On Thu, 30 Oct 2014, Michal Sojka <sojkam1@fel.cvut.cz> wrote: > Hi all, > > this is v5 of the search --output=address series. It obsoletes v4 > (id:1414421455-3037-1-git-send-email-sojkam1@fel.cvut.cz). > > I addresses comments from Mark and Tomi. Based on the comments to v4 > and earlier versions, patches 1-4 should be ready for merging. Patch 5 > is a non-controversial part of the controversial --filter-by patch and > could be probably merged after review. I have looked at Patches 1-5 and tested. These look good to me. +1 My only query is in the text output: should the name part be printed as a quoted string. For example currently I get a line of the form Bloggs, Fred <fred@example.com> and I think in theory I could have a real name "Fred <stupid> Bloggs" which this would print without the quotes. For the other formats it is much less of a problem because the name and address are clearly separated. I am happy with an answer of the form "for robust parseable results use a structured format" which is what we say for search for example. I just thought I would mention it in case you thought the quoted form was more useful for consumers. Best wishes Mark > > Patch 6 needs at least a review and patch 7 needs more discussion. > > Changes from v4: > > - patch changed to commit in commit messages > - opt->format changed to format > - Added comments to process_* functions > - duplicite changed to duplicate > - check_duplicate changed to is_duplicate > - Deduplication was split into two commits: basic deduplication > without a command line option and configurable deduplication with > --fiter-by. > > Changes from v3: > > - `o' renamed to `opt'. > - Conversion of --output from keyword to keyword-flags is now a > separate patch. > - Structured output formats print name and address separately. > - Added test for --format=json. > - Changed --filter-by default to nameaddr. In v2, the default was > addrfold, in v3 the default was no filtering at all. I believe that > Mark's suggestion to make nameaddr the default is good trade off. > - Added new --output=count > - Minor style fixes > - Few typos fixed > - There is no way to output unfiltered (duplicite) addresses. > Hopefully, the introduction of --output=count is sufficient > replacement for this "feature". > > Cheers, > -Michal > > > Jani Nikula (1): > cli: Add support for parsing keyword-flag arguments > > Michal Sojka (6): > cli: search: Refactor passing of command line options > cli: search: Convert --output to keyword-flag argument > cli: search: Add --output={sender,recipients} > cli: search: Do not output duplicate addresses > cli: search: Add --output=count > cli: search: Add --filter-by option to configure address filtering > > command-line-arguments.c | 6 +- > command-line-arguments.h | 1 + > completion/notmuch-completion.bash | 8 +- > completion/notmuch-completion.zsh | 4 +- > doc/man1/notmuch-search.rst | 66 ++++++- > notmuch-search.c | 388 +++++++++++++++++++++++++++++-------- > test/T090-search-output.sh | 137 +++++++++++++ > test/T095-search-filter-by.sh | 64 ++++++ > test/T410-argument-parsing.sh | 3 +- > test/arg-test.c | 9 + > 10 files changed, 604 insertions(+), 82 deletions(-) > create mode 100755 test/T095-search-filter-by.sh > > -- > 2.1.1 > > _______________________________________________ > notmuch mailing list > notmuch@notmuchmail.org > http://notmuchmail.org/mailman/listinfo/notmuch