| 1 |
Following Sebastian's pyflakes and testing suggestions turned up a few |
| 2 |
bugs in the _get_uris patch. Changes since v1: |
| 3 |
|
| 4 |
Patch #1: |
| 5 |
|
| 6 |
* Fixed: |
| 7 |
|
| 8 |
"""<summary> |
| 9 |
|
| 10 |
to: |
| 11 |
|
| 12 |
""" |
| 13 |
<summary> |
| 14 |
|
| 15 |
and converted some from tabs to spaces where I'd missed them in v1. |
| 16 |
|
| 17 |
Patch #3: |
| 18 |
|
| 19 |
* Fixed docstrings as for patch #1. |
| 20 |
* Fix 'myuri' → 'uri' in _get_file_uri_tuples. |
| 21 |
* Fix 'cmirr' → 'm_uri' in _expand_mirror's urlunparse calls. |
| 22 |
* Fix 'cmirr' → 'locmirr' in _expand_mirror's third_party_mirrors branch. |
| 23 |
* Move “No known mirror” error from _get_uris to _expand_mirror. |
| 24 |
* Calculate 'restrict', 'restrict_fetch', 'restrict_mirror', and |
| 25 |
'force_mirror' in _get_uris. |
| 26 |
* Remove 'force_mirror' from fetch, because it's only used in _get_uris. |
| 27 |
* Fix 'custom_mirrors' → 'custommirrors' in _get_uris invocation. |
| 28 |
|
| 29 |
After these changes, runtests.sh passes with only TODO messages on |
| 30 |
Python 2.7, 3.2, and 3.3, although I'm not sure how thoroughly the |
| 31 |
test suite covers fetch. There are also a few unrelated: |
| 32 |
|
| 33 |
/usr/lib64/python3.3/xml/etree/ElementTree.py:1726: |
| 34 |
DeprecationWarning: This method of XMLParser is deprecated. |
| 35 |
Define doctype() method on the TreeBuilder target. |
| 36 |
parser.feed(data) |
| 37 |
|
| 38 |
warnings in the 3.3 results. |
| 39 |
|
| 40 |
For folks who prefer Git fetches to the emailed patches, my current |
| 41 |
series is at: |
| 42 |
|
| 43 |
git://tremily.us/portage.git fetch-refactor |
| 44 |
|
| 45 |
For those who prefer Git diffs to the above “changes since v1”, the v1 |
| 46 |
version of this series is tagged in my repository as |
| 47 |
fetch-refactor-v1. |
| 48 |
|
| 49 |
Sebastian, I'd normally cc you directly on v2, but I'm not sure what |
| 50 |
the Portage team's conventions are here. Let me know if there is a |
| 51 |
convention, or if you have a personal preference on direct mail vs the |
| 52 |
list. I think projects that encourage cc-ing tend to have reasonable |
| 53 |
numbers of NNTP readers, and I don't know where the |
| 54 |
gentoo-portage-dev@ population falls on that issue. |
| 55 |
|
| 56 |
W. Trevor King (3): |
| 57 |
pym/portage/package/ebuild/fetch.py: Factor out |
| 58 |
_get_checksum_failure_max_tries |
| 59 |
pym/portage/package/ebuild/fetch.py: Factor out _get_fetch_resume_size |
| 60 |
pym/portage/package/ebuild/fetch.py: Factor out _get_uris |
| 61 |
|
| 62 |
pym/portage/package/ebuild/fetch.py | 318 +++++++++++++++++++++--------------- |
| 63 |
1 file changed, 189 insertions(+), 129 deletions(-) |
| 64 |
|
| 65 |
-- |
| 66 |
1.8.5.2.8.g0f6c0d1 |
Archives