On 22/02/18 16:47, Doug Goldstein wrote: > On 2/22/18 6:54 AM, Andrew Cooper wrote: >> On 22/02/18 05:52, Doug Goldstein wrote: >>> These changes should make it possible to support modern Pythons as well >>> as the oldest Python 2 still supported. >>> >>> Signed-off-by: Doug Goldstein <car...@cardoe.com> >> To the overall effect, this is definitely a good thing. Some queries >> however... >> >>> diff --git a/xen/tools/compat-build-header.py >>> b/xen/tools/compat-build-header.py >>> index 32421b645b..546371225d 100755 >>> --- a/xen/tools/compat-build-header.py >>> +++ b/xen/tools/compat-build-header.py >>> @@ -23,4 +23,4 @@ pats = [ >>> for line in sys.stdin.readlines(): >>> for pat in pats: >>> line = re.subn(pat[0], pat[1], line)[0] >>> - print line.rstrip() >>> + sys.stdout.write(line.rstrip() + '\n') >> Is there anything wrong with print(line.rstrip()) which is the more >> common way of doing this? > The other scripts in the repo avoided using print as a function which > was only added to Python 2.6 and newer if you use from __future__ import > print_function. The README in the repo still says Python 2.3 and newer > so I had assumed if I broke things prior to Python 2.6 it would be a > show stopper. While Python 2.3 was released July 29th 2003 and Python > 2.6 was released Oct 1st 2008 I figured changing this requirement was > going to be met with the same difficultly I encountered last time > suggesting that we only support versions of build tools released in the > past decade.
The print function, or lack thereof, doesn't matter. The above works with all of python 2, because it parses as "print_stmt, expression of type str" The only bit where it goes wonky is if you try to use any of the optional print function parameters, at which point Py2 will complain. >>> diff --git a/xen/tools/compat-build-source.py >>> b/xen/tools/compat-build-source.py >>> index 595bc3ff58..8101290ebe 100755 >>> --- a/xen/tools/compat-build-source.py >>> +++ b/xen/tools/compat-build-source.py >>> @@ -26,4 +26,4 @@ for pat in pats: >>> for line in sys.stdin.readlines(): >>> for pat in pats: >>> line = re.sub(pat[0], pat[1], line) >>> - print line.rstrip() >>> + sys.stdout.write(line.rstrip() + '\n') >>> diff --git a/xen/tools/gen-cpuid.py b/xen/tools/gen-cpuid.py >>> index 613b909c3d..d64f257816 100755 >>> --- a/xen/tools/gen-cpuid.py >>> +++ b/xen/tools/gen-cpuid.py >>> @@ -3,6 +3,13 @@ >>> >>> import sys, os, re >>> >>> +if (sys.version_info > (3, 0)): >>> + def long(x): >>> + return x >> Strictly speaking, return int(x), however... >> >>> + >>> + def xrange(x): >>> + return range(x) >>> + >>> class Fail(Exception): >>> pass >>> >>> @@ -98,13 +105,13 @@ def parse_definitions(state): >>> def featureset_to_uint32s(fs, nr): >>> """ Represent a featureset as a list of C-compatible uint32_t's """ >>> >>> - bitmap = 0L >>> + bitmap = long(0) >>> for f in fs: >>> - bitmap |= 1L << f >>> + bitmap |= long(1) << f >> Having tested this out, I think the L suffixes can just be dropped: >> >> Python 2.4.3 (#1, Sep 21 2011, 20:06:00) >> [GCC 4.1.2 20080704 (Red Hat 4.1.2-51)] on linux2 >> Type "help", "copyright", "credits" or "license" for more information. >>>>> bitmap = 0 >>>>> type(bitmap) >> <type 'int'> >>>>> for x in xrange(80): bitmap |= 1 << x >> ... >>>>> type(bitmap) >> <type 'long'> >> I don't recall why I wrote it like this, but I don't think its >> necessary. As a result, you shouldn't need to define long above. >> >> ~Andrew > Python 2.4 and newer will properly extend for you but Python 2.3 will > not and I was aiming to still support Python 2.3. Now this is all > theoretical since I can't find any boxes with Python 2.3 anywhere and > I'm not interested in building it. Oh - lovely :( So we either keep that as it was, or decide that continuing to support a 15 year old thing is bordering on the excessive. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
