransform_sample(filename, baseline, inmap, y): "Transform a map line." # Sample to illustrate how map-transformation hooks are called. # The baseline argument will be the starting line number of the map. # The inmap argument will be a 2D string array containing the # entire map. y will be the vertical coordinate of the map line. # You pass a list of these as the second argument of translator(). raise maptransform_error(filename, baseline+y+1, "unrecognized map element at line %d" % (y,)) tagstack = [] # For tracking tag nesting def outermap(func, inmap): "Apply a transformation based on neighborhood to the outermost ring." # Top and bottom rows for i in xrange(len(inmap[0])): inmap[0][i] = func(inmap[0][i]) inmap[len(inmap)-1][i] = func(inmap[len(inmap)-1][i]) # Leftmost and rightmost columns excluding top and bottom rows for i in xrange(1, len(inmap)-1): inmap[i][0] = func(inmap[i][0]) inmap[i][len(inmap[0])-1] = func(inmap[i][len(inmap[0])-1]) def translator(filename, mapxforms, textxform): "Apply mapxform to map lines and textxform to non-map lines." global tagstack gzipped = filename.endswith(".gz") if gzipped: unmodified = gzip.open(filename).readlines() else: unmodified = file(filename).readlines() # Pull file into an array of lines, CR-stripping as needed mfile = [] map_only = filename.endswith(".map") terminator = "\n" for line in unmodified: if line.endswith("\n"): line = line[:-1] if line.endswith("\r"): line = line[:-1] if not stripcr: terminator = '\r\n' mfile.append(line) if "map_data" in line: map_only = False # Process line-by-line lineno = baseline = 0 cont = False validate = True unbalanced = False newdata = [] refname = None while mfile: if not map_only: line = mfile.pop(0) if verbose >= 3: sys.stdout.write(line + terminator) lineno += 1 # Check for one certain error condition if line.count("{") and line.count("}"): refname = line[line.find("{"):line.rfind("}")] # Ignore all-caps macro arguments. if refname == refname.upper(): pass elif 'mask=' in line and not (refname.endswith("}") or refname.endswith(".mask")): print \ '"%s", line %d: fatal error, mask file without .mask extension (%s)' \ % (filename, lineno+1, refname) sys.exit(1) # Exclude map_data= lines that are just 1 line without # continuation, or which contain {}. The former are # pathological and the parse won't handle them, the latter # refer to map files which will be checked separately. if map_only or (("map_data=" in line or "mask=" in line) and line.count('"') in (1, 2) and line.count('""') == 0 and line.count("{") == 0 and line.count("}") == 0 and not within('time')): outmap = [] add_border = True add_usage = True have_header = have_delimiter = False maskwarn = False maptype = None if map_only: if filename.endswith(".mask"): maptype = "mask" else: maptype = "map" else: leadws = leader(line) if "map_data" in line: maptype = "map" elif "mask" in line: maptype = "mask" baseline = lineno cont = True if not map_only: fields = line.split('"') if fields[1].strip(): mfile.insert(0, fields[1]) if len(fields) == 3: mfile.insert(1, '"') if verbose >= 3: print "*** Entering %s mode on:" % maptype print mfile # Gather the map header (if any) and data lines savedheaders = [] while cont and mfile: line = mfile.pop(0) if verbose >= 3: sys.stdout.write(line + terminator) lineno += 1 # This code supports ignoring comments and header lines if len(line) == 0 or line[0] == '#' or '=' in line: if '=' in line: have_header = True if 'border_size' in line: add_border = False if "usage" in line: add_usage = False usage = line.split("=")[1].strip() if usage == 'mask': add_border = False if filename.endswith(".map"): print "warning: usage=mask in file with .map extension" elif usage == 'map': if filename.endswith(".mask"): print "warning: usage=map in file with .mask extension" if len(line) == 0: have_delimiter = True savedheaders.append(line + terminator) continue if '"' in line: cont = False if verbose >= 3: print "*** Exiting map mode." line = line.split('"')[0] if line: if ',' in line: fields = line.split(",") else: fields = [x for x in line] outmap.append(fields) if not maskwarn and maptype == 'map' and "_f" in line: print \ '"%s", line %d: warning, fog in map file' \ % (filename, lineno+1) maskwarn = True # Checking the outmap length here is a bit of a crock; # the one-line map we don't want to mess with is in the # NO_MAP macro. if len(outmap) == 1: add_border = add_usage = False # Deduce the map type if not map_only: if maptype == "map": newdata.append(leadws + "map_data=\"") elif maptype == "mask": newdata.append(leadws + "mask=\"") original = copy.deepcopy(outmap) for transform in mapxforms: for y in xrange(len(outmap)): transform(filename, baseline, outmap, y) if maptype == "mask": add_border = False if add_border: print '%s, "line %d": adding map border...' % \ (filename, baseline) newdata.append("border_size=1" + terminator) have_header = True # Start by duplicating the current outermost ring outmap = [outmap[0]] + outmap + [outmap[-1]] for i in xrange(len(outmap)): outmap[i] = [outmap[i][0]] + outmap[i] + [outmap[i][-1]] # Strip villages out of the edges outermap(lambda n: re.sub(r"\^V[a-z]+", "", n), outmap) # Strip keeps out of the edges outermap(lambda n: re.sub(r"K([a-z]+)", r"C\1", n), outmap) # Strip the starting positions out of the edges outermap(lambda n: re.sub(r"[1-9] ", r"", n), outmap) # Turn big trees on the edges to ordinary forest hexes outermap(lambda n: n.replace(r"Gg^Fet", r"Gs^Fp"), outmap) if add_usage: print '%s, "line %d": adding %s usage header...' % \ (filename, baseline, maptype) newdata.append("usage=" + maptype + terminator) have_header = True newdata += savedheaders if have_header and not have_delimiter: newdata.append(terminator) for y in xrange(len(outmap)): newdata.append(",".join(outmap[y]) + terminator) # All lines of the map are processed, add the appropriate trailer if not map_only: newdata.append("\"" + terminator) elif "map_data=" in line and (line.count("{") or line.count("}")): newline = line refre = re.compile(r"\{@?([^A-Z].*)\}").search(line) if refre: mapfile = refre.group(1) if not mapfile.endswith(".map") and is_map(mapfile): newline = newline.replace(mapfile, mapfile + ".map") newdata.append(newline + terminator) if newline != line: if verbose > 0: print 'wmllint: "%s", line %d: %s -> %s.' % (filename, lineno, line, newline) elif "map_data=" in line and line.count('"') > 1: print 'wmllint: "%s", line %d: one-line map.' % (filename, lineno) newdata.append(line + terminator) else: # Handle text (non-map) lines. It can use within(). newline = textxform(filename, lineno, line) newdata.append(newline + terminator) fields = newline.split("#") trimmed = fields[0] destringed = re.sub('"[^"]*"', '', trimmed) # Ignore string literals comment = "" if len(fields) > 1: comment = fields[1] # Now do warnings based on the state of the tag stack. if not unbalanced: for instance in re.finditer(r"\[\/?\+?([a-z][a-z_]*[a-z])\]", destringed): tag = instance.group(1) attributes = [] closer = instance.group(0)[1] == '/' if not closer: tagstack.append((tag, {})) else: if len(tagstack) == 0: print '"%s", line %d: closer [/%s] with tag stack empty.' % (filename, lineno+1, tag) elif tagstack[-1][0] != tag: print '"%s", line %d: unbalanced [%s] closed with [/%s].' % (filename, lineno+1, tagstack[-1][0], tag) else: if validate: validate_on_pop(tagstack, tag, filename, lineno) tagstack.pop() if tagstack: for instance in re.finditer(r'([a-z][a-z_]*[a-z])\s*=(.*)', trimmed): attribute = instance.group(1) value = instance.group(2) if '#' in value: value = value.split("#")[0] tagstack[-1][1][attribute] = value.strip() if "wmllint: validate-on" in comment: validate = True if "wmllint: validate-off" in comment: validate = False if "wmllint: unbalanced-on" in comment: unbalanced = True if "wmllint: unbalanced-off" in comment: unbalanced = False if "wmllint: match" in comment: comment = comment.strip() try: fields = comment.split("match ", 1)[1].split(" with ", 1) if len(fields) == 2: notepairs.append((fields[0], fields[1])) except IndexError: pass # It's an error if the tag stack is nonempty at the end of any file: if tagstack: print '"%s", line %d: tag stack nonempty (%s) at end of file.' % (filename, lineno, tagstack) tagstack = [] if iswml(filename): # Perform checks that are purely local. This is an # optimization hack to reduce parsing overhead. for nav in WmllintIterator(newdata, filename): try: (key, prefix, value, comment) = parse_attribute(nav.text) local_sanity_check(filename, nav, key, prefix, value, comment) except TypeError: key = prefix = value = comment = None local_sanity_check(filename, nav, key, prefix, value, comment) # Perform file-global semantic sanity checks newdata = global_sanity_check(filename, newdata) # OK, now perform WML rewrites newdata = hack_syntax(filename, newdata) # Run everything together filetext = "".join(newdata) transformed = filetext else: # Map or mask -- just run everything together transformed = "".join(newdata) # Simple check for unbalanced macro calls unclosed = None linecount = 1 startline = None quotecount = 0 display_state = False singleline = False for i in xrange(len(transformed)): if transformed[i] == '\n': if singleline: singleline = False if not display_state and quotecount % 2 and transformed[i:i+2] != "\n\n" and transformed[i-1:i+1] != "\n\n": print '"%s", line %d: nonstandard word-wrap style within message' % (filename, linecount) linecount += 1 elif transformed[i-7:i] == "message" and transformed[i] == '=': singleline = True elif re.match(" *wmllint: *display +on", transformed[i:]): display_state = True elif re.match(" *wmllint: *display +off", transformed[i:]): display_state = False elif transformed[i] == '"' and not display_state: quotecount += 1 if quotecount % 2 == 0: singleline = False # Return None if the transformation functions made no changes. if "".join(unmodified) != transformed: return transformed else: return None def inner_spellcheck(nav, value, spelldict): "Spell-check an attribute value or string." # Strip off translation marks if value.startswith("_"): value = value[1:].strip() # Strip off line continuations, they interfere with string-stripping value = value.strip() if value.endswith("+"): value = value[:-1].rstrip() # Strip off string quotes value = string_strip(value) # Discard extraneous stuff value = value.replace("...", " ") value = value.replace("\"", " ") value = value.replace("\\n", " ") value = value.replace("/", " ") value = value.replace("@", " ") value = value.replace(")", " ") value = value.replace("(", " ") value = value.replace("\xe2\x80\xa6", " ") # UTF-8 ellipsis value = value.replace("\xe2\x80\x94", " ") # UTF-8 em dash value = value.replace("\xe2\x80\x93", " ") # UTF-8 en dash value = value.replace("\xe2\x80\x95", " ") # UTF-8 horizontal dash value = value.replace("\xe2\x88\x92", " ") # UTF-8 minus sign value = value.replace("\xe2\x80\x99", "'") # UTF-8 right single quote value = value.replace("\xe2\x80\x98", "'") # UTF-8 left single quote value = value.replace("\xe2\x80\x9d", " ") # UTF-8 right double quote value = value.replace("\xe2\x80\x9c", " ") # UTF-8 left double quote value = value.replace("\xe2\x80\xa2", " ") # UTF-8 bullet value = value.replace("◦", "") # Why is this necessary? value = value.replace("''", "") value = value.replace("female^", " ") value = value.replace("male^", " ") value = value.replace("teamname^", " ") value = value.replace("team_name^", " ") value = value.replace("UI^", " ") value = value.replace("^", " ") if '<' in value: value = re.sub(".*< ref>", "", value) value = re.sub("<[^>]+>text='([^']*)'<[^>]+>", r"\1", value) value = re.sub("<[0-9,]+>", "", value) # Fold continued lines value = re.sub(r'" *\+\s*_? *"', "", value) # It would be nice to use pyenchant's tokenizer here, but we can't # because it wants to strip the trailing quotes we need to spot # the Dwarvish-accent words. for token in value.split(): # Try it with simple lowercasing first lowered = token.lower() if d.check(lowered): continue # Strip leading punctuation and grotty Wesnoth highlighters # Last char in this regexp is to ignore concatenation signs. while lowered and lowered[0] in " \t(`@*'%_+": lowered = lowered[1:] # Not interested in interpolations or numeric literals if not lowered or lowered.startswith("$"): continue # Suffix handling. Done in two passes because some # Dwarvish dialect words end in a single quote while lowered and lowered[-1] in "_-*).,:;?!& \t": lowered = lowered[:-1] if lowered and spelldict.check(lowered): continue; while lowered and lowered[-1] in "_-*').,:;?!& \t": lowered = lowered[:-1] # Not interested in interpolations or numeric literals if not lowered or lowered.startswith("$") or lowered[0].isdigit(): continue # Nuke balanced string quotes if present lowered = string_strip(lowered) if lowered and spelldict.check(lowered): continue # No match? Strip possessive suffixes and try again. elif lowered.endswith("'s") and spelldict.check(lowered[:-2]): continue # Hyphenated compounds need all their parts good if "-" in lowered: parts = lowered.split("-") if [w for w in parts if not w or spelldict.check(w)] == parts: continue # Modifier literals aren't interesting if re.match("[+-][0-9]", lowered): continue # Match various onomatopoetic exclamations of variable form if re.match("hm+", lowered): continue if re.match("a+[ur]*g+h*", lowered): continue if re.match("(mu)?ha(ha)*", lowered): continue if re.match("ah+", lowered): continue if re.match("no+", lowered): continue if re.match("no+", lowered): continue if re.match("um+", lowered): continue if re.match("aw+", lowered): continue if re.match("o+h+", lowered): continue if re.match("s+h+", lowered): continue nav.printError('possible misspelling "%s"' % token) def spellcheck(fn, d): "Spell-check a file using an Enchant dictionary object." local_spellings = [] # Accept declared spellings for this file # and for all directories above it. up = fn while True: if not up or is_root(up): break else: local_spellings += declared_spellings.get(up,[]) up = os.path.dirname(up) local_spellings = [w for w in local_spellings if not d.check(w)] #if local_spellings: # print "%s: inherited local spellings: %s" % (fn, local_spellings) map(d.add_to_session, local_spellings) # Process this individual file for nav in WmllintIterator(filename=fn): #print "element=%s, text=%s" % (nav.element, `nav.text`) # Recognize local spelling exceptions if not nav.element and "#" in nav.text: comment = nav.text[nav.text.index("#"):] words = re.search("wmllint: local spellings? (.*)", comment) if words: for word in words.group(1).split(): word = word.lower() if not d.check(word): d.add_to_session(word) local_spellings.append(word) else: nav.printError("spelling '%s' already declared" % word) #if local_spellings: # print "%s: with this file's local spellings: %s" % (fn,local_spellings) for nav in WmllintIterator(filename=fn): # Spell-check message and story parts if nav.element in spellcheck_these: # Special case, beyond us until we can do better filtering.. # There is lots of strange stuff in text- attributes in the # helpfile(s). if nav.element == 'text=' and '[help]' in nav.ancestors(): continue # Remove pango markup if "<" in nav.text or ">" in nav.text or '&' in nav.text: nav.text = pangostrip(nav.text) # Spell-check the attribute value (key, prefix, value, comment) = parse_attribute(nav.text) if "no spellcheck" in comment: continue inner_spellcheck(nav, value, d) # Take exceptions from the id fields if nav.element == "id=": (key, prefix, value, comment) = parse_attribute(nav.text) value = string_strip(value).lower() if value and not d.check(value): d.add_to_session(value) local_spellings.append(value) #if local_spellings: # print "%s: slated for removal: %s" % (fn, local_spellings) for word in local_spellings: try: d.remove_from_session(word) except AttributeError: print "Caught AttributeError when trying to remove %s from dict" % word vctypes = (".svn", ".git", ".hg") def interesting(fn): "Is a file interesting for conversion purposes?" return fn.endswith(".cfg") or is_map(fn) or issave(fn) def allcfgfiles(dir): "Get the names of all interesting files under dir." datafiles = [] if not os.path.isdir(dir): if interesting(dir): if not os.path.exists(dir): sys.stderr.write("wmllint: %s does not exist\n" % dir) else: datafiles.append(dir) else: for root, dirs, files in os.walk(dir): for vcsubdir in vctypes: if vcsubdir in dirs: dirs.remove(vcsubdir) for name in files: if interesting(os.path.join(root, name)): datafiles.append(os.path.join(root, name)) datafiles.sort() # So diffs for same campaigns will cluster in reports return map(os.path.normpath, datafiles) def help(): sys.stderr.write("""\ Usage: wmllint [options] [dir] Convert Battle of Wesnoth WML from older versions to newer ones. Also validates WML to check for errors. Takes any number of directories as arguments. Each directory is converted. If no directories are specified, acts on the current directory. Mode options: Changes wmllint from default conversion mode. Only one mode can be chosen. -h, --help Emit this help message and quit. -d, --dryrun List changes (-v) but don't perform them. -c, --clean Clean up -bak files. -D, --diff Display diffs between converted and unconverted files. -r, --revert Revert the conversion from the -bak files. Other options: -v, --verbose -v lists changes. -v -v names each file before it's processed. -v -v -v shows verbose parse details. -m, --missing Don't warn about tags without side= keys now applying to all sides. -s, --stripcr Convert DOS-style CR/LF to Unix-style LF. -K, --known Suppress check for unknown unit types, recruits, races, scenarios, etc. -S, --nospellcheck Suppress spellchecking -Z, --stringfreeze Suppress repair attempts of newlines in messages For more about wmllint, including how to prevent unwanted conversions and false positive warnings with magic comments, read the introduction in the wmllint file itself. See also: http://wiki.wesnoth.org/Maintenance_tools. """) if __name__ == '__main__': try: (options, arguments) = getopt.getopt(sys.argv[1:], "cdDhmnrsvKSZ", [ "clean", "diffs", "dryrun", "help", "missing", "revert", "stripcr", "verbose", "known", "nospellcheck", "stringfreeze", ]) # -f --future has been removed; there have been no experimental conversions since 1.4 # -p --progress has been removed; similar to existing -v -v except getopt.GetoptError: help() sys.stderr.write('\nAn option you have entered is invalid. Review options and try again.') sys.exit(1) clean = False diffs = False dryrun = False missingside = True revert = False stringfreeze = False stripcr = False verbose = 0 dospellcheck = True inconsistency = False for (switch, val) in options: if switch in ('-h', '--help'): help() sys.exit(0) elif switch in ('-c', '--clean'): clean = True elif switch in ('-d', '--dryrun'): dryrun = True elif switch in ('-D', '--diffs'): diffs = True elif switch in ('-m', '--missing'): missingside = False elif switch in ('-r', '--revert'): revert = True elif switch in ('-s', '--stripcr'): stripcr = True elif switch in ('-Z', '--stringfreeze'): stringfreeze = True elif switch in ('-v', '--verbose'): verbose += 1 elif switch in ('-S', '--nospellcheck'): dospellcheck = False elif switch in ('-K', '--known'): inconsistency = True if dryrun: verbose = max(1, verbose) if clean and revert: sys.stderr.write("wmllint: can't do clean and revert together.\n") sys.exit(1) post15 = False def hasdigit(str): for c in str: if c in "0123456789": return True return False def texttransform(filename, lineno, line): "Resource-name transformation on text lines." original = line # Perform line changes if "wmllint: noconvert" not in original: for (old, new) in linechanges + mapchanges: line = line.replace(old, new) # Perform any base terrain string conversions needed in # [terrain_type] aliasof=, mvt_alias=, and def_alias= attributes. if under("terrain_type"): match = re.search(r"\b(?:aliasof|mvt_alias|def_alias)\s*=(.*)$", line) if match: aliases = match.group() for (old, new) in aliaschanges: aliases = aliases.replace(old, new) line = line.replace(match.group(), aliases) # Perform tag renaming for 1.5. Note: this has to happen before # the sanity check, which assumes [unit] has already been # mapped to [unit_type]. Also, beware that this test will fail to # convert any unit definitions not in conventionally-named # directories -- this is necessary in order to avoid stepping # on SingleUnitWML in macro files. The post15 flag expresses whether # we've seen a [unit_type] and can therefore assume the files have # undergone 1.4 -> 1.5 conversion. global post15 if "units" in filename and not post15: if '[unit_type]' in line: post15 = True else: line = line.replace("[unit]", "[unit_type]") line = line.replace("[+unit]", "[+unit_type]") line = line.replace("[/unit]", "[/unit_type]") # Handle SingleUnitWML or Standard Unit Filter or SideWML # Also, when macro calls have description= in them, the arg is # a SUF being passed in. if tagstack and ((under("unit") and not "units" in filename) or \ standard_unit_filter() or \ under("side") or \ re.search("{[A-Z]+.*description=.*}", line)): if "id" not in tagstack[-1][1] and "_" not in line: line = re.sub(r"\bdescription\s*=", "id=", line) if "name" not in tagstack[-1][1]: line = re.sub(r"user_description\s*=", "name=", line) if "generate_name" not in tagstack[-1][1]: line = re.sub(r"generate_description\s*=", "generate_name=", line) # Now, inside objects... if under("object") and "description" not in tagstack[-1][1]: line = re.sub(r"user_description\s*=", "description=", line) # Alas, WML variable references cannot be converted so # automatically. if ".description" in line: print '"%s", line %d: .description may need hand fixup' % \ (filename, lineno) if ".user_description" in line: print '"%s", line %d: .user_description may need hand fixup' % \ (filename, lineno) # In unit type definitions if under("unit_type") or under("female") or under("unit"): line = line.replace("unit_description=", "description=") line = line.replace("advanceto=", "advances_to=") # Inside themes if within("theme"): line = line.replace("[unit_description]", "[unit_name]") # Report the changes if verbose > 0 and line != original: msg = "%s, line %d: %s -> %s" % \ (filename, lineno, original.strip(), line.strip()) print msg return line try: # If a backslash comes before a quote, it will be interpreted as an # escape to a literal quote rather than a Windows directory delimiter, # causing Windows to find "no such file or directory". This block deals # with this issue, but it is impossible to handle all cases if multiple # (intended) arguments are involved. We also activate globbing on # Windows, if there is a wildcard. if arguments and sys.platform == 'win32': wildcard = False ugly = False newargs = [] for i,arg in enumerate(arguments): if not wildcard and '*' in arg: wildcard = True from glob import glob if '"' in arg: if len(arguments)-i > 1: ugly = True break test = arg.split('"', 1) if " " in test[1].lstrip(): ugly = True break sys.stderr.write('\n\nWARNING!! A backslash followed by a quote (\\") is interpreted not as a directory separator and an argument delimiter, but as an escape to a literal quote. Two quotes together ("") will also be interpreted as including a literal quote. Your system sees the file/directory you are targeting as:\n\n%s\n\nAlthough wmllint believes it can resolve this particular instance, please do not repeat this in the future. (If you are using a final backslash at the end of your argument, it is not necessary; also, frontslashes will also be recognized as directory separators.)\n\n\n'%arg) arguments.remove(arg) arg = re.sub(r'([^ ])"([^ ])', r'\1\\\2', arg) for new in arg.rstrip('"').split('"'): arguments.insert(i, new.strip()) i += 1 sys.stderr.write('wmllint: resolving address as: %s\n' % new.strip()) if ugly: print >>sys.stderr, """ WARNING!! A backslash followed by a quote (\\"), or two quotes (""), is misinterpreted by your system to mean you want a literal quote character, not a path encloser: %s After exiting this message: hit the up arrow key, edit your command, and press Enter. """ % re.sub(r'"', '-->>"<<--', arg) moreugly = raw_input('Press "H" if you need more help, or Enter to exit: ') if moreugly.startswith(('h', 'H')): print >>sys.stderr, """ Explanation: Windows' use of the backslash as a directory separator is clashing with the use of the backslash as an escape. As an escape, the backslash tells your system that you want a normally special character to be its literal self (or sometimes, that you want a normally ordinary character to have a special meaning). Your system interprets '\\"' as an escape for a literal quote character. Two quotes together are also interpreted as a literal quote. '"Campaign\\"' is interpreted as 'Campaign"' instead of 'Campaign\\'. '"My Folder\\"Campaign' is interpreted as 'My Folder"Campaign' (not 'My Folder\Campaign'). '"My Folder\Campaign\\" "My Folder\Another_Campaign"' is interpreted as 'My Folder\Campaign" My' and 'Folder\Another_Campaign'. In your case, your system interprets your arguments as: %s Solutions: (1) If your problem is adjoining quotes, delete (or move) one of them ( "folder\Campaign"" -> "folder\Campaign" ) (2) If you are using a final backslash at the end of a directory address, it is not necessary ( "folder\Campaign\\" -> "folder\Campaign" -- but NOT "folder\Campaign\\"file.cfg -> "folder\Campaign"file.cfg ) (3) If you add a second backslash, your system will then escape the backslash instead of the quote ( "folder\Campaign\\" -> "folder\Campaign\\\\" ) (4) Frontslashes will also be recognized as directory separators ( "folder\Campaign\\" -> "folder/Campaign" ) (5) If there are no spaces in your address, it is not necessary to use quotes ( "folder\Campaign\\" -> folder\Campaign -- but NOT "My Folder\Campaign\\" -> My Folder\Campaign ) (6) You can move the affected quote ( M"y Folder\\"Campaign -> M"y Fold"er\Campaign -- but NOT Fil"es and Folders\My Folder\\"Campaign -> Fil"es and Folders\M"y Folder\Campaign ) Reminder: (a) Hit the up arrow key. The up and down arrows move through your command history; one press of the up arrow will take you to your last command. (b) Edit your command. Use any of the solutions described above. (c) Press Enter """ % repr(arguments)[1:-1] sys.exit(2) if wildcard: for arg in arguments: for wild in glob(arg): newargs.append(wild) if newargs: arguments = newargs else: sys.stderr.write('wmllint: wildcard did not match any files or directories (%s)\n'%arguments) sys.exit(1) if not arguments: arguments = ["."] for dir in arguments: ofp = None for fn in allcfgfiles(dir): if verbose >= 2: print fn + ":" backup = fn + "-bak" if clean or revert: # Do housekeeping if os.path.exists(backup): if clean: print "wmllint: removing %s" % backup if not dryrun: os.remove(backup) elif revert: print "wmllint: reverting %s" % backup if not dryrun: if sys.platform == 'win32': os.remove(fn) os.rename(backup, fn) elif diffs: # Display diffs if os.path.exists(backup): fromdate = time.ctime(os.stat(backup).st_mtime) todate = time.ctime(os.stat(fn).st_mtime) fromlines = open(backup, 'U').readlines() tolines = open(fn, 'U').readlines() diff = difflib.unified_diff(fromlines, tolines, backup, fn, fromdate, todate, n=3) sys.stdout.writelines(diff) else: if "~" in fn: print "wmllint: ignoring %s, the campaign server won't accept it." % fn continue # Do file conversions try: changed = translator(fn, [maptransform], texttransform) if changed: print "wmllint: converting", fn if not dryrun: if sys.platform == 'win32' and os.path.exists(backup): os.remove(backup) os.rename(fn, backup) if fn.endswith(".gz"): ofp = gzip.open(fn, "w") ofp.write(changed) ofp.close() else: ofp = open(fn, "w") ofp.write(changed) ofp.close() #except maptransform_error, e: # sys.stderr.write("wmllint: " + `e` + "\n") except: sys.stderr.write("wmllint: internal error on %s\n" % fn) (exc_type, exc_value, exc_traceback) = sys.exc_info() raise exc_type, exc_value, exc_traceback if not clean and not diffs and not revert: # Consistency-check everything we got from the file scans if not inconsistency: consistency_check() # Attempt a spell-check if dospellcheck: try: import enchant d = enchant.Dict("en_US") checker = d.provider.desc if checker.endswith(" Provider"): checker = checker[:-9] print "# Spell-checking with", checker for word in declared_spellings["GLOBAL"]: d.add_to_session(word.lower()) for dir in arguments: ofp = None for fn in allcfgfiles(dir): if verbose >= 2: print fn + ":" spellcheck(fn, d) except ImportError: sys.stderr.write("wmllint: spell check unavailable, install python-enchant to enable\n") except KeyboardInterrupt: print "Aborted" # wmllint ends here