reassign 442100 aspell-nl tag 442100 patch thanks On Sat, Sep 15, 2007 at 10:06:38AM +0200, Sven Joachim wrote: > reassign 442100 dictionaries-common > thanks > > Robbert Haarman <[EMAIL PROTECTED]> writes: > > > Package: emacs > > Version: 21.4a+1-3etch1 > > Severity: normal > > > > > > Emacs spell checking (both ispell-buffer and flyspell-buffer) fails with > > Invalid regexp: "Invalid range end" when the language is set to dutch > > (M-x ispell-change-language RET nederlands RET). > > > > This may be the same problem as in bug #351846 > > (/usr/bin/emacs-21.2: ispell w/dutch dictionary gives Invalid regexp: > > "Invalid range end") and/or bug #296734 (idutch: emacs integration > > broken), both of which are supposed to have been fixed. > > > > However, rather than using ispell and idutch, I am using aspell > > (0.60.4-4) and aspell-nl (0.1e-44). > > > > I can spell check files fine (both English and Dutch) when I run aspell > > from the command line. I can also spell check from Emacs if the language > > is set to english. The only combination that fails is spell checking > > from Emacs with the language set to nederlands (which is Dutch for > > "Dutch"). > > I can reproduce this on sid with emacs22, the error message being > slightly different: Invalid regexp: "Range striding over charsets". > > With emacs-snapshot and Emacs' version of ispell.el, however, > spell-checking with aspell-nl works fine, so I think the bug is in > dictionaries-common's version of ispell.el; reassigning the bug.
Strange, things are working for me in etch with aspell-nl 1:0.1e-44 and emacs21 21.4a+1-3etch1. Tried with etch dictionaries-common (0.70.10) as well as with most recent version. However I fully reproduce the reported bug variant with emacs22 as well as the original report with emacs 21.4a+1-5.1, both under sid. Anyway the underlying problem is that handling ranges of 8bit chars from emacs have been a moving target for some time. Ranges like [\300-\305] were previously allowed, as well as ranges like [À-Å]. [\300-\305] support was removed first and seems that [À-Å] support has been removed with emacs22 and probably latests emacs21 versions. This problem should also appear with ispell, since currently the same casechars description is being used for both, and is emacs who complains, not the spellchecker. The fix is to use the full chars list, as in attached patch, based on emacs-cvs changes in Revision 1.154, (ispell-dictionary-alist-4): Rewrite the CASECHARS and NOT-CASECHARS regular expressions of the "nederlands" and "nederlands8" dictionaries to prevent a "Range striding over charsets" error. Reassigning to aspell-nl and tagging +patch. [Thijs, by the way, why is your 957D58CF key not in Debian?] -- Agustin
diff -u dutch-1.00/debian/changelog dutch-1.00/debian/changelog --- dutch-1.00/debian/changelog +++ dutch-1.00/debian/changelog @@ -1,3 +1,10 @@ +dutch (1:1.00-1.1) unstable; urgency=low + + * debian/idutch.info-ispell: Used explicit chars (in octal) + instead of ranges (Closes: #442100). + + -- Agustin Martin Domingo <[EMAIL PROTECTED]> Tue, 18 Sep 2007 17:12:40 +0200 + dutch (1:1.00-1) unstable; urgency=low [ Kurt Roeckx & Thijs Kinkhorst ] diff -u dutch-1.00/debian/idutch.info-ispell dutch-1.00/debian/idutch.info-ispell --- dutch-1.00/debian/idutch.info-ispell +++ dutch-1.00/debian/idutch.info-ispell @@ -1,8 +1,8 @@ Language: nederlands (Dutch) Hash-Name: dutch Emacsen-Name: nederlands -Casechars: [A-Za-zÀ-ÅÇÈ-ÏÒ-ÖÙ-Üà-åçè-ïñò-öù-ü] -Not-Casechars: [^A-Za-zÀ-ÅÇÈ-ÏÒ-ÖÙ-Üà-åçè-ïñò-öù-ü] +Casechars: [A-Za-z\300\301\302\303\304\305\307\310\311\312\313\314\315\316\317\322\323\324\325\326\331\332\333\334\340\341\342\343\344\345\347\350\351\352\353\354\355\356\357\361\362\363\364\365\366\371\372\373\374] +Not-Casechars: [^A-Za-z\300\301\302\303\304\305\307\310\311\312\313\314\315\316\317\322\323\324\325\326\331\332\333\334\340\341\342\343\344\345\347\350\351\352\353\354\355\356\357\361\362\363\364\365\366\371\372\373\374] Otherchars: ['] Many-Otherchars: yes Ispell-Args: -C -d dutch @@ -13,8 +13,8 @@ Language: nederlands (Dutch) Hash-Name: dutch Emacsen-Name: nederlands8 -Casechars: [A-Za-zÀ-ÅÇÈ-ÏÒ-ÖÙ-Üà-åçè-ïñò-öù-ü] -Not-Casechars: [^A-Za-zÀ-ÅÇÈ-ÏÒ-ÖÙ-Üà-åçè-ïñò-öù-ü] +Casechars: [A-Za-z\300\301\302\303\304\305\307\310\311\312\313\314\315\316\317\322\323\324\325\326\331\332\333\334\340\341\342\343\344\345\347\350\351\352\353\354\355\356\357\361\362\363\364\365\366\371\372\373\374] +Not-Casechars: [^A-Za-z\300\301\302\303\304\305\307\310\311\312\313\314\315\316\317\322\323\324\325\326\331\332\333\334\340\341\342\343\344\345\347\350\351\352\353\354\355\356\357\361\362\363\364\365\366\371\372\373\374] Otherchars: ['] Many-Otherchars: yes Additionalchars: àâäçèéêëîïñöóûüÀÂÄÇÈÉÊËÎÏÑÖÓÛÜ
