(catching up with old mails)
On Thu, Jun 15, 2006 at 09:35:49PM +0200, Nicolas François wrote:
Here is the status of the Man module.
dir1\dir2 IGN OK OK2 WOK1 WOK2 WOK3 PBS WDIFF
IGN 0 722 1076 45 110 215 102 12
OK 0 428 0 0 0 0 0 0
OK2 0 0 2202 0 0 0 0 0
WOK1 0 0 0 181 0 0 0 0
WOK2 0 0 0 0 454 0 0 0
WOK3 0 0 4 1 1 399 0 0
PBS 0 2 3 1 0 3 339 1
WDIFF 0 2 9 1 2 3 0 21
total: 6339 | 6339
Legend so that other can read it too:
OK Man pages for which diff does not see a difference.
OK2 diff does not see a difference after converting hyphens
to minus signs, `` to ", and '' to " (in both pages).
WOK1 wdiff doesn't see any difference after the same modifications
WOK2 This tries to detect changes in the hyphenation of words.
WOK3 This removes minus signs, and thus detects more changes in
hyphenation. It also mask font changes.
PBS po4a preferred to stop processing the man page (unsupported,...)
WDIFF These are probably bugs in po4a or in the man page
IGN man pages po4a refused to operate on (e.g. were generated by Pod::Man)
And then, the table means that they are 1076 pages on nekral's system that
the version 0.25 ignores and which are in the OK2 category (minor
difference) with the current CVS. 9 pages where WDIFF (probable bug) and are
now OK2 too.
There are no more ignored pages (I used -o generated).
Basically, except from the IGN line, it is nice when a page is going to the
left in this table. The WDIFF column is the one we want to avoid (po4a
process the page without warning, but the normalized page is diffrent from
the original page).
The generated pages do not lower the stats of the man module.
(Moreover, they permitted to detect some issues and to fix some of the
WDIFF of po4a 0.25)
That's quite impressive indeed.
How did you deal with the definitions added by the Pod::Man, for example?
You check that their definition match what you expect from them and then
That's a very good job!
Thou shalt check the array bounds of all strings (indeed, all arrays), for
surely where thou typest ``foo'' someone someday shall type
-- Fifth commandment of the C programmer