>[...] I just fixed a problem with jreload
>(ClassFormatError due to truncated class data because when loading the
>class from the stream, it does a single read instead of looping reads till
Yeah, we have several of those left. I believe the only truely solid way
of reading a zipfile entry is when it's done as in
>so I'll submit that patch using SF sometime today.
>I assume that when doing a "cvs diff -u", I should do a "clean" build
>first so that none of the build cruft shows up in the patch?
Or trim the cruft from the patch with an editor.
>Also, if I know that I only changed a single source file, can I do the
>"cvs diff -c" on just that file?
>Does the current directory matter when
>doing the diff, or can I be in the directory containing the file? i.e.
>which of the following is preferred:
>"cvs diff -c jython/org/python/core/imp.java"
>"cd jython/org/python/core; cvs diff -c imp.java"
For a single file or single directory patch, I prefer the second
example. Either way, we'll manage.
>> If you *really* want to help out, you should have made a SF bug report
>> with your observation (there is still time to make amends <wink>) and a
>Heh, ok, if I get a chance today I'll do it, lots to get done before my
>vacation starts so no guarantees =)