Tsantilas Christos wrote:
Amos Jeffries wrote:
Finally got around to running it myself on just src/*.*
For me md5checker detects bungled files. I haven't looked further than
those files to see if other directories are bunged as well.
md5checker.sh assures us the printable byte order is identical, yes?
This script just removes all spaces,tabs and newlines from the files
(converted and original) and computes its md5. It fails if a comment has
moved to an other position.
eg the code:
int i;/*Test comment*/
converted to:
/*Test comment*/
int i;
access_log.cc - converts the prepared array of log format tokens
removing the \n placed to make the list human readable.
This is a big issue. The array entries which are properly documented
show up worst.
yep true, if you are using astyle 1.21. I test it with astyle 1.22 too
and looks OK know :-)
Okay, in that case we need some form of version test to ensure the
installed astyle version is usable. I guess I'll have to locate a 1.22.
FWIW: I checked my astyle version against the one listed in formater.pl
to check that it was new enough. The fail still shows 1.21.
gopher.cc - I couldn't find it. But the diff itself could hide a lot.
Some fairly large chunks of code being shifted left by astyle and get
fully shuffled by a diff.
Of the rest all I could see was comments being shifted inside { .
which is okay.
dnsserver.cc
HttpHdrRange.cc
snmp_agent.cc
Amos
Amos
--
Please use Squid 2.7.STABLE4 or 3.0.STABLE9
