We should check the source code for commonly mispelled words such as 'dependancy', 'bugzilla' (lower case), 'arbitary' and so on. This would pick up everything, comments, variables names, etc. Hopefully we don't rely on any software or standards that require us to misspell words ('referer' obviously), which would cause problems. This is only looking for specific words, it is not a dictionary check in any sense, and we certainly wouldn't go looking for spellings that are different between different English dialects. The background for this is bug #142792.

t/006spellcheck.t has been added to check for this. Reassigning to timeless who was going to generate a list of things to look for and fix.

> We should check the source code for commonly mispelled words such as > 'dependancy', 'bugzilla' (lower case), 'arbitary' and so on. Matty: why do you think that coding this will ever be the best use of any Bugzilla engineer's time? :-) Gerv

"varsion" productmenu.js:161:// - f: a form containing product, component, varsion and template/en/default/search/form.html.tmpl:242: # - f: a form containing product, component, varsion and

I propose the following patch to the spellchecker: Index: t/006spellcheck.t =================================================================== RCS file: /cvsroot/mozilla/webtools/bugzilla/t/006spellcheck.t,v retrieving revision 1.3 diff -u -r1.3 006spellcheck.t --- t/006spellcheck.t 10 Jan 2003 23:51:38 -0000 1.3 +++ t/006spellcheck.t 3 Jul 2003 23:33:16 -0000 @@ -35,6 +35,11 @@ paramater existance existant +extention +necces +reciev +seperate +trafic ); $testcount = scalar(@Support::Files::testitems) * scalar(@evilwords); But in fact you may be interested in the following script: http://fgouget.free.fr/typos/typos I developed it independently to automate the search for typos in the Wine source but it works very well on any kind of source: the WineHQ web site, the Linux kernel, Bugzilla, etc. So I'll let you integrate it into 006spellcheck.t if you are interested. You can also find a patch fixing a bunch of spelling errors detected by my script at the following URL: http://fgouget.free.fr/tmp/bugzilla-20030703.diff

Please, have mercy, make it stop! The current list already bloats the textual output of runtests.sh --verbose by 3 x the number of CGIs. Turning that into 8 x the number of CGIs would have the sole effect of making me scroll three times as far to get past it before I got to the output I wanted in my buffer. The gain achieved by this test was marginal at best, even originally - and the more words you add, the more inconvenient it becomes. Gerv

I agree; the test either need to output only one line per .cgi or (my preference) be eliminated. To keep adding words (w/each word being another run through all .cgi's) is not ideal, IMHO. The former could probably be accomplished using a method similar to what's in http://fgouget.free.fr/typos/typos (mentioned in comment 4). The method for accomplishing the later should be obvious.

One line per CGI is fine, why does this need to be eliminated? It performs a function, and we have no reason to remove it. The performance is fine, even on my old 533 box it runs quite speedily. Running over the same CGI before moving onto the next one would probably speed it up even more, as might doing a |-separated regexp search if it became a problem. Premature optimisation, people.

> It performs a function, and we have no reason to remove it. Has it ever really caught a spelling mistake that you (or anyone else) were about to check in? (And would it be a big deal if it did?) Gerv

This has been bugging me for a while. Currently, if I runtests --verbose, I have to scroll back up an extra 394 lines if the failure I want to see if before test 6. This is only going to get worse if we add more words. (The fact that it also takes away five seconds of my life whenever I run runtests is secondary.) Can we please turn this test into a small shell script on mattyt's machine that he runs manually every so often? Gerv

or how about adding a --spelling flag to runtests.pl to include that test, otherwise it's skipped? Tinderbox could continue to run it of course. The output is way too verbose. We should have one test success for each file or one test success for each word being tested for. Not both.

We could have a flag - but I think it's a nice property of the tests that they all the ones we have get run by default. (This leads to my conclusion that we shouldn't have a test that no-one runs.) Having it run by tinderbox but not by developers just means there's a small chance of unspottable tree-breakage - in the unlikely event that someone does make one of those five spelling mistakes. Gerv

Bugzilla 2.20 feature set is now frozen as of 15 Sept 2004. Anything flagged enhancement that hasn't already landed is being pushed out. If this bug is otherwise ready to land, we'll handle it on a case-by-case basis, please set the blocking2.20 flag to '?' if you think it qualifies.

Now after bug 257111, the issue of comment 5 seems to me to have gone away, so I think we can do it.

Checking in defparams.pl; /cvsroot/mozilla/webtools/bugzilla/defparams.pl,v <-- defparams.pl new revision: 1.145; previous revision: 1.144 done Checking in Bugzilla/Config.pm; /cvsroot/mozilla/webtools/bugzilla/Bugzilla/Config.pm,v <-- Config.pm new revision: 1.31; previous revision: 1.30 done Checking in Bugzilla/DB.pm; /cvsroot/mozilla/webtools/bugzilla/Bugzilla/DB.pm,v <-- DB.pm new revision: 1.17; previous revision: 1.16 done Checking in t/006spellcheck.t; /cvsroot/mozilla/webtools/bugzilla/t/006spellcheck.t,v <-- 006spellcheck.t new revision: 1.5; previous revision: 1.4 done